Training models in network
The method addresses inefficiencies in federated learning by dynamically selecting entities for AI model updates and using additional training data to enhance convergence and adaptability in heterogeneous network environments.
Patent Information
- Application Number
- PCT/CN2023/143038
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2023-12-29
- Publication Date
- 2025-07-03
Smart Images

Figure CN2023143038_03072025_PF_FP_ABST
Abstract
Description
TRAINING MODELS IN NETWORKFIELD
[0001] Example embodiments of the present disclosure generally relate to the field of communications, and in particular, to methods, apparatuses, systems, a non-transitory computer readable medium, a chip, and a computer program product for communications.BACKGROUND
[0002] In some scenarios, model training (e.g. artificial intelligence (AI) model training) is needed for a plurality of network entities (e.g. devices, apparatuses or functions) . For example, in an application, a group of network entities, e.g. a user equipment (UE) , a server, a network function, a radio access network (RAN) node, etc., each of these network entities has a local model and further has (or host) a private dataset. The local models may be heterogeneous, i.e. different in their structures. In the application, each of these models is to be trained using a collection of the private datasets, without requiring the private datasets to leave their hosting network entities. For example, a solution for training models of a plurality of entities is using federated learning. Federated learning aims to train a common model for all the network entities. In federated learning, a number of network entities (e.g. UEs, servers, network functions, RAN nodes, etc., or a combination thereof) collectively train a common AI model using their private datasets in multiple iterations. In an iteration, each of the network entities sends a local version of the model to an aggregator node. The local version of the model sent from a network entity to the aggregator node is obtained by the network entity through a local training, i.e. by training the model using the network entity’s private dataset based on, for example, a random initialization of the model or a global version of the mode.SUMMARY
[0003] In general, example embodiments of the present disclosure provide a solution for training one or more models in a network.
[0004] It is to be understood that the summary section is not intended to identify key or essential features of embodiments of the present disclosure, nor is it intended to be used to limit the scope of the present disclosure. Other features of the present disclosure will become easily comprehensible through the following description.
[0005] In a first aspect, there is provided a method implemented at a first apparatus. In the method, the first apparatus receives an indication for indicating the first apparatus to perform model updating of one or more models associated with first one or more apparatuses, wherein the first apparatus is one of the first one or more apparatuses. The first apparatus transmits a first notification for notifying that the model updating of the one or more models is completed. In this way, personalized local models may be supported, and the entities for training the models may be dynamically selected according to the network dynamics, thus a selection space is large and it is helpful to train the models.
[0006] In some embodiments, the indication comprises at least one of the following: a list of model identifying information for identifying the one or more models; version information of the one or more models, wherein each of the one or more models has corresponding version information; or one or more status indications, wherein each of the one or more status indications indicates a status of a model among the one or more models, and the status of the model comprises frozen or unfrozen. In this way, an entity obtain the indication may update the models based on the indication in an iteration, thus the entities for training the models may be dynamically selected in each iteration.
[0007] In some embodiments, the indication is carried on a first message, the first message is from a second apparatus, and the first message further comprises values of model parameters of the one or more models. In this way, the indication may be indicated together with values of model parameters of the models to notifying the model updating.
[0008] In some embodiments, the indication is carried on a first message, the first message is from a second apparatus, and the method further comprises: obtaining, at the first apparatus, values of model parameters of the one or more models, wherein the values of model parameters is pre-configured at the first apparatus. Thus, the values of model parameters may be obtained in a pre-configured way.
[0009] In some embodiments, the second apparatus is a part of a data plane of a core network, or the second apparatus is connected to a data plane of a core network. In this way, there is a flexible manner to implement the second apparatus.
[0010] In some embodiments, the indication is transmitted from the second apparatus and via a data plane function to the first apparatus. In this way, a suitable way may be utilized for transmitting the indication, and transmission efficiency is increased.
[0011] In some embodiments, the indication is from a third apparatus, and the method further comprises: receiving, at the first apparatus and from a second apparatus, values of model parameters of the one or more models. In this way, the indication and values of model parameters of the models may be transmitted separately based on actual requirements, and transmission efficiency is increased.
[0012] In some embodiments, the third apparatus is a part of a control plane of the core network. Thus the model updating may be indicated via the control plane, to increase efficiency.
[0013] In some embodiments, at least one of the following: the indication is transmitted from the third apparatus to the first apparatus; or the indication is transmitted from the third apparatus and via a control plane function to the first apparatus. In this way, there is a flexible way to transmit the indication.
[0014] In some embodiments, the method further comprises: updating, at the first apparatus, at least one model among the one or more models. Thus, local models may be trained in a personalized way.
[0015] In some embodiments, the first notification further comprises at least one of the following: a list of model identifying information for identifying the one or more models; or one or more status indications, wherein each of the one or more status indications indicates a status of a model among the one or more models, and the status of the model comprises frozen or unfrozen. Thus, it is helpful for model training in the next iteration.
[0016] In some embodiments, a frozen model is a converged model; or an unfrozen model is an unconverged model. In this way, it is helpful for performing model training based on the model status in the next iteration.
[0017] In some embodiments, at least one of the following: the first notification is transmitted from the first apparatus to a third apparatus; or the first notification is transmitted from the first apparatus and via a control plane function to the third apparatus. In this way, the third apparatus may manage, control or influence the control plane’s decision, e.g. decision on traffic routing, to help to train the models in the next iteration.
[0018] In some embodiments, the method further comprises: receiving, at the first apparatus, a second message for notifying that the first apparatus is selected for performing the model updating. In this way, the entities for training the models may be dynamically selected according to the network dynamics in each iteration, thus a selection space is large, and the training efficiency is improved.
[0019] In some embodiments, the method further comprises: transmitting, at the first apparatus, a positive acknowledgement (ACK) indicating confirmation of the selection. In this way, the first apparatus may decide whether to agree on the selection based on its condition, and it is helpful to improve the training efficiency.
[0020] In some embodiments, the second message is from a fourth apparatus, and the positive ACK is transmitted to the fourth apparatus, wherein the fourth apparatus is configured to perform a policy control function; or the second message is from a fifth apparatus, and the positive ACK is transmitted to the fifth apparatus, wherein the fifth apparatus is a network controller. In this way, a selection of the first apparatus may be notified and confirmed in a flexible way.
[0021] In some embodiments, the method further comprises: receiving, at the first apparatus and from a third apparatus, a third message for notifying the first apparatus to transmit at least one latest version of the one or more models obtained through the model updating; and transmitting, at the first apparatus, the at least one latest version. In this way, the first apparatus may know the apparatus to which the latest version (s) of the model (s) may be transmitted, thus, the entities for storing the at least one latest version may be dynamically selected in each iteration.
[0022] In some embodiments, at least one of the following: the third message is received from the third apparatus via a control plane function; or the at least one latest version is transmitted via a data plane function, wherein the data plane function is configured to transport the at least one latest version to a sixth apparatus among second one or more apparatuses. In this way, a suitable way may be utilized for transmitting the third message, and transmission efficiency is increased.
[0023] In some embodiments, the sixth apparatus is the second apparatus. In this way, there may be the same or different apparatuses to help train the models in each iteration.
[0024] In some embodiments, the control plane function comprises at least one of the following: a network exposure function (NEF) , a policy control function (PCF) , a network storage function (NSF) , a network controller (NWC) , a path management function (PMF) or an access and mobility management function (AMF) . In this way, the control plane function for the solution of the present embodiments may be implemented flexibly.
[0025] In some embodiments, the third message further comprises: model identifying information for identifying at least one model associated with the at least one latest version. In this way, the first apparatus may be notified to send latest version of some of the at least one model, thus transmission efficiency is improved.
[0026] In some embodiments, the first apparatus is one of a terminal apparatus, a sever or a network function. In this way, the entities for model training may be diverse.
[0027] In a second aspect, there is provided a method implemented at a second apparatus. In the method, the second apparatus receives, from a third apparatus, a fourth message for indicating whether additional training data should be used by the second apparatus to perform model updating of one or more models associated with first one or more apparatuses, wherein the second apparatus is one of second one or more apparatuses, and the additional training data is different from training data for performing model updating at the first one or more apparatuses. The second apparatus updates the one or more models based on receiving the fourth message. In this way, personalized local models may be supported, and the entities for training the models may be dynamically selected, in addition, negative impact of data heterogeneity may be mitigated.
[0028] In some embodiments, the fourth message comprises the following: a list of model identifying information for identifying the one or more models; version information of the one or more models, wherein each of the one or more models has corresponding version information; and one or more status indications, wherein each of the one or more status indications indicates a status of a model among the one or more models, and the status of the model comprises frozen or unfrozen. In this way, the negative impact of data heterogeneity of the private datasets may be mitigated.
[0029] In some embodiments, the fourth message further indicates where to obtain the additional training data and / or how to select the additional training data in the event that the additional training data should be used. In this way, the second apparatus does not need to store the additional training data locally to save storage space.
[0030] In some embodiments, updating the one or more models comprises: in the event of the additional training data should be used by the second apparatus to perform the model updating, using the additional training data to update at least one model with status indication indicating unfrozen among the one or more models. In this way, the negative impact of data heterogeneity of the private datasets may be mitigated.
[0031] In some embodiments, the method further comprises: transmitting, at the second apparatus and to a first apparatus among the first one or more apparatuses, an indication for indicating the first apparatus to perform the model updating of the one or more models. In this way, the indication may be transmitted to a dynamically selected entity for training the models.
[0032] In some embodiments, the indication is carried on a first message, and the first message further comprises values of model parameters of the one or more models. In this way, the indication may be indicated together with values of model parameters of the models to notifying the model updating.
[0033] In some embodiments, the indication comprises the following: a list of model identifying information for identifying the one or more models; version information of the one or more models, wherein each of the one or more models has corresponding version information; and one or more status indications, wherein each of the one or more status indications indicates a status of a model among the one or more models, and the status of the model comprises frozen or unfrozen. In this way, personalized local models may be supported, and the entities for training the models may be dynamically selected in each iteration.
[0034] In some embodiments, a frozen model is a converged model; or an unfrozen model is an unconverged model. In this way, it is helpful for performing model training based on the model status in the next iteration.
[0035] In some embodiments, the method further comprises: receiving, at the second apparatus and from the third apparatus, a second notification that the second apparatus is selected for receiving at least one latest version of the one or more models from the first apparatus, wherein the at least one latest version is obtained through the model updating performed by the first apparatus. In this way, the second apparatus selected for receiving at least one latest version of the one or more models may be dynamically selected for each iteration.
[0036] In some embodiments, the second notification further comprises: model identifying information for identifying at least one model associated with the at least one latest version. In this way, receiving the at least one latest version of the one or more models is configurable.
[0037] In some embodiments, the method further comprises: receiving, at the second apparatus and from the third apparatus, a configuration for the second apparatus to obtain one or more latest versions from a further apparatus among the second one or more apparatuses, wherein the one or more latest versions are among a plurality of versions of the one or more models excluding the at least one latest version. In this way, the second apparatus may obtain the one or more latest versions from another apparatus, and it is configurable.
[0038] In some embodiments, the first apparatus is one of a terminal apparatus, a sever or a network function. In this way, the entities for model training may be diverse.
[0039] In some embodiments, the second apparatus is a part of a data plane of a core network; or the second apparatus is connected to a data plane of a core network. In this way, there is a flexible manner to implement the second apparatus.
[0040] In some embodiments, the third apparatus is a part of a control plane of the core network. Thus the model updating may be indicated via the control plane, to increase efficiency.
[0041] In a third aspect, there is provided a method implemented at a third apparatus. In the method, the third apparatus transmits, to a second apparatus, a fourth message for indicating whether additional training data should be used by the second apparatus to perform model updating of one or more models associated with first one or more apparatuses, wherein the second apparatus is one of second one or more apparatuses, and the additional training data is different from training data for performing model updating at the first one or more apparatuses. In this way, personalized local models may be supported, and the network entities for training the models may be dynamically selected, in addition, negative impact of data heterogeneity may be mitigated.
[0042] In some embodiments, the method further comprises: determining, at the third apparatus, one of the first one or more apparatuses as a first apparatus for performing the model updating. In this way, the network entities for training the models may be dynamically selected.
[0043] In some embodiments, the method further comprises: determining, at the third apparatus, the second apparatus by selecting an application location associated with the second apparatus among at least one application location. In this way, the second apparatus may be selected based on the application location to improve the training efficiency.
[0044] In some embodiments, the method further comprises: transmitting, at the third apparatus and to the second apparatus, a configuration for the second apparatus to obtain version information of the one or more models from a further apparatus among the second one or more apparatuses. In this way, the second apparatus may obtain the one or more latest versions from another apparatus, and it is configurable.
[0045] In some embodiments, the method further comprises: transmitting, at the third apparatus and to a first apparatus among the first one or more apparatuses, an indication for indicating the first apparatus to perform the model updating. The entities for training the models may be dynamically indicated and selected.
[0046] In some embodiments, the indication comprises at least one of the following: a list of model identifying information for identifying the one or more models; version information of the one or more models, wherein each of the one or more models has corresponding version information; or one or more status indications, wherein each of the one or more status indications indicates a status of a model among the one or more models, and the status of the model comprises frozen or unfrozen. In this way, an entity obtain the indication may update the models based on the indication in an iteration, thus the entities for training the models may be dynamically selected in each iteration.
[0047] In some embodiments, the indication is transmitted from the third apparatus to the first apparatus, or the indication is transmitted from the third apparatus and via a control plane function to the first apparatus. In this way, in this way, a suitable way may be utilized for transmitting the indication, and transmission efficiency is increased.
[0048] In some embodiments, the fourth message comprises at least one of the following: a list of model identifying information for identifying the one or more models; version information of the one or more models, wherein each of the one or more models has corresponding version information; or one or more status indications, wherein each of the one or more status indications indicates a status of a model among the one or more models, and the status of the model comprises frozen or unfrozen.
[0049] In some embodiments, a frozen model is a converged model; or an unfrozen model is a unconverged model. In this way, it is helpful for performing model training based on the model status in the next iteration.
[0050] In some embodiments, the fourth message further indicates where to obtain the additional training data and / or how to select the additional training data in the event that the additional training data should be used. In this way, the second apparatus does not need to store the additional training data locally to save storage space.
[0051] In some embodiments, the method further comprises: receiving, at the third apparatus and from a first apparatus among the first one or more apparatuses, a first notification for notifying that the model updating of one or more models performed by the first apparatus is completed; and determining, based on the first notification, at the third apparatus and from the second one or more apparatuses, a sixth apparatus for receiving at least one latest version of the one or more models, wherein the at least one latest version is obtained through model updating performed by the first apparatus, and a latest version corresponds to an updated model. In this way, the entities for training the models may be dynamically selected according to the network dynamics, thus a selection space is large and it is helpful to train the models.
[0052] In some embodiments, the method further comprises: transmitting, at the third apparatus and to the sixth apparatus, a second notification that the sixth apparatus is selected for receiving the at least one latest version from the first apparatus. In this way, the apparatus selected for receiving at least one latest version of the one or more models may be dynamically selected and indicated for each iteration.
[0053] In some embodiments, the second notification further comprises: model identifying information for identifying at least one model associated with the at least one latest version. In this way, the first apparatus may be notified to send latest version of some of the at least one model, thus transmission efficiency is improved, and the model identifying information may be dynamically indicated.
[0054] In some embodiments, the method further comprises: transmitting, at the third apparatus and to the sixth apparatus, a configuration for the sixth apparatus to obtain one or more latest versions from a further apparatus among the second one or more apparatuses, wherein the one or more latest versions are among a plurality of versions of the one or more models excluding the at least one latest version. In this way, the sixth apparatus may obtain the one or more latest versions from another apparatus, and it is configurable.
[0055] In some embodiments, the method further comprises: transmitting, at the third apparatus and to a first apparatus among the first one or more apparatuses, a third message for notifying the first apparatus to transmit at least one latest version of the one or more models obtained through the model updating to a sixth apparatus among the second one or more apparatuses. In this way, the first apparatus may know the apparatus to which the latest version (s) of the model (s) may be transmitted, thus, the entities for storing the at least one latest version may be dynamically selected in each iteration.
[0056] In some embodiments, the sixth apparatus is the second apparatus. In this way, there may be the same or different apparatuses to help train the models in each iteration.
[0057] In some embodiments, the first apparatus is one of a terminal apparatus, a sever or a network function. In this way, the entities for model training may be diverse.
[0058] In some embodiments, the second apparatus is a part of a data plane of a core network; or the second apparatus is connected to a data plane of a core network. In this way, there is a flexible manner to implement the second apparatus.
[0059] In some embodiments, the third apparatus is a part of a control plane of the core network. Thus the model updating may be indicated via the control plane, to increase efficiency.
[0060] In a fourth aspect, there is provided a method implemented at a third apparatus. In the method, the third apparatus transmits, to a fourth apparatus, a fifth message for requesting to influence traffic routing, wherein influencing the traffic routing is for selecting a second apparatus among second one or more apparatuses and at least one candidate apparatus among first one or more apparatuses, the second apparatus and the at least one candidate apparatus jointly perform model updating of one or more models associated with the first one or more apparatuses, and the fourth apparatus is configured to perform a policy control function; and receives, from a fifth apparatus, a sixth message for indicating the second apparatus and the at least one candidate apparatus, wherein the fifth apparatus is configured to be a network controller. In this way, personalized local models may be supported, and the entities for training the models may be dynamically selected according to the network dynamics, thus a selection space is large and it is helpful to train the models.
[0061] In some embodiments, the method further comprises: determining, at the third apparatus, the second apparatus by selecting an application location associated with the second apparatus among at least one application location. In this way, an apparatus corresponding to a suitable application location may be selected as the second apparatus, and the efficiency of training may be improved.
[0062] In some embodiments, the fifth message comprises at least one of the following: information of data traffic identifying data traffic of the at least one application associated with the second one or more apparatuses, wherein the data traffic is to be routed during the traffic routing; information of the first one or more apparatuses; information of at least one location of the at least one application; information of one or more requirements associated with the traffic routing; or information of traffic filtering associated with the data traffic. In this way, it is helpful to determine a suitable traffic routing for the process of model training.
[0063] In some embodiments, the method further comprises: transmitting, at the third apparatus and to the fifth apparatus, feedback for the sixth message, wherein the feedback indicates a positive acknowledgement (ACK) or a negative ACK to the selection for the application location associated with the second apparatus. In this way, the third apparatus may manage / control the second apparatus.
[0064] In some embodiments, the method further comprises: determining, at the third apparatus, a candidate apparatus as a first apparatus among the at least one candidate apparatus, wherein information identifying the first apparatus is included in the positive ACK. In this way, the third apparatus may manage / control the first apparatus.
[0065] In some embodiments, the first apparatus is one of a terminal apparatus, a sever or a network function. In this way, the entities for model training may be diverse.
[0066] In some embodiments, the second apparatus is a part of a data plane of a core network; or the second apparatus is connected to a data plane of a core network. In this way, there is a flexible manner to implement the second apparatus.
[0067] In some embodiments, the third apparatus is a part of a control plane of the core network. Thus the model updating may be indicated via the control plane, to increase efficiency.
[0068] In a fifth aspect, there is provided a method implemented at a fourth apparatus. In the method, the fourth apparatus receives, from a third apparatus, a fifth message for requesting to influence traffic routing, wherein influencing the traffic routing is for selecting a second apparatus among second one or more apparatuses and at least one candidate apparatus among first one or more apparatuses, the second apparatus and the at least one candidate apparatus jointly perform model updating of one or more models associated with the first one or more apparatuses, and the fourth apparatus is configured to perform a policy control function; and generates, based on the fifth message, at least one policy on influencing the traffic routing. In this way, personalized local models may be supported, and the entities for training the models may be dynamically selected according to the network dynamics, thus a selection space is large and it is helpful to train the models.
[0069] In some embodiments, the method further comprises at least one of: transmitting, at the fourth apparatus, the at least one policy to a fifth apparatus configured to be a network controller for determining a second apparatus and the at least one candidate apparatus; or determining, at the fourth apparatus and based on the at least one policy, the second apparatus and the at least one candidate apparatus. In this way, the at least one policy may be generated and performed by a same apparatus, or generated and performed by different apparatuses in a flexibly way.
[0070] In some embodiments, the at least one policy comprises at least one of the following: information of data traffic to be routed in the traffic routing; information of at least one location of at least one application, wherein one of the at least one application is associated with the second apparatus; information of one or more requirements associated with the traffic routing; or information of the first one or more apparatuses. In this way, the at least one policy is configurable, and may include various information according to requirement (s) .
[0071] In some embodiments, the method further comprises: transmitting, at the fourth apparatus and to the third apparatus, a sixth message for indicating a second apparatus and the at least one candidate apparatus. In this way, the third apparatus may manage / control the second apparatus.
[0072] In some embodiments, the sixth message indicates a data plane management event associated with a selection of an application location associated with the second apparatus among at least one application location. In this way, the second apparatus may be selected based on the application location.
[0073] In some embodiments, the method further comprises: receiving, at the fourth apparatus and from the third apparatus, feedback for the sixth message, wherein the feedback indicates a positive acknowledgement (ACK) for confirmation of the selection or a negative ACK for rejection of the selection. In this way, the third apparatus may manage / control the second apparatus.
[0074] In some embodiments, the feedback is the positive ACK, and the positive ACK further comprises: information identifying a candidate apparatus as a first apparatus among the at least one candidate apparatus. In this way, the third apparatus may manage / control the first apparatus.
[0075] In some embodiments, the method further comprises: transmitting, at the fourth apparatus and to the first apparatus, a second message for notifying that the first apparatus is selected to perform the model updating. In this way, the first apparatus may determine whether to agree on the selection based on its condition, and it is helpful to improve the training efficiency.
[0076] In some embodiments, the method further comprises: receiving, at the fourth apparatus and from the first apparatus, a positive ACK for confirmation of performing the model updating or a negative ACK for rejection of performing the model updating. In this way, the first apparatus may notify whether it agrees on the selection based on its condition, so that the fourth apparatus may further determine whether to select a new first apparatus.
[0077] In some embodiments, the method further comprises: transmitting, at the fourth apparatus and to a seventh apparatus for selecting a data plane path for the traffic routing, traffic routing information indicating the traffic routing. In this way, a suitable data plane path is selected so that the information may be transmitted via the data plane.
[0078] In some embodiments, the method further comprises at least one of the following: receiving, at the fourth apparatus and from the seventh apparatus, a third notification that a data plane path has been configured for the traffic routing; or receiving, at the fourth apparatus and from the seventh apparatus, a response to the transmission of the traffic routing information, wherein the response comprises the third notification that the data plane path has been configured for the traffic routing. The way of implementing the notification is flexible.
[0079] In some embodiments, the traffic routing information comprises the at least one of the following: information identifying an application location associated with the second apparatus among at least one application location; information identifying a first apparatus, wherein the first apparatus is determined among the at least one candidate apparatus; information associated with the at least one policy; information of a data plane path for the traffic routing. In this way, the fourth apparatus may know the configured traffic routing path.
[0080] In some embodiments, the first apparatus is one of a terminal apparatus, a sever or a network function. In this way, the entities for model training may be diverse.
[0081] In some embodiments, one of the second one or more apparatuses is a part of a data plane of a core network; or one of the second one or more apparatuses is connected to a data plane of a core network. In this way, there is a flexible manner to implement the second apparatus.
[0082] In some embodiments, the third apparatus is a part of a control plane of the core network. Thus the model updating may be indicated via the control plane, to increase efficiency.
[0083] In a sixth aspect, there is provided a method implemented at a fifth apparatus. In the method, the fifth apparatus receives, from a fourth apparatus, at least one policy on influencing traffic routing, wherein influencing the traffic routing is for selecting a second apparatus among second one or more apparatuses and at least one candidate apparatus among first one or more apparatuses, the second apparatus and the at least one candidate apparatus jointly perform model updating of one or more models associated with the first one or more apparatuses, the fourth apparatus is configured to perform a policy control function, and the fifth apparatus is configured to be a network controller; selects, based on the at least one policy, the second apparatus and the at least one candidate apparatus; and transmits, to a third apparatus, a sixth message for indicating the second apparatus and the at least one candidate apparatus, wherein the third apparatus is configured to control the second one or more apparatuses. In this way, personalized local models may be supported, and the entities for training the models may be dynamically selected according to the network dynamics, thus a selection space is large and it is helpful to train the models.
[0084] In some embodiments, the at least one policy comprises at least one of the following: information of data traffic to be routed in the traffic routing; information of at least one location of at least one application, wherein one of the at least one application is associated with the second apparatus; information of one or more requirements associated with the traffic routing; or information of the first one or more apparatuses. In this way, the at least one policy is configurable, and may include various information according to requirement (s) .
[0085] In some embodiments, the sixth message indicates a data plane management event associated with a selection of an application location associated with the second apparatus among at least one application location. In this way, the second apparatus may be selected based on the application location.
[0086] In some embodiments, the method further comprises: receiving, at a fifth apparatus and from the third apparatus, feedback for the sixth message, wherein the feedback indicates a positive acknowledgement (ACK) for confirmation of the selection or a negative ACK for rejection of the selection. In this way, the third apparatus may manage / control the second apparatus.
[0087] In some embodiments, the feedback is the positive ACK, and the positive ACK further comprises: information identifying a candidate apparatus as a first apparatus among the at least one candidate apparatus. In this way, the third apparatus may manage / control the fifth apparatus.
[0088] In some embodiments, the method further comprises: transmitting, at a fifth apparatus and to the first apparatus, a second message for notifying that the first apparatus is selected to perform the model updating. In this way, the first apparatus may determine whether to agree on the selection based on its condition, and it is helpful to improve the training efficiency.
[0089] In some embodiments, the method further comprises: receiving, at a fifth apparatus and from the first apparatus, a positive ACK for confirmation of performing the model updating or a negative ACK for rejection of performing the model updating. In this way, the first apparatus may notify whether it agrees on the selection based on its condition, so that the fifth apparatus may further determine whether to select a new first apparatus.
[0090] In some embodiments, the method further comprises: transmitting, at a fifth apparatus and to a seventh apparatus for selecting a data plane path for the traffic routing, traffic routing information indicating the traffic routing. In this way, a suitable data plane path is selected so that the information may be transmitted via the data plane.
[0091] In some embodiments, the traffic routing information comprises the at least one of the following: information identifying an application location associated with the second apparatus among at least one application location; information identifying a first apparatus, wherein the first apparatus is determined among the at least one candidate apparatus; information associated with the at least one policy; information of a data plane path for the traffic routing. In this way, the fifth apparatus may know the configured traffic routing path.
[0092] In some embodiments, the method further comprises at least one of the following: receiving, at a fifth apparatus and from the seventh apparatus, a third notification that a data plane path has been configured for the traffic routing; or receiving, at a fifth apparatus and from the seventh apparatus, a response to the transmission of the traffic routing information, wherein the response comprises the third notification that the data plane path has been configured for the traffic routing. The way of implementing the notification is flexible.
[0093] In some embodiments, the fifth apparatus and the fourth apparatus are integrated in a single network entity. In this way, the transmission signaling is simplified.
[0094] In some embodiments, first apparatus is one of a terminal apparatus, a sever or a network function. In this way, the entities for model training may be diverse.
[0095] In some embodiments, one of the second one or more apparatuses is a part of a data plane of a core network; or one of the second one or more apparatuses is connected to a data plane of a core network. In this way, there is a flexible manner to implement the second apparatus.
[0096] In some embodiments, the third apparatus is a part of a control plane of the core network. Thus the model updating may be indicated via the control plane, to increase efficiency.
[0097] In a seventh aspect, there is provided a first apparatus. The first apparatus comprises a transceiver and a processor communicatively coupled with the transceiver. The processor is configured to: receive an indication for indicating the first apparatus to perform model updating of one or more models associated with first one or more apparatuses, wherein the first apparatus is one of the first one or more apparatuses; and transmit a first notification for notifying that the model updating of the one or more models is completed.
[0098] In an eighth aspect, there is provided second apparatus. The second apparatus comprises a transceiver and a processor communicatively coupled with the transceiver. The processor is configured to: receive, from a third apparatus, a fourth message for indicating whether additional training data should be used by the second apparatus to perform model updating of one or more models associated with first one or more apparatuses, wherein the second apparatus is one of second one or more apparatuses, and the additional training data is different from training data for performing model updating at the first one or more apparatuses; and update the one or more models based on receiving the fourth message.
[0099] In a ninth aspect, there is provided third apparatus. The third apparatus comprises a transceiver and a processor communicatively coupled with the transceiver. The processor is configured to: transmit, to a second apparatus, a fourth message for indicating whether additional training data should be used by the second apparatus to perform model updating of one or more models associated with first one or more apparatuses, wherein the second apparatus is one of second one or more apparatuses, and the additional training data is different from training data for performing model updating at the first one or more apparatuses.
[0100] In a tenth aspect, there is provided third apparatus. The third apparatus comprises a transceiver and a processor communicatively coupled with the transceiver. The processor is configured to: transmit, to a fourth apparatus, a fifth message for requesting to influence traffic routing, wherein influencing the traffic routing is for selecting a second apparatus among second one or more apparatuses and at least one candidate apparatus among first one or more apparatuses, the second apparatus and the at least one candidate apparatus jointly perform model updating of one or more models associated with the first one or more apparatuses, and the fourth apparatus is configured to perform a policy control function; and receive, from a fifth apparatus, a sixth message for indicating the second apparatus and the at least one candidate apparatus, wherein the fifth apparatus is configured to be a network controller.
[0101] In an eleventh aspect, there is provided fourth apparatus. The fourth apparatus comprises a transceiver and a processor communicatively coupled with the transceiver. The processor is configured to: receive, from a third apparatus, a fifth message for requesting to influence traffic routing, wherein influencing the traffic routing is for selecting a second apparatus among second one or more apparatuses and at least one candidate apparatus among first one or more apparatuses, the second apparatus and the at least one candidate apparatus jointly perform model updating of one or more models associated with the first one or more apparatuses, and the fourth apparatus is configured to perform a policy control function; and generate, based on the fifth message, at least one policy on influencing the traffic routing.
[0102] In a twelfth aspect, there is provided fifth apparatus. The fifth apparatus comprises a transceiver and a processor communicatively coupled with the transceiver. The processor is configured to: receive, from a fourth apparatus, at least one policy on influencing traffic routing, wherein influencing the traffic routing is for selecting a second apparatus among second one or more apparatuses and at least one candidate apparatus among first one or more apparatuses, the second apparatus and the at least one candidate apparatus jointly perform model updating of one or more models associated with the first one or more apparatuses, the fourth apparatus is configured to perform a policy control function, and the fifth apparatus is configured to be a network controller; select, based on the at least one policy, the second apparatus and the at least one candidate apparatus; and transmit, to a third apparatus, a sixth message for indicating the second apparatus and the at least one candidate apparatus, wherein the third apparatus is configured to control the second one or more apparatuses.
[0103] In a thirteenth aspect, there is provided a system. The system comprises a first apparatus, a second apparatus, a third apparatus, a fourth apparatus and a fifth apparatus. The third apparatus is configured to transmit a fourth message for indicating whether additional training data should be used by the second apparatus to perform model updating of one or more models associated with first one or more apparatuses, wherein the second apparatus is one of second one or more apparatuses, and the additional training data is different from training data for performing model updating at the first one or more apparatuses; the second apparatus is configured to receive, from the third apparatus, the fourth message, and update the one or more models based on receiving the fourth message; the first apparatus is configured to receive, from the second apparatus or the third apparatus, an indication for indicating the first apparatus to perform the model updating, and transmit a first notification for notifying that the model updating of the one or more models is completed, wherein the first apparatus is one of the first one or more apparatuses; the third apparatus is further configured to control the second one or more apparatuses, and transmit, to the fourth apparatus, a fifth message for requesting to influence traffic routing, wherein influencing the traffic routing is for selecting the second apparatus and at least one candidate apparatus among the first one or more apparatuses, the second apparatus and the at least one candidate apparatus jointly perform the model updating; the fourth apparatus is configured to receive the fifth message from the third apparatus, to generate, based on the fifth message, at least one policy on influencing the traffic routing, and transmit the at least one policy to the fifth apparatus; and the fifth apparatus is configured to be a network controller and to receive the at least one policy from the fourth apparatus, and the fifth apparatus is further configured to select, based on the at least one policy, the second apparatus and the at least one candidate apparatus, and transmit to the third apparatus, a sixth message for indicating the second apparatus and the at least one candidate apparatus.
[0104] In a fourteenth aspect, there is provided a method implemented at a system. The system comprises a first apparatus, a second apparatus, a third apparatus, a fourth apparatus and a fifth apparatus. In the method, the third apparatus transmits a fourth message for indicating whether additional training data should be used by a second apparatus to perform model updating of one or more models associated with first one or more apparatuses, wherein the second apparatus is one of second one or more apparatuses, and the additional training data is different from training data for performing model updating at the first one or more apparatuses; the second apparatus receives, from the third apparatus, the fourth message, and updates the one or more models based on receiving the fourth message; the first apparatus receives, from the second apparatus or the third apparatus, an indication for indicating the first apparatus to perform the model updating, and transmits a first notification for notifying that the model updating of the one or more models is completed, wherein the first apparatus is one of the first one or more apparatuses; the third apparatus controls the second one or more apparatuses, and transmits, to a fourth apparatus, a fifth message for requesting to influence traffic routing, wherein influencing the traffic routing is for selecting the second apparatus and at least one candidate apparatus among the first one or more apparatuses, and the second apparatus and the at least one candidate apparatus jointly perform the model updating; the fourth apparatus receives the fifth message from the third apparatus, to generate, based on the fifth message, at least one policy on influencing the traffic routing, and transmits the at least one policy to a fifth apparatus; and the fifth apparatus receives the at least one policy from the fourth apparatus, and the fifth apparatus is further configured to be a network controller and configured to select, based on the at least one policy, the second apparatus and the at least one candidate apparatus, and transmit to the third apparatus, a sixth message for indicating the second apparatus and the at least one candidate apparatus.
[0105] In a fifteenth aspect, there is provided a system. The system comprises at least one of the following: the first apparatus of the seventh aspect, the second apparatus of the eighth aspect, the third apparatus of the ninth aspect or the tenth aspect, the fourth apparatus of the eleventh aspect, or the fifth apparatus of the twelfth aspect.
[0106] In a sixteenth aspect, there is provided a non-transitory computer readable medium comprising computer program stored thereon, the computer program, when executed on at least one processor, causing the at least one processor to perform the method of any one of the first aspect to the sixth aspect.
[0107] In a seventeenth aspect, there is provided a chip comprising at least one processing circuit configured to perform the method of any one of the first aspect to the sixth aspect.
[0108] In an eighteenth aspect, there is provided a computer program product tangibly stored on a computer-readable medium and comprising computer-executable instructions which, when executed, cause an apparatus to perform the method of any one of the first aspect to the sixth aspect.BRIEF DESCRIPTION OF THE DRAWINGS
[0109] Some example embodiments will now be described with reference to the accompanying drawings, in which:
[0110] FIG. 1A illustrates an example environment in which some embodiments of the present disclosure can be implemented;
[0111] FIG. 1B illustrates an example communication system in which some embodiments of the present disclosure can be implemented;
[0112] FIG. 1C illustrates an example component structure which some embodiments of the present disclosure can applied;
[0113] FIG. 1D illustrates an example module structure which some embodiments of the present disclosure can applied;
[0114] FIG. 2A illustrates an example process according to some embodiments of the present disclosure;
[0115] FIG. 2B illustrates another example process according to some other embodiments of the present disclosure;
[0116] FIG. 3 illustrates an example training process according to some embodiments of the present disclosure;
[0117] FIG. 4A illustrates an example process of an iteration of model evolution according to some embodiments of the present disclosure;
[0118] FIG. 4B illustrates an example system architecture according to some embodiments of the present disclosure;
[0119] FIG. 5 illustrates an example procedure for an iteration of model evolution according to some other embodiments of the present disclosure;
[0120] FIG. 6 illustrates an example procedure of request for influencing traffic routing according to some embodiments of the present disclosure;
[0121] FIG. 7 illustrates a flowchart of an example method implemented at a first apparatus according to some embodiments of the present disclosure;
[0122] FIG. 8 illustrates a flowchart of an example method implemented at a second apparatus according to some embodiments of the present disclosure;
[0123] FIG. 9 illustrates a flowchart of an example method implemented at a third apparatus according to some embodiments of the present disclosure;
[0124] FIG. 10 illustrates a flowchart of an example method implemented at a third apparatus according to some other embodiments of the present disclosure;
[0125] FIG. 11 illustrates a flowchart of an example method implemented at a fourth apparatus according to some embodiments of the present disclosure;
[0126] FIG. 12 illustrates a flowchart of an example method implemented at a fifth apparatus according to some embodiments of the present disclosure;
[0127] FIG. 13 illustrates a flowchart of an example method implemented at a system according to some embodiments of the present disclosure; and
[0128] FIG. 14 illustrates a simplified block diagram of a device that is suitable for implementing example embodiments of the present disclosure.
[0129] Throughout the drawings, the same or similar reference numerals represent the same or similar elements.DETAILED DESCRIPTION
[0130] Principles of the present disclosure will now be described with reference to some example embodiments. It is to be understood that these embodiments are described for the purpose of illustration and help those skilled in the art to understand and implement the present disclosure, without suggesting any limitation as to the scope of the disclosure. Embodiments of the present disclosure described herein can be implemented in various manners other than the ones described below.
[0131] In the following description and claims, unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skills in the art to which this disclosure belongs.
[0132] References in the present disclosure to “one embodiment” , “an embodiment” , “an example embodiment” , and the like indicate that the embodiment described may include a particular feature, structure, or characteristic, but it is not necessary that every embodiment includes the particular feature, structure, or characteristic. The term ‘another embodiment’ is to be read as ‘at least one other embodiment. ’ Moreover, such phrases are not necessarily referring to the same embodiment. Further, when a particular feature, structure, or characteristic is described in connection with an embodiment, it is submitted that it is within the knowledge of one skilled in the art to affect such feature, structure, or characteristic in connection with other embodiments whether or not explicitly described.
[0133] It shall be understood that although the terms “first” and “second” etc. may be used herein to describe various elements, these elements should not be limited by these terms. These terms are used to distinguish one element from another. For example, a first element could be termed a second element, and similarly, a second element could be termed a first element, without departing from the scope of example embodiments. As used herein, the term “and / or” includes any and all combinations of one or more of the listed terms. Other definitions, explicit and implicit, may be included below.
[0134] The terminology used herein is for the purpose of describing particular embodiments and is not intended to be limiting of example embodiments. As used herein, the singular forms “a” , “an” and “the” are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms “comprises” , “comprising” , “has” , “having” , “includes” and / or “including” , when used herein, specify the presence of stated features, elements, and / or components etc., but do not preclude the presence or addition of one or more other features, elements, components and / or combinations thereof.
[0135] In the following description and claims, unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skills in the art to which this disclosure belongs.
[0136] As used herein, the term ‘terminal apparatus’ refers to a terminal device or a module / chip in the terminal device above. The terminal device may refer to any device having wireless or wired communication capabilities. Examples of the terminal device include, but not limited to, user equipment (UE) , personal computers, desktops, mobile phones, cellular phones, smart phones, personal digital assistants (PDAs) , portable computers, tablets, wearable devices, internet of things (IoT) devices, Ultra-reliable and Low Latency Communications (URLLC) devices, Internet of Everything (IoE) devices, machine type communication (MTC) devices, device on vehicle for V2X communication where X means pedestrian, vehicle, or infrastructure / network, devices for Integrated Access and Backhaul (IAB) , Small Data Transmission (SDT) , mobility, Multicast and Broadcast Services (MBS) , positioning, dynamic / flexible duplex in commercial networks, reduced capability (RedCap) , Space borne vehicles or Air borne vehicles in Non-terrestrial networks (NTN) including Satellites and High Altitude Platforms (HAPs) encompassing Unmanned Aircraft Systems (UAS) , eXtended Reality (XR) devices including different types of realities such as Augmented Reality (AR) , Mixed Reality (MR) and Virtual Reality (VR) , the unmanned aerial vehicle (UAV) commonly known as a drone which is an aircraft without any human pilot, devices on high speed train (HST) , or image capture devices such as digital cameras, sensors, gaming devices, music storage and playback appliances, or Internet appliances enabling wireless or wired Internet access and browsing and the like. The ‘terminal device’ can further has ‘multicast / broadcast’ feature, to support public safety and mission critical, V2X applications, transparent IPv4 / IPv6 multicast delivery, IPTV, smart TV, radio services, software delivery over wireless, group communications and IoT applications. It may be also incorporated one or multiple Subscriber Identity Module (SIM) as known as Multi-SIM. The term “terminal device” can be used interchangeably with a UE, a mobile station, a subscriber station, a mobile terminal, a user terminal, a wireless device or a reduced capability terminal device.
[0137] As used herein, the apparatus other than the terminal apparatus, such as the server, the network function, the apparatus as the part of a data plane / control plane of a core network, or the radio access network (RAN) node, etc. may be referred to as network apparatus. The network apparatus may be a network device or a module / chip of the network device above. The term “network device” refers to a device which is capable of providing or hosting a cell or coverage where terminal devices can communicate. Examples of a network device include, but not limited to, a Node B (NodeB or NB) , an evolved NodeB (eNodeB or eNB) , a next generation NodeB (gNB) , a transmission reception point (TRP) , a remote radio unit (RRU) , a radio head (RH) , a remote radio head (RRH) , an IAB node, a low power node such as a femto node, a pico node, a reconfigurable intelligent surface (RIS) , Network-controlled Repeaters, and the like. In some other embodiments, the term “network device” may refer to a device at core network side, for example, the network device may be a core network side entity / element, e.g. a network function in a control plane or a network function in a data plane. In other embodiments, the term “network device” may refer to a device in a data network, for example, the network device may be a data network side entity / element, e.g. a network server, an application server.
[0138] The terminal device or the network device may have Artificial intelligence (AI) or Machine learning capability. It generally includes a model which has been trained from numerous collected data for a specific function, and can be used to predict some information. As an example, the terminal or the network device may work on several frequency ranges, e.g. FR1 (410 MHz –7125 MHz) , FR2 (24.25 GHz to 71 GHz) , 71 GHz to 114 GHz, and frequency band larger than 100 GHz as well as Tera Hertz (THz) . It can further work on licensed / unlicensed / shared spectrum. The terminal device may have more than one connections with the network devices under Multi-Radio Dual Connectivity (MR-DC) application scenario. The terminal device or the network device can work on full duplex, flexible duplex and cross division duplex modes.
[0139] The network device may have the function of network energy saving, Self-Organizing Networks (SON) / Minimization of Drive Tests (MDT) . The terminal may have the function of power saving.
[0140] The embodiments of the present disclosure may be performed in test equipment, e.g. signal generator, signal analyzer, spectrum analyzer, network analyzer, test terminal device, test network device, channel emulator.
[0141] The embodiments of the present disclosure may be performed according to any generation communication protocols either currently known or to be developed in the future. Examples of the communication protocols include, but not limited to the fourth generation (4G) , 4.5G, the fifth generation (5G) communication protocols, 5.5G, 5G-Advanced networks, Wireless Fidelity (WiFi) network, Ultra Wideband (UWB) network, or the sixth generation (6G) networks.
[0142] In some examples, values, procedures, or apparatus are referred to as ‘best, ’ ‘lowest, ’ ‘highest, ’ ‘minimum, ’ ‘maximum, ’ or the like. It will be appreciated that such descriptions are intended to indicate that a selection among many used functional alternatives can be made, and such selections need not be better, smaller, higher, or otherwise preferable to other selections.
[0143] The term “circuitry” used herein may refer to hardware circuits and / or combinations of hardware circuits and software. For example, the circuitry may be a combination of analog and / or digital hardware circuits with software / firmware. As a further example, the circuitry may be any portions of hardware processors with software including digital signal processor (s) , software, and memory (ies) that work together to cause an apparatus, such as a terminal device or a network device, to perform various functions. In a still further example, the circuitry may be hardware circuits and or processors, such as a microprocessor or a portion of a microprocessor, that requires software / firmware for operation, but the software may not be present when it is not needed for operation. As used herein, the term circuitry also covers an implementation of merely a hardware circuit or processor (s) or a portion of a hardware circuit or processor (s) and its (or their) accompanying software and / or firmware.
[0144] Some embodiments of the present disclosure relate to training of one or more models (such as AI models, also known as artificial intelligence (AI) or machine learning (ML) , AI / ML, models) in a network. In an example application, each of a group of entities, such as UE, server, network function, RAN node, etc., has a local model and further has (or hosts) a private dataset. The local models may be heterogeneous, i.e. different in their structures. In the application, each of these models is to be trained using the private datasets, without requiring the private datasets to leave their hosting entities. Data in the private datasets may be highly heterogeneous (biased) , and the heterogeneity negatively impacts the training performance. It is shown that use of additional training data (e.g. data from a public dataset that is not privacy-sensitive) can improve the training performance. There is thus a need for systems and methods of using the additional training data during the training process (i.e. during the training of the local models) .
[0145] The solution of the present disclosure is different from some existing solutions (e.g a solution using federated learning. In the solution for AI models training in which federated learning is used, a number of entities (e.g. UEs, servers, network functions, RAN nodes, etc., or a combination thereof) collectively train a common AI model using their private datasets in multiple iterations. In an iteration, each of the entities sends a local version of the model to an aggregator node. The local version of the model sent from an entity to the aggregator node is obtained by the entity through a local training, i.e. by training the model using the entity’s private dataset based on, for example, a random initialization of the model or a global version of the mode. The global version of the model is received from an aggregator node in a previous iteration. The aggregator node receives local versions of the model from the entities and aggregates the local versions of the model, e.g. using FederatedAveraging / FedAverage, to update a global version of the model. FederatedAveraging / FedAverage is a communication efficient algorithm for the distributed training with an enormous number of clients. The aggregator node then sends the updated global version of the model (e.g. including values or gradients of model parameters) to the entities, each of which performs the local training accordingly for the next iteration. When the goal version of the model converges, the entities are notified by the aggregator node, and the learning procedure finishes.
[0146] Federated learning aims to train a common model for all the entities, and it does not support personalized local models as the solution of some embodiments of the present disclosure does. Federated learning requires all the type-1 PSFs (processing service functions) or majority / many of them (i.e. selection space is small) to participate in each iteration of the learning (corresponding to model evolution in some embodiments of the present disclosure) , and there is not much room for the learning to adapt to network dynamics such as resource availability and performance variation. Whereas, the solution of some embodiments of the present disclosure involve one type-1 PSF in each iteration of model evolution, which is dynamically selected (selection space is large) according to the network dynamics. Furthermore, Federated learning suffers from biased private datasets, and it does not provide a solution to mitigating associated negative impact as the solution of some embodiments of the present disclosure does. In some embodiments, a type-1 PSF and a type-2 PSF are different processing service functions. The type-1 PSF has its own local model to be trained. In some embodiments, during model training of the type-1 PSFs, the network not only acts as a channel, but also assists in training the models, for example, some nodes of the network may provide additional training data, and such nodes may be referred to as type-2 PSFs. The type-2 PSFs does not have (or is not associated with) its own local model to be trained, but may help train the local models of the type-1 PSFs.
[0147] For illustrative purposes, principles and example embodiments of the present disclosure will be described below with reference to FIGs. 1A-13. However, it is to be noted that these embodiments are given to enable the skilled in the art to understand inventive concepts of some embodiments of the present disclosure and implement the solution as proposed herein, and not intended to limit scope of the present disclosure in any way.
[0148] FIG. 1A illustrates an example environment in which some embodiments of the present disclosure can be implemented. Referring to FIG. 1A, as an illustrative example without limitation, a simplified schematic illustration of a communication system is provided. The communication system 100 comprises a radio access network 120. The radio access network 120 may be a next generation (e.g. sixth generation (6G) or later) radio access network, or a legacy (e.g. 5G, 4G, 3G or 2G) radio access network. One or more communication electric device (ED) 110a-120j (generically referred to as 110) may be interconnected to one another or connected to one or more network nodes (170a, 170b, generically referred to as 170) in the radio access network 120. A core network 130 may be a part of the communication system and may be dependent or independent of the radio access technology used in the communication system 100. Also the communication system 100 comprises a public switched telephone network (PSTN) 140, the internet 150, and other networks 160. In some embodiments, the one or more communication electric device (ED) 110a-120j or network function of the core network 130 may be as type-1 PSF, such as type-1 PSF 510 in FIG. 5, or type-1 PSF A 610, etc.. In some embodiments. In some embodiments, part of the data plane of the core network 130 may be as type-2 PSF, such as a first type-2 PSF 540, a second type-2 PSF 550, or a type-2 PSF B 640, etc..
[0149] FIG. 1B illustrates an example communication system 100-1 in which some embodiments of the present disclosure can be implemented. FIG. 1B illustrates an example communication system 100-1. In general, the communication system 100-1 enables multiple wireless or wired elements to communicate data and other content. The purpose of the communication system 100-1 may be to provide content, such as voice, data, video, and / or text, via broadcast, multicast and unicast, etc. The communication system 100-1 may operate by sharing resources, such as carrier spectrum bandwidth, between its constituent elements. The communication system 100-1 may include a terrestrial communication system and / or a non-terrestrial communication system. The communication system 100-1 may provide a wide range of communication services and applications (such as earth monitoring, remote sensing, passive sensing and positioning, navigation and tracking, autonomous delivery and mobility, etc. ) . The communication system 100-1 may provide a high degree of availability and robustness through a joint operation of the terrestrial communication system and the non-terrestrial communication system. For example, integrating a non-terrestrial communication system (or components thereof) into a terrestrial communication system can result in what may be considered a heterogeneous network comprising multiple layers. Compared to conventional communication networks, the heterogeneous network may achieve better overall performance through efficient multi-link joint operation, more flexible functionality sharing, and faster physical layer link switching between terrestrial networks and non-terrestrial networks.
[0150] The terrestrial communication system and the non-terrestrial communication system could be considered sub-systems of the communication system. In the example shown, the communication system 100-1 includes electronic devices (ED) 110a-110d (generically referred to as ED 110) , radio access networks (RANs) 120a-120b, non-terrestrial communication network 120c, a core network 130, a public switched telephone network (PSTN) 140, the internet 150, and other networks 160. The RANs 120a-120b include respective base stations (BSs) 170a-170b, which may be generically referred to as terrestrial transmit and receive points (T-TRPs) 170a-170b. The non-terrestrial communication network 120c includes an access node 120c, which may be generically referred to as a non-terrestrial transmit and receive point (NT-TRP) 172.
[0151] Any ED 110 may be alternatively or additionally configured to interface, access, or communicate with any other T-TRP 170a-170b and NT-TRP 172, the internet 150, the core network 130, the PSTN 140, the other networks 160, or any combination of the preceding. In some examples, ED 110a may communicate an uplink and / or downlink transmission over an interface 190a with T-TRP 170a. In some examples, the EDs 110a, 110b and 110d may also communicate directly with one another via one or more sidelink air interfaces 190b. In some examples, ED 110d may communicate an uplink and / or downlink transmission over an interface 190c with NT-TRP 172.
[0152] The air interfaces 190a and 190b may use similar communication technology, such as any suitable radio access technology. For example, the communication system 100-1 may implement one or more channel access methods, such as code division multiple access (CDMA) , time division multiple access (TDMA) , frequency division multiple access (FDMA) , orthogonal FDMA (OFDMA) , or single-carrier FDMA (SC-FDMA) in the air interfaces 190a and 190b. The air interfaces 190a and 190b may utilize other higher dimension signal spaces, which may involve a combination of orthogonal and / or non-orthogonal dimensions.
[0153] The air interface 190c can enable communication between the ED 110d and one or multiple NT-TRPs 172 via a wireless link or simply a link. For some examples, the link is a dedicated connection for unicast transmission, a connection for broadcast transmission, or a connection between a group of EDs and one or multiple NT-TRPs for multicast transmission.
[0154] The RANs 120a and 120b are in communication with the core network 130 to provide the EDs 110a 110b, and 110c with various services such as voice, data, and other services. The RANs 120a and 120b and / or the core network 130 may be in direct or indirect communication with one or more other RANs (not shown) , which may or may not be directly served by core network 130, and may or may not employ the same radio access technology as RAN 120a, RAN 120b or both. The core network 130 may also serve as a gateway access between (i) the RANs 120a and 120b or EDs 110a 110b, and 110c or both, and (ii) other networks (such as the PSTN 140, the internet 150, and the other networks 160) . In addition, some or all of the EDs 110a 110b, and 110c may include functionality for communicating with different wireless networks over different wireless links using different wireless technologies and / or protocols. Instead of wireless communication (or in addition thereto) , the EDs 110a 110b, and 110c may communicate via wired communication channels to a service provider or switch (not shown) , and to the internet 150. PSTN 140 may include circuit switched telephone networks for providing plain old telephone service (POTS) . Internet 150 may include a network of computers and subnets (intranets) or both, and incorporate protocols, such as Internet Protocol (IP) , Transmission Control Protocol (TCP) , User Datagram Protocol (UDP) . EDs 110a 110b, and 110c may be multimode devices capable of operation according to multiple radio access technologies, and incorporate multiple transceivers necessary to support such.
[0155] FIG. 1C illustrates an example component structure 100-2 which some embodiments of the present disclosure can applied. FIG. 1C illustrates another example of an ED 110 and a base station 170a, 170b and / or 170c. The ED 110 is used to connect persons, objects, machines, etc. The ED 110 may be widely used in various scenarios, for example, cellular communications, device-to-device (D2D) , vehicle to everything (V2X) , peer-to-peer (P2P) , machine-to-machine (M2M) , machine-type communications (MTC) , internet of things (IOT) , virtual reality (VR) , augmented reality (AR) , industrial control, self-driving, remote medical, smart grid, smart furniture, smart office, smart wearable, smart transportation, smart city, drones, robots, remote sensing, passive sensing, positioning, navigation and tracking, autonomous delivery and mobility, etc.
[0156] Each ED 110 represents any suitable end user device for wireless operation and may include such devices (or may be referred to) as a user equipment / device (UE) , a wireless transmit / receive unit (WTRU) , a mobile station, a fixed or mobile subscriber unit, a cellular telephone, a station (STA) , a machine type communication (MTC) device, a personal digital assistant (PDA) , a smartphone, a laptop, a computer, a tablet, a wireless sensor, a consumer electronics device, a smart book, a vehicle, a car, a truck, a bus, a train, or an IoT device, an industrial device, or apparatus (e.g. communication module, modem, or chip) in the forgoing devices, among other possibilities. Future generation EDs 110 may be referred to using other terms. The base station 170a and 170b is a T-TRP and will hereafter be referred to as T-TRP 170. Also shown in FIG. 1C, a NT-TRP will hereafter be referred to as NT-TRP 172. Each ED 110 connected to T-TRP 170 and / or NT-TRP 172 can be dynamically or semi-statically turned-on (i.e., established, activated, or enabled) , turned-off (i.e., released, deactivated, or disabled) and / or configured in response to one of more of: connection availability and connection necessity.
[0157] The ED 110 includes a transmitter 201 and a receiver 203 coupled to one or more antennas 204. One antenna 204 is illustrated. One, some, or all of the antennas may alternatively be panels. The transmitter 201 and the receiver 203 may be integrated, e.g. as a transceiver. The transceiver is configured to modulate data or other content for transmission by at least one antenna 204 or network interface controller (NIC) . The transceiver is also configured to demodulate data or other content received by the at least one antenna 204. Each transceiver includes any suitable structure for generating signals for wireless or wired transmission and / or processing signals received wirelessly or by wire. Each antenna 204 includes any suitable structure for transmitting and / or receiving wireless or wired signals.
[0158] The ED 110 includes at least one memory 208. The memory 208 stores instructions and data used, generated, or collected by the ED 110. For example, the memory 208 could store software instructions or modules configured to implement some or all of the functionality and / or embodiments described herein and that are executed by the processing unit (s) 210. Each memory 208 includes any suitable volatile and / or non-volatile storage and retrieval device (s) . Any suitable type of memory may be used, such as random access memory (RAM) , read only memory (ROM) , hard disk, optical disc, subscriber identity module (SIM) card, memory stick, secure digital (SD) memory card, on-processor cache, and the like.
[0159] The ED 110 may further include one or more input / output devices (not shown) or interfaces (such as a wired interface to the internet 150 in FIG. 1A) . The input / output devices permit interaction with a user or other devices in the network. Each input / output device includes any suitable structure for providing information to or receiving information from a user, such as a speaker, microphone, keypad, keyboard, display, or touch screen, including network interface communications.
[0160] The ED 110 further includes a processor 210 for performing operations including those related to preparing a transmission for uplink transmission to the NT-TRP 172 and / or T-TRP 170, those related to processing downlink transmissions received from the NT-TRP 172 and / or T-TRP 170, and those related to processing sidelink transmission to and from another ED 110. Processing operations related to preparing a transmission for uplink transmission may include operations such as encoding, modulating, transmit beamforming, and generating symbols for transmission. Processing operations related to processing downlink transmissions may include operations such as receive beamforming, demodulating and decoding received symbols. Depending upon the embodiment, a downlink transmission may be received by the receiver 203, possibly using receive beamforming, and the processor 210 may extract signaling from the downlink transmission (e.g. by detecting and / or decoding the signaling) . An example of signaling may be a reference signal transmitted by NT-TRP 172 and / or T-TRP 170. In some embodiments, the processor 276 implements the transmit beamforming and / or receive beamforming based on the indication of beam direction, e.g. beam angle information (BAI) , received from T-TRP 170. In some embodiments, the processor 210 may perform operations relating to network access (e.g. initial access) and / or downlink synchronization, such as operations relating to detecting a synchronization sequence, decoding and obtaining the system information, etc. In some embodiments, the processor 210 may perform channel estimation, e.g. using a reference signal received from the NT-TRP 172 and / or T-TRP 170.
[0161] Although not illustrated, the processor 210 may form part of the transmitter 201 and / or receiver 203. Although not illustrated, the memory 208 may form part of the processor 210.
[0162] The processor 210, and the processing components of the transmitter 201 and receiver 203 may each be implemented by the same or different one or more processors that are configured to execute instructions stored in a memory (e.g. in memory 208) . Alternatively, some or all of the processor 210, and the processing components of the transmitter 201 and receiver 203 may be implemented using dedicated circuitry, such as a programmed field-programmable gate array (FPGA) , a graphical processing unit (GPU) , or an application-specific integrated circuit (ASIC) .
[0163] The T-TRP 170 may be known by other names in some implementations, such as a base station, a base transceiver station (BTS) , a radio base station, a network node, a network device, a device on the network side, a transmit / receive node, a Node B, an evolved NodeB (eNodeB or eNB) , a Home eNodeB, a next Generation NodeB (gNB) , a transmission point (TP) ) , a site controller, an access point (AP) , or a wireless router, a relay station, a remote radio head, a terrestrial node, a terrestrial network device, or a terrestrial base station, base band unit (BBU) , remote radio unit (RRU) , active antenna unit (AAU) , remote radio head (RRH) , central unit (CU) , distribute unit (DU) , positioning node, among other possibilities. The T-TRP 170 may be macro BSs, pico BSs, relay node, donor node, or the like, or combinations thereof. The T-TRP 170 may refer to the forging devices or apparatus (e.g. communication module, modem, or chip) in the forgoing devices.
[0164] In some embodiments, the parts of the T-TRP 170 may be distributed. For example, some of the modules of the T-TRP 170 may be located remote from the equipment housing the antennas of the T-TRP 170, and may be coupled to the equipment housing the antennas over a communication link (not shown) sometimes known as front haul, such as common public radio interface (CPRI) . Therefore, in some embodiments, the term T-TRP 170 may also refer to modules on the network side that perform processing operations, such as determining the location of the ED 110, resource allocation (scheduling) , message generation, and encoding / decoding, and that are not necessarily part of the equipment housing the antennas of the T-TRP 170. The modules may also be coupled to other T-TRPs. In some embodiments, the T-TRP 170 may actually be a plurality of T-TRPs that are operating together to serve the ED 110, e.g. through coordinated multipoint transmissions.
[0165] The T-TRP 170 includes at least one transmitter 252 and at least one receiver 254 coupled to one or more antennas 256. One antenna 256 is illustrated. One, some, or all of the antennas may alternatively be panels. The transmitter 252 and the receiver 254 may be integrated as a transceiver. The T-TRP 170 further includes a processor 260 for performing operations including those related to: preparing a transmission for downlink transmission to the ED 110, processing an uplink transmission received from the ED 110, preparing a transmission for backhaul transmission to NT-TRP 172, and processing a transmission received over backhaul from the NT-TRP 172. Processing operations related to preparing a transmission for downlink or backhaul transmission may include operations such as encoding, modulating, precoding (e.g. MIMO precoding) , transmit beamforming, and generating symbols for transmission. Processing operations related to processing received transmissions in the uplink or over backhaul may include operations such as receive beamforming, and demodulating and decoding received symbols. The processor 260 may also perform operations relating to network access (e.g. initial access) and / or downlink synchronization, such as generating the content of synchronization signal blocks (SSBs) , generating the system information, etc. In some embodiments, the processor 260 also generates the indication of beam direction, e.g. BAI, which may be scheduled for transmission by scheduler 253. The processor 260 performs other network-side processing operations described herein, such as determining the location of the ED 110, determining where to deploy NT-TRP 172, etc. In some embodiments, the processor 260 may generate signaling, e.g. to configure one or more parameters of the ED 110 and / or one or more parameters of the NT-TRP 172. Any signaling generated by the processor 260 is sent by the transmitter 252. Note that “signaling” , as used herein, may alternatively be called control signaling. Dynamic signaling may be transmitted in a control channel, e.g. a physical downlink control channel (PDCCH) , and static or semi-static higher layer signaling may be included in a packet transmitted in a data channel, e.g. in a physical downlink shared channel (PDSCH) .
[0166] A scheduler 253 may be coupled to the processor 260. The scheduler 253 may be included within or operated separately from the T-TRP 170, which may schedule uplink, downlink, and / or backhaul transmissions, including issuing scheduling grants and / or configuring scheduling-free ( “configured grant” ) resources. The T-TRP 170 further includes a memory 258 for storing information and data. The memory 258 stores instructions and data used, generated, or collected by the T-TRP 170. For example, the memory 258 could store software instructions or modules configured to implement some or all of the functionality and / or embodiments described herein and that are executed by the processor 260.
[0167] Although not illustrated, the processor 260 may form part of the transmitter 252 and / or receiver 254. Also, although not illustrated, the processor 260 may implement the scheduler 253. Although not illustrated, the memory 258 may form part of the processor 260.
[0168] The processor 260, the scheduler 253, and the processing components of the transmitter 252 and receiver 254 may each be implemented by the same or different one or more processors that are configured to execute instructions stored in a memory, e.g. in memory 258. Alternatively, some or all of the processor 260, the scheduler 253, and the processing components of the transmitter 252 and receiver 254 may be implemented using dedicated circuitry, such as a FPGA, a GPU, or an ASIC.
[0169] Although the NT-TRP 172 is illustrated as a drone as an example, the NT-TRP 172 may be implemented in any suitable non-terrestrial form. Also, the NT-TRP 172 may be known by other names in some implementations, such as a non-terrestrial node, a non-terrestrial network device, or a non-terrestrial base station. The NT-TRP 172 includes a transmitter 272 and a receiver 274 coupled to one or more antennas 280. One antenna 280 is illustrated. One, some, or all of the antennas may alternatively be panels. The transmitter 272 and the receiver 274 may be integrated as a transceiver. The NT-TRP 172 further includes a processor 276 for performing operations including those related to: preparing a transmission for downlink transmission to the ED 110, processing an uplink transmission received from the ED 110, preparing a transmission for backhaul transmission to T-TRP 170, and processing a transmission received over backhaul from the T-TRP 170. Processing operations related to preparing a transmission for downlink or backhaul transmission may include operations such as encoding, modulating, precoding (e.g. MIMO precoding) , transmit beamforming, and generating symbols for transmission. Processing operations related to processing received transmissions in the uplink or over backhaul may include operations such as receive beamforming, and demodulating and decoding received symbols. In some embodiments, the processor 276 implements the transmit beamforming and / or receive beamforming based on beam direction information (e.g. BAI) received from T-TRP 170. In some embodiments, the processor 276 may generate signaling, e.g. to configure one or more parameters of the ED 110. In some embodiments, the NT-TRP 172 implements physical layer processing, but does not implement higher layer functions such as functions at the medium access control (MAC) or radio link control (RLC) layer. As this is an example, more generally, the NT-TRP 172 may implement higher layer functions in addition to physical layer processing.
[0170] The NT-TRP 172 further includes a memory 278 for storing information and data. Although not illustrated, the processor 276 may form part of the transmitter 272 and / or receiver 274. Although not illustrated, the memory 278 may form part of the processor 276.
[0171] The processor 276 and the processing components of the transmitter 272 and receiver 274 may each be implemented by the same or different one or more processors that are configured to execute instructions stored in a memory, e.g. in memory 278. Alternatively, some or all of the processor 276 and the processing components of the transmitter 272 and receiver 274 may be implemented using dedicated circuitry, such as a programmed FPGA, a GPU, or an ASIC. In some embodiments, the NT-TRP 172 may actually be a plurality of NT-TRPs that are operating together to serve the ED 110, e.g. through coordinated multipoint transmissions.
[0172] The T-TRP 170, the NT-TRP 172, and / or the ED 110 may include other components, but these have been omitted for the sake of clarity.
[0173] FIG. 1D illustrates an example module structure which some embodiments of the present disclosure can applied. One or more operations (or steps) of the embodiment methods provided herein may be performed by corresponding units or modules, according to FIG. 1D. FIG. 1D illustrates units or modules in a device, such as in ED 110, in T-TRP 170, or in NT-TRP 172. For example, a signal may be transmitted by a transmitting unit or a transmitting module. For example, a signal may be transmitted by a transmitting unit or a transmitting module. A signal may be received by a receiving unit or a receiving module. A signal may be processed by a processing unit or a processing module. Other steps (or operations) may be performed by an artificial intelligence (AI) or machine learning (ML) module. The respective units or modules may be implemented using hardware, one or more components or devices that execute software, or a combination thereof. For instance, one or more of the units or modules may be an integrated circuit, such as a programmed FPGA, a GPU, or an ASIC. It will be appreciated that where the modules are implemented using software for execution by a processor for example, they may be retrieved by a processor, in whole or part as needed, individually or together for processing, in single or multiple instances, and that the modules themselves may include instructions for further deployment and instantiation.
[0174] Additional details regarding the EDs 110, T-TRP 170, and NT-TRP 172 are known to those of skill in the art. As such, these details are omitted here.
[0175] FIG. 2A illustrates an example process 200-1 according to some embodiments of the present disclosure. As shown in FIG. 2A, a first apparatus 221, a second apparatus 220 and a third apparatus 230 are involved in the process 200-1. In some examples, the first apparatus 221 may be a terminal apparatus, or a sever, or a network function. The server is a computer or system that provides resources, data, services, or programs to other computers, known as clients, over a network. The network function (NF) is a functional building block within a network infrastructure, which has well-defined external interfaces and a well-defined functional behavior, for example, the network function may be a network node or a physical appliance. The network function may also be logical network entity that can be deployed or located at a network location or on a physical network node. In some embodiments, the network function is a virtual network function which can be instantiated in a cloud computing environment (e.g. a data center) . In some embodiments, an example of the first apparatus 221 may refer to a type-1 PSF 510 in FIG. 5 or a type-1 PSF A 610 in FIG. 6. The terminal apparatus may be a terminal device or a module / chip of a terminal device. A PSF is a processing service function. In some examples, the second apparatus 220 is a part of a data plane of a core network, or the second apparatus 220 is connected to a data plane of a core network. In some embodiments, the data plane is known as a user plane, e.g. the user plane of the 3GPP 5G system. In some examples, the data plane may be a collection of network functions which are used to process and / or transmit data traffic. A type-2 PSF may be an example of the second apparatus 220, e.g. the type-2 PSF 403 or the first type-2 PSF 540 or the type-2 PSF B 640 below. In some embodiments, the third apparatus 230 may be a part of a control plane of the core network. The control plane may be a collection of network functions that control how the system operates, including how the data traffic is processed and / or transmitted. In some embodiments, the third apparatus 230 may be a third apparatus 231 in FIG. 2B. In some embodiments, one or more other apparatuses (not shown in FIG. 2A) are involved in the process 200-1, such as a fourth apparatus, a fifth apparatus, and / or a sixth apparatus. The fourth apparatus, the fifth apparatus, and / or the sixth apparatus may be the same apparatus as a fourth apparatus 240, a fifth apparatus 250, and / or a sixth apparatus mentioned in a process 200-2 (with reference to FIG. 2B) , respectively. Some optional operations below are not shown in FIG. 2A, one or more of optional operations may further refer to FIG. 2B.
[0176] The first apparatus 221 is one of the first one or more apparatuses, and the second apparatus 220 is one of second one or more apparatuses. One (may be any one) of the second one or more apparatuses may be a part of a data plane of a core network or be connected to a data plane of a core network. The first apparatus 221 may be selected among the first one or more apparatuses for performing model updating of the one or more models associated with the first one or more apparatuses. One of the first one or more apparatuses may be determined, by the third apparatus 230, as the first apparatus 221 for performing the model updating. A model associated with one of the first one or more apparatuses (e.g. the first apparatus) herein may refer to the model is a local model of an apparatus (i.e. the one of the first one or more apparatuses) , for example, the model is stored or maintained by a device including this apparatus.
[0177] The second apparatus 220 may be selected among the second one or more apparatuses. The third apparatus 230 may transmit to the fourth apparatus, a fifth message (e.g. the fifth message 265 in FIG. 2B below) , and the fourth apparatus may receive the fifth message from the third apparatus 230. The fifth message is for requesting to influence traffic routing, in which influencing the traffic routing is for selecting the second apparatus 220 among the second one or more apparatuses and at least one candidate apparatus among the first one or more apparatuses. The second apparatus 220 and the at least one candidate apparatus may jointly perform the model updating of the one or more models associated with the first one or more apparatuses, and the fourth apparatus is configured to perform a policy control function. The at least one candidate apparatus comprise the first apparatus 221. For each iteration, an apparatus as the first apparatus 221 may be selected among the first one or more apparatuses, and the at least one candidate apparatus are a collection of apparatuses selected as the first apparatuses 221 for each iteration. In some examples, the fifth message may comprise the following information, such as information of data traffic identifying data traffic of the at least one application (in some embodiment, the at least one application may be associated with the second one or more apparatuses) , in which the data traffic is to be routed during the traffic routing; information of the first one or more apparatuses; information of at least one location of the at least one application (in some embodiments, the at least one location of the at least one application corresponds to the second one or more apparatuses) ; information of one or more requirements associated with the traffic routing; or information of traffic filtering associated with the data traffic; or any combination thereof. In some embodiments, the application may be an application program and / or an application server having its associated data traffic, and the application is associated with an apparatus (i.e. one of the second one or more apparatuses) may refer to the location of the application corresponds to this apparatus. In some examples, the number of the application may be one or more than one, that is there may be the at least one application.
[0178] As mentioned above, the at least one application may be associated with the second one or more apparatuses (e.g. type-2 PSFs) , specifically, a location of an application among the at least one application corresponds to a type-2 PSF (e.g. the type-2 PSF 403 or the first type-2 PSF 540 or the type-2 PSF B 640 below) . In addition, one of the at least one candidate apparatus, for example, the type-1 PSF (e.g. the type-1 PSF 510 in FIG. 5 or the type-1 PSF A 610 in FIG. 6) , may join the training of the model by accessing the application. In other words, the application may be associated with a type-2 PSF (e.g. the type-2 PSF corresponds to the location of the application) and may be accessed by a type-1 PSF (e.g. the type-1 PSF may send and / or receive the data traffic related to the application) . In the fifth message above, the data traffic identified by the information of the data traffic may include data traffic to be routed to (or targeting) a type-1 PSF (referred to as downlink traffic) and data traffic to be routed to (or targeting) a type-2 PSF (referred to as uplink traffic) . Based on the fifth message, the fourth apparatus may generate at least one policy on influencing the traffic routing. In some embodiments, the at least one policy may comprise the following information: information of data traffic to be routed in the traffic routing, i.e. the information of data traffic identifying data traffic of the at least one application associated with the second one or more apparatuses, as included in the fifth message; the information of the at least one location of at least one application, wherein one of the at least one application is associated with the second apparatus, for example, a type-2 PSF corresponds to an location of the application; the information of one or more requirements associated with the traffic routing; or information of the first one or more apparatuses; or any combination thereof. The information comprised in the at least one policy may be the same as which comprised in the fifth message. The details of the information comprised in the fifth message may further refer to the corresponding descriptions of the request transported from the PSC 560 to the PCF 630, in the operation 603, in FIG. 6. The at least one policy may also referred to as one or multiple policies in the operation 605 below.
[0179] Further, in some examples, the fourth apparatus may transmit the at least one policy to a fifth apparatus (e.g. the fifth apparatus 250 in FIG. 2B below) . The fifth apparatus is configured to be a network controller for determining, based on the at least one policy, the second apparatus 220 and the at least one candidate apparatus. The at least one candidate apparatus comprise the first apparatus 221. In some examples, the at least one candidate apparatus comprise more than one candidate apparatuses, and one of the at least one candidate apparatus will be selected as the first apparatus 221. In some other examples, the at least one candidate apparatus comprise only one candidate apparatus, and this candidate apparatus will be as the first apparatus 221, thus, the fifth apparatus is configured to be a network controller for determining the second apparatus 220 and the first apparatus 221 (i.e. the only one candidate apparatus) . On the fifth apparatus side, after determining the second apparatus 220 and the at least one candidate apparatus, the fifth apparatus may transmit and the third apparatus 230 may receive a sixth message (may further refer to the sixth message 285 in FIG. 2B) for indicating the second apparatus 220 and the at least one candidate apparatus. After receiving the sixth message, the third apparatus 230 may transmit a feedback for the sixth message to the fifth apparatus, and accordingly, the fifth apparatus may receive the feedback from the third apparatus 230. The feedback for the sixth message may indicate a positive acknowledgement (ACK) or a negative ACK to the selection for the application location associated with the second apparatus 220. Specifically, the positive ACK may be for confirmation of the selection, and the negative ACK may be for rejection of the selection. In some examples, the selection indicates that multiple second apparatuses (including the second apparatus 220) are selected, then the third apparatus 230 may, based on the selection, determine the second apparatus 220 by selecting an application location associated with the second apparatus 220 among at least one application location. In some examples, the information of the at least one application may be included in the at least one policy, and the fifth apparatus may receive this information. In some examples, the third apparatus 230 may further determine a candidate apparatus as the first apparatus 221 among the at least one candidate apparatus, and information (e.g. an ID or a network address) identifying the first apparatus 221 may be included in the positive ACK for the sixth message. In some other examples, if there is only one candidate apparatus (i.e. the first apparatus 221) , then the sixth message above is for indicating the second apparatus 220 and the first apparatus 221. In such examples, the first apparatus 221 is indicated in the sixth message, and the third apparatus 230 doesn’t determine the first apparatus 221.
[0180] In some other examples, the fifth apparatus may be integrated in the fourth apparatus, then the fourth apparatus has the function of the fifth apparatus, that is, the fourth apparatus may determine, based on the at least one policy, the second apparatus 220 and the at least one candidate apparatus. In such examples, the transmission of the at least one policy is inner implementation of the fourth apparatus. Similar to the examples above, the at least one candidate apparatus comprise the first apparatus 221. In some examples, the at least one candidate apparatus comprise more than candidate apparatuses, and one of the at least one candidate apparatus will be selected as the first apparatus 221. In some other examples, the at least one candidate apparatus comprise only one candidate apparatus, and this candidate apparatus will be as the first apparatus 221, thus, the fourth apparatus may determine the second apparatus 220 and the first apparatus 221 (i.e. the only one candidate apparatus) . The sixth message may be transmitted from the fourth apparatus to the third apparatus 230, that is, the fourth apparatus transmits to the third apparatus 230, the sixth message for indicating the second apparatus 221 and the at least one candidate apparatus. After receiving the sixth message, the third apparatus 230 may transmit a feedback for the sixth message to the fourth apparatus, and the fourth apparatus may receive the feedback from the third apparatus 230. The feedback for the sixth message may indicates a positive acknowledgement (ACK) for confirmation of the selection or a negative ACK for rejection of the selection. If the feedback for the sixth message is the positive ACK, the positive ACK for the sixth message further comprises information identifying a candidate apparatus as the first apparatus 221 among the at least one candidate apparatus. In some other examples, if there is only one candidate apparatus (i.e. the first apparatus 221) , then the sixth message above is for indicating the second apparatus 220 and the first apparatus 221. In such examples, the first apparatus 221 is indicated in the sixth message, and the fourth apparatus doesn’t determine the first apparatus 221.
[0181] In some examples, the fifth apparatus and the fourth apparatus are integrated in a single network entity. In some embodiments, the sixth message from the fourth apparatus or the fifth apparatus may indicate a data plane management event associated with a selection of an application location associated with the second apparatus 220 among at least one application location. The data plane management event is for indicating a selection of an application location. The fourth apparatus and the fifth apparatus are not shown in FIG. 2A. In some examples, the fourth apparatus may be a fourth apparatus 240, and the fifth apparatus may be a fifth apparatus 250 in FIG. 2B. In some examples, the fifth apparatus may be a NWC, e.g. the NWC 520 in FIG. 5 or FIG. 6. In some examples, the fourth apparatus may be a PCF, e.g. the PCF 630 in FIG. 6. In some examples, the fifth apparatus (e.g. a NWC) may be integrated with the fourth apparatus (e.g. PCF) .
[0182] The first apparatus 221 may receive a second message for notifying that the first apparatus 221 is selected for performing the model updating. In some examples, the second message may be from the fourth apparatus or the fifth apparatus. In some embodiments, the fifth apparatus is a network controller. The fifth apparatus may transmit to the first apparatus 221, and the first apparatus 221 may receive from the fifth apparatus, the second message. In some examples, after receiving the second message, the first apparatus 221 may transmit a positive acknowledgement (ACK) indicating confirmation of the selection to the fifth apparatus, and on the fifth apparatus side, the fifth apparatus may receive the positive ACK for the confirmation of performing the model updating, in other words, the positive ACK is for confirming updating the model. Alternatively, in some embodiments, the fourth apparatus is configured to perform a policy control function, and the fourth apparatus and the fifth apparatus are integrated as one entity. The fourth apparatus may transmit to the first apparatus 221, and the first apparatus 221 may receive from the fourth apparatus, the second message. In some examples, after receiving the second message, the first apparatus 221 may transmit a positive ACK indicating confirmation of the selection to the fourth apparatus, and on the fourth apparatus side, the fourth apparatus may receive the positive ACK for the confirmation of performing the model updating.
[0183] The first apparatus 221 may receive an indication for instructing the first apparatus 221 to perform model updating of one or more models associated with first one or more apparatuses. In some embodiments, the received indication may be an indication 225 from the second apparatus 220, for example, the second apparatus 220 may transmit (201a) , to the first apparatus 221 among the first one or more apparatuses, the indication 225 for indicating the first apparatus 221 to perform the model updating of the one or more models. The second apparatus 220 may obtain the indication 225 from the third apparatus 230, for example, the third apparatus 230 may transmit and the second apparatus 220 may receive, a message including the indication 225. In some examples, the message including the indication 225 may be the same message as a fourth message 245 below. On the first apparatus 221 side, the first apparatus 221 receives (203a) the indication 225. In some examples, the indication 225 may be carried on a first message from the second apparatus 220, and the first message may further comprise values of model parameters of the one or more models. In some other examples, the indication 225 may be carried on a first message from the second apparatus 220, and the first apparatus 221 may obtain values of model parameters of the one or more models. In some examples, the values of model parameters may be pre-configured at the first apparatus 221. In some other examples, the values of model parameters may be dynamically provided by the second apparatus 220, for example, the values of model parameters may be provided together with the indication 225, alternatively, the the values of model parameters may be included in the indication 225. In some examples, the indication 225 may be transmitted from the second apparatus 220 and via a data plane function to the first apparatus 221.
[0184] In some other embodiments, the received indication may be an indication 235 from the third apparatus 230, for example, the third apparatus 230 may transmit (205) , to the first apparatus 221 among the first one or more apparatuses, the indication 235 for indicating the first apparatus 221 to perform the model updating, in other words, the third apparatus may transmit the indication 235 to the first apparatus 221 not via the second apparatus 221. On the first apparatus 221 side, the first apparatus 221 receives (207) the indication 235. Additionally, in some examples, the first apparatus 221 may further receive, from the second apparatus 220, the values of model parameters of the one or more models. In some examples, the indication 235 may be transmitted from the third apparatus 230 to the first apparatus 221, for example, the indication 235 is transmitted not via a control plane function. In some other examples, the indication 235 may be transmitted from the third apparatus 230 and via a control plane function to the first apparatus 221. In some embodiments, the control plane function may comprise a network exposure function (NEF) , a policy control function (PCF) , a network storage function (NSF) , a network controller (NWC) , a path management function (PMF) or an access and mobility management function (AMF) , or the combination of one or more functions above.
[0185] In some embodiments, the received indication, for example, the indication 225 or the indication 235, may comprise a list of model identifying information for identifying the one or more models; version information of the one or more models; one or more status indications; or any combination thereof. Each of the one or more models has corresponding version information. Each of the one or more status indications indicates a status of a model among the one or more models, and the status of the model comprises frozen or unfrozen. In some embodiments, a frozen model is a converged model (in other words, the model has converged) , additionally or alternatively, an unfrozen model is an unconverged model (in other words, the model has not converged) . During the model updating of the one or more models, the status of a model in the one or more models may change, e.g. change from the frozen status to unfrozen status, or from the unfrozen status to the forgone status. For example, when or after the first apparatus 221 performs the model updating, the status of the model associated with the first apparatus 221 may change.
[0186] On the second apparatus 220 side, in some embodiments, the model updating may also be performed at the second apparatus 220. For example, the second apparatus 220 does not have (or is not associated with) a local model to train, but the second apparatus 220 may help train local models of the first one or more apparatuses. In some examples, the second apparatus 220 may perform the model updating using additional training data. The additional training data may be different from training data for performing the model updating at the first one or more apparatuses.
[0187] In some embodiments, the third apparatus 230 may transmit (209) the fourth message 245 to the second apparatus 220, and accordingly, the second apparatus 220 may receive (211) the fourth message 245 from the third apparatus 230. The fourth message 245 is for indicating whether the additional training data should be used by the second apparatus 220 to perform the model updating of one or more models associated with the first one or more apparatuses. In some embodiments, the fourth message 245 may further indicate where to obtain the additional training data, and / or how to select the additional training data if the additional training data should be used. As mentioned above, in some examples, the fourth message 245 may also be used for indicating the first apparatus 221 to perform the model updating. In such examples, operations 209 and 211 is performed prior to operations 201a and 203a, as shown in Fig. 2A. In some examples, the fourth message 245 may comprise a list of model identifying information for identifying the one or more models, version information of the one or more models in which each of the one or more models has corresponding version information, and one or more status indications, in which each of the one or more status indications indicates a status of a model among the one or more models, and the status of the model comprises frozen or unfrozen. Based on receiving the fourth message 245, the second apparatus 220 updates (213) the one or more models. In some examples, in a process of updating the one or more models, if the additional training data should be used by the second apparatus 220 to perform the model updating, the second apparatus 220 uses the additional training data to update at least one model with status indication indicating unfrozen (i.e. not converged) among the one or more models.
[0188] As mentioned above, the first apparatus 221 may be indicated to perform the model updating. The first apparatus 221 will accordingly perform the model updating (i.e. according to the indication 225 or 235) . In a process of performing the model updating, the first apparatus 221 may update at least one model among the one or more models. After completing the model updating, the first apparatus 221 may transmit (215) a first notification 255 for notifying that the model updating of the one or more models is completed. In some examples, the first notification 255 may be transmitted from the first apparatus 221 to the third apparatus 230. In some other examples, the first notification 255 may be transmitted from the first apparatus 221 and via the control plane function (e.g. the NWC or the AMF, etc. ) to the third apparatus 230. On the third apparatus 230 side, the third apparatus 230 may receive (217) the first notification 255. In some embodiments, the first notification 255 may further comprise a list of model identifying information for identifying the one or more models, or one or more status indications, or the combination thereof.
[0189] In some embodiments, based on the first notification 255, the third apparatus 230 may determine, from the second one or more apparatuses, a sixth apparatus for receiving at least one latest version of the one or more models. The at least one latest version is obtained through the model updating performed by the first apparatus 221, and a latest version corresponds to an updated model.
[0190] Further, the third apparatus 230 may transmit a third message (for example a control command transmitted at operation 517 below) to the first apparatus 221. The third message is for notifying the first apparatus 221 to transmit at least one latest version of the one or more models obtained through the model updating, for example, the at least one latest version of the one or more models may be transmitted to the sixth apparatus among the second one or more apparatuses. In some examples, the sixth apparatus may be the second apparatus 220, that is, the sixth apparatus and the second apparatus 220 may be the same apparatus. On the first apparatus 221 side, the first apparatus 221 receives the third message from the third apparatus 230, and then transmits the at least one latest version. In some examples, the third apparatus 230 transmits to the first apparatus 221, and the first apparatus 221 receives from the third apparatus 230, the third message via a control plane function. In some other examples, the third apparatus 230 transmits and the first apparatus 221 receives the third message via a data plane function. In some embodiments, the first apparatus 221 transmits and a data plane function receives the at least one latest version, and the data plane function is configured to transport the at least one latest version to the sixth apparatus among second one or more apparatuses, and the first apparatus 221 may not know which apparatus among the second one or more apparatuses is the sixth apparatus. In some examples, the third message may further comprise model identifying information for identifying at least one model associated with the at least one latest version.
[0191] As mentioned above, in some examples, the sixth apparatus may be the same as the second apparatus 220, in such examples, the third apparatus 230 may transmit a second notification, and the second apparatus 220 may receive the second notification from the third apparatus 230. In some other examples, the sixth apparatus is different from the second apparatus 220, in such examples, third apparatus 230 may transmit the second notification, and the sixth apparatus may receive the second notification from the third apparatus 230. The second notification is for notifying that the second apparatus 220 or the sixth apparatus is selected for receiving at least one latest version of the one or more models from the first apparatus 221, in which the at least one latest version is obtained through the model updating performed by the first apparatus 221. In some examples, the second notification further comprises model identifying information for identifying at least one model associated with the at least one latest version.
[0192] In some embodiments, the third apparatus 230 may transmit and the second apparatus 220 may receive, a configuration for the second apparatus 220 to obtain one or more latest versions from a further apparatus among the second one or more apparatuses. In some embodiments, the third apparatus 230 may transmit and the sixth apparatus may receive, a configuration for the sixth apparatus to obtain one or more latest versions from a further apparatus among the second one or more apparatuses. In the embodiments above, the one or more latest versions obtained from the further apparatus above are among a plurality of versions of the one or more models excluding the at least one latest version to be received by the second apparatus 220 or the sixth apparatus above.
[0193] FIG. 2B illustrates another example process according to some other embodiments of the present disclosure. FIG. 2B illustrates an example process 200-2 according to some embodiments of the present disclosure. As shown in FIG. 2B, a third apparatus 231, a fourth apparatus 240 and a fifth apparatus 250 are involved in the process 200-2. In some embodiments, the third apparatus 231, the four apparatus 240, and the fifth apparatus 250 may be the same as the third apparatus 230, the four apparatus, and the fifth apparatus in the process 200-1, respectively. In some embodiments, one or more other apparatuses (not shown in FIG. 2B) are involved in the process 200-2, such as a first apparatus, a second apparatus, and / or a sixth apparatus, and these apparatuses may also be the same as the first apparatus 221, the second apparatus 220, and / or the sixth apparatus mentioned in a process 200-1 (with reference to FIG. 2A) , respectively.
[0194] In the process 200-2, the third apparatus 231 may transmit (202) a fifth message 265 to the fourth apparatus 240. The fifth message 265 is for requesting to influence traffic routing. In some examples, the fifth message 265 may comprise information of data traffic identifying data traffic of the at least one application associated with second one or more apparatuses (e.g. the second one or more apparatuses in the process 200-1) , in which the data traffic is to be routed during the traffic routing; information of first one or more apparatuses (e.g. the first one or more apparatuses in the process 200-1) ; information of at least one location of the at least one application associated with the second apparatus; information of one or more requirements associated with the traffic routing; information of traffic filtering associated with the data traffic; or the combination thereof. Influencing the traffic routing is for selecting the second apparatus among second one or more apparatuses and at least one candidate apparatus among first one or more apparatuses. The second apparatus and the at least one candidate apparatus jointly perform model updating of one or more models associated with the first one or more apparatuses, and the fourth apparatus 240 is configured to perform a policy control function. On the fourth apparatus 240 side, the fourth apparatus 240 receives (204b) the fifth message 265 from the third apparatus 231. Further, based on the fifth message 265, the fourth apparatus 240 may generate (206) at least one policy 275 on influencing the traffic routing. In some embodiments, the fourth apparatus 240 may transmit (208b) the at least one policy 275 to the fifth apparatus 250 configured to be a network controller for determining (selecting) the second apparatus and the at least one candidate apparatus. On the fifth apparatus 250 side, the fifth apparatus 250 may receive (212) , from the fourth apparatus 240, the at least one policy 275 on influencing traffic routing. Then the fifth apparatus 250 may select (214) , based on the at least one policy 275, the second apparatus and the at least one candidate apparatus. Further, the fifth apparatus 250 may transmit (216) , to the third apparatus 231, a sixth message 285 for indicating the second apparatus and the at least one candidate apparatus, in which the third apparatus may be configured to control or manage the second one or more apparatuses. On the third apparatus 231 side, the third apparatus 231 may receive (218) the sixth message 285 from the fifth apparatus. For more details of the fifth message 265, the at least one policy 275, and the sixth message 285 may refer to the corresponding descriptions in the process 200-1 above.
[0195] Similar to the process 200-1, in some embodiments, the fourth apparatus 240 may transmit, to the first apparatus, a second message for notifying that the first apparatus is selected to perform the model updating. Then the fourth apparatus 240 may receive, from the first apparatus, a positive ACK for confirmation of performing the model updating or a negative ACK for rejection of performing the model updating. In some other embodiments, the fifth apparatus 250 may transmit a second message to the first apparatus to notify that the first apparatus is selected to perform the model updating, and then the fifth apparatus may receive, from the first apparatus, a positive ACK for confirmation of performing the model updating or a negative ACK for rejection of performing the model updating. If the negative ACK for rejection of performing the model updating is received by the fourth apparatus 240 or the fifth apparatus 250, a new apparatus different from the first apparatus transmitting the rejection above may be selected as a new first apparatus to perform the model updating. The first apparatus to perform the model updating may also be referred to as a target first apparatus. In the embodiments of the present disclosure, the operations related to the first apparatus 221 in the process 200-1 and the operations related to the model updating by the first apparatus in the process 200-2 are performed by the target first apparatus.
[0196] In some embodiments, the fourth apparatus 240 and the fifth apparatus 250 are integrated as a single network entity, that is, from the single network entity side, the fourth apparatus 240 has the function of the fifth apparatus 250, for example, the fourth apparatus 240 may transmit traffic routing information indicating the traffic routing to a seventh apparatus. The seventh apparatus is for selecting a data plane path for the traffic routing. In some examples, the fourth apparatus 240 may receive, from the seventh apparatus, a third notification that the data plane path has been configured for the traffic routing. In some other examples, the fourth apparatus 240 may receive, from the seventh apparatus, a response to the transmission of the traffic routing information. The response comprises the third notification that the data plane path has been configured for the traffic routing.
[0197] In some other embodiments, the fourth apparatus 240 and the fifth apparatus 250 are not integrated. The fifth apparatus 250 may transmit, to the seventh apparatus, traffic routing information indicating the traffic routing. The fifth apparatus 250 may receive, from the seventh apparatus, a third notification that a data plane path has been configured for the traffic routing, or may receive, from the seventh apparatus, a response to the transmission of the traffic routing information, in which the response comprises the third notification that the data plane path has been configured for the traffic routing. In some embodiments, the traffic routing information may comprise information (e.g. an ID or a network address) identifying an application location corresponding to the second apparatus 220 among at least one application location; information (e.g. an ID or a network address) identifying the first apparatus; information associated with (or included in) the at least one policy 275; or information of a data plane path (e.g. IDs of DPFs in the data plane path, and information describing how the DPFs are connected along the data plane path) for the traffic routing; or any combination thereof.
[0198] In some other embodiments, the fifth apparatus 250 may be integrated within / implemented by the seventh apparatus. Thus, in this integrated single network entity, interaction between the fifth apparatus 250 and the seventh apparatus may be internal process of the seventh apparatus (is therefore optional in the procedure) .
[0199] It should be noted that the process 200-2 may further comprise one or more of other operations performed by the first apparatus 221, the second apparatus 220, the third apparatus 230, the fourth apparatus, the fifth apparatus, and / or the sixth apparatus in the process 200-1, and in some embodiments, the process 200-1 and the process 200-2 may be combined to be implemented.
[0200] As mentioned above, each of a group of network entities, such as UE, server, network function, RAN node, etc., or a combination thereof, has a local AI model (or simply put, a local model) and a private dataset. The local models may be heterogeneous, i.e. different in their structures. Among the private datasets, different private datasets may include different data (diversified data) . The goal is to train each of these local models (e.g. the local models described in the embodiments related to FIGs. 2A and 2B) using the private datasets collectively (benefiting from the data diversity) , without requiring the private datasets to leave their hosting network entities (protecting the data privacy) . Data in the private datasets may be highly heterogeneous (biased, or unbalanced) , and the data heterogeneity can negatively impact the training performance. In some examples of the present disclosure, the group of network entities may be an example of the first one or more apparatuses, e.g. type-1 PSFs (processing service functions) in some embodiments, for example, type-1 PSFs 1 to 5 in FIG. 4A, or type-1 PSFs mentioned in FIGs 5 and 6, etc. . In some examples, one of the first one or more apparatuses, e.g. the first apparatus 221, and an example of the first apparatus 221 is a target type-1 PSF. The systems and methods of the embodiments of the present disclosure may achieve the above-described goal. In some examples, during the training of the local models, the solution of the embodiments of the present disclosure may use additional training data to mitigate the negative impact from biased private datasets. The additional training data are selected from one or multiple datasets that are different from the private datasets and are transparent to the type-1 PSFs. The additional training data is selected by a processing service controller (PSC) and is used at or by one or multiple type-2 PSFs. As such, use of the additional training data does not increase complexity at type-1 PSFs, which is important especially for type-1 PSFs that are wireless devices such as UEs. A type-2 PSF is a network entity (e.g. server, network function, RAN node, etc. ) that does not have (or is not associated with) a local model to train, but will help train local models of the type-1 PSFs. The type-2 PSF may be an example of the second apparatus (e.g. the second apparatus 220) in the embodiments described above.
[0201] FIG. 3 illustrates an example training process 300 according to some embodiments of the present disclosure, as shown in FIG. 3, the training process 300 of models may comprise three phases: a model collection phase 310, a model evolution phase 330 and a model distribution phase 340. The training process 300 starts with the model collection phase 310 and ends with the model distribution phase 340. The model evolution phase 330 is performed iteratively between the model collection phase 310 and the model distribution phase 340. In the model evolution phase 330, the model updating (as mentioned in some embodiments above) may be performed. Each of the type-1 PSFs (an example of the first one or more apparatuses above) , may join the training process 300 by accessing an application corresponding to the training. When accessing the application, a type-1 PSF may send (e.g. operation 519 in FIG. 5) and / or receive (e.g. operation 609 in FIG. 6) data traffic related to the application. The data traffic is referred to as application traffic. In the model evolution phase 330, a type-1 PSF as a target type-1 PSF among the type-1 PSFs may perform the operations performed by the first apparatus 221 in FIG. 2A. A type-2 PSF may perform the operations performed by the second apparatus 220 in FIG. 2A. During the model evolution phase 330, there are some interactions among the entities, such as the type-1 PSF, the type-2 PSF, and other entities (including a PSC (an example of the third apparatus 231) , a PCF (an example of the fourth apparatus 230, an NWC (an example of the fifth apparatus 250) , etc. ) . The details of the interactions may refer to FIG. 2A and FIG. 2B.
[0202] In the model collection phase 310, each of the type-1 PSFs trains its local model using its private dataset and then sends the local model to the type-2 PSF. The trained local model (or referred to as the updated local model) may be sent to the type-2 PSF at the end of an iteration during the training process, and for different iterations, the type-2 PSFs for receiving the models may be the same or different. The local model and the private dataset are considered associated or corresponding to the type-1 PSF. The type-1 PSF may indicate whether the local model is frozen or not (i.e. whether the local model is frozen ( (i.e. converged) ) or unfrozen (i.e. not converged) ) when sending the local model to the type-2 PSF. All the type-1 PSFs send their local models to the same type-2 PSF. These local models are considered at their initial version and collected at the type-2 PSF.
[0203] In the model evolution phase 330, the model evolution may be performed distributively at different entities, such as, at least one type-1 PSF and the type-2 PSF, and the model evolution procedure may comprise a plurality of iterations. The type-2 PSF maintains state information of each local model, which indicates whether the local model is frozen or not (i.e. whether the local model is frozen or unfrozen) and is received from the corresponding type-1 PSF. After the model collection phase 310 or after an iteration of the model evolution phase 330, according to the state information, the type-2 PSF evaluates a termination condition of the training process and determines whether the termination condition is met 320. The termination condition is met, for example, when all local models are frozen, or when the number of local models that are frozen is larger or not smaller than certain threshold value, or when the ratio of local models frozen to those not frozen is larger or not smaller than certain threshold value. If the termination condition is met, the type-2 PSF triggers / performs the model distribution phase 340. Otherwise, the type-2 PSF triggers / performs a (next) iteration of the model evolution phase 330. In an iteration of the model evolution phase 330, a local model will not (or should not) evolve if the local mode is frozen, and will evolve otherwise. When the local model does not evolve, the local model stays unchanged, that is, values of model parameters of the local model remain unchanged. When the local model evolves, the local model evolves from current version to a next version, wherein value of at least one model parameter of the local model changed (e.g. from a source value, which is associated to the current version, to a target value, which is associated to the next version) . The local model may evolve from current version to the next version via an intermediate version, for example, the value of the at least one model parameter changes from the source value to the target value via an intermediate value that is associated to the intermediate version.
[0204] In the model distribution phase 340, the type-2 PSF notifies each type-1 PSF that the model training process terminates. When notifying a type-1 PSF that the model training process 300 terminates, the type-2 PSF may send a latest version of the local model associated to the type-1 PSF to the type-1 PSF. The sending may be optional, for example, when the type-1 PSF already has the latest version of the associated local model.
[0205] In the embodiments of FIG. 3, it has been mentioned that the model evolution phase 330 is performed iteratively between the model collection phase 310 and the model distribution phase 340. The details of an iteration of the model evolution phase 330 may further refer to the embodiments below. FIG. 4A illustrates an example process 400 of an iteration (referred to as current iteration) of model evolution according to some embodiments of the present disclosure. As mentioned above, the type-2 PSF collects local models of a plurality of type-1 PSFs. There are five models are shown as an example. The five models, such as M1, M2, M3, M4, and M5, corresponding to type-1 PSF 1, type-1 PSF 2, type-1 PSF 3, type-1 PSF 4, and type-1 PSF 5, respectively. These local model are initially trained at corresponding type-1 PSFs, respectively. In an operation 402, as shown in FIG. 4A, the model evolution is performed at the type-2 PSF, and in some examples, additional training data may be used for the model evolution. After the model evolution performed at the type-2 PSF, intermediate versions of the local models are obtained. In an operation 404, the type-2 PSF sends the intermediate versions of the local models to a target type-1 PSF among the plurality of type-1 PSFs. As an example, the target type-1 PSF is the type-1 PSF 1, and the intermediate versions of the local models may be M’1, M’2, M’3, M’4, and M’5. The five intermediate versions of the models, i.e. M’1, M’2, M’3, M’4, and M’5, corresponding to the type-1 PSF 1, the type-1 PSF 2, the type-1 PSF 3, the type-1 PSF 4, and the type-1 PSF 5, respectively. In an operation 406, the model evolution is performed at the target type-1 PSF, and updated local models are obtained. As shown in an operation 408, the M1, M2, M3, M4, and M5 in this operation refer to the updated local models or evolved local models, and these models are sent from the target type-1 PSF to the type-2 PSF, and may evolve further through a next iteration of the model evolution.
[0206] Specifically, in the current iteration, the target type-1 PSF is selected from the group of type-1 PSFs with respect to a number of factors, with the goal of optimizing network performance and training performance. The factors taken into account when selecting a type-1 PSF as the target type-1 PSF may comprise: network conditions (e.g. throughput, delay, rate) , or status (e.g. loading, computing capability) of the type-1 PSF, or history (e.g. number of times) of the type-1 PSF being selected as target type-1 PSF, or any combination thereof. Network conditions may be statistical or instantons, related to the type-1 PSF. The network conditions may be informed by or obtained from a network function that maintains (estimates or analyzes) information about the network conditions, e.g. NWDAF in the 3GPP 5G system as described in 3GPP TS 23.501 V18.2.0. Status of the type-1 PSF may be statistical or instantons. The status may be informed by or obtained from the type-1 PSF. The history of the type-1 PSF being selected as target type-1 PSF may be maintained locally, or informed by or obtained from a storage function.
[0207] One type-1 PSF (e.g. type-1 PSF 1 in FIG. 4A) may be selected as the target type-1 PSF by a network controller (NWC) , e.g. the NWC described in the system architecture in association to the FIG. 4B. A different type-1 PSF may be selected as the target type-1 PSF in different iterations of the model evolution. In the current iteration, if a local model is not frozen (as indicated by the corresponding type-1 PSF) , the type-2 PSF updates the local model, e.g. by knowledge transfer and / or model training, using additional data, such that the local model evolves to an intermediate version. The type-2 PSF sends the intermediate version of the local model to the target type-1 PSF, and the type-2 PSF may further indicate to the target type-1 PSF that the local model is not frozen. If a local model is frozen (as indicated by the corresponding type-1 PSF) , the type-2 PSF will not update the local model, and the type-2 PSF will send the current version of the local model to the target type-1 PSF as the intermediate version. When sending the current version of the local model to the target type-1 PSF, the type-2 PSF may further indicate to the target type-1 PSF that the local model is frozen. In some embodiments, the additional data may be used to evolve the local model from current version to the intermediate version as described above includes a group of data samples. The data samples are selected from a dataset that may be different from any of the private datasets. The dataset may be a public dataset, and may or may not be located at the type-2 PSF. The data samples are selected according to information about the target type-1 PSF’s private dataset, e.g. dataset size, label distribution, if the information is available (e.g. provided by the target type-1 PSF when joining the training process) , such that, for example, the number of data samples satisfies certain condition, e.g. not larger or smaller than a ratio to the private dataset size, or per-label data sample distribution satisfies certain condition, e.g., being correlated to per-label data distribution in the private data set. In some embodiments, the data samples may be selected by a processing service controller (PSC) , e.g. the PSC described in the system architecture in association to the FIG. 4B.
[0208] In the current iteration, the target type-1 PSF receives local models (possibly including its own local model, i.e the local model associated to the target type-1 PSF) from the type-2 PSF and may update some or all of the received local models and its own model using its private dataset. The received local models by the target type-1 PSF may be intermediate version of the models determined during a previous iteration. When updating a local model (whether a received one or its own one) , the type-1 PSF may uses other local models as well (e.g. by transferring knowledge from the other local models to the local model) . If a local model is frozen, the type-1 PSF will not update the local model, but may still use the local model to update other local models that are not frozen. If a local model is not frozen, the target type-1 PSF may update the local module, e.g. by continuing the knowledge transfer and / or training, using its private dataset, such that the local model evolves to the next version (from the intermediate version) . If a local model is updated, the target type-1 PSF sends the updated local model to the type-2 PSF. If a local model is not updated, the target type-1 PSF may indicate to the type-2 PSF that the local model is not updated. The target type-1 PSF may indicate to the type-2 PSF whether or not its own local model is frozen. The target type-1 PSF determines whether or not its own local model is frozen according to the convergence status of the local model. For example, if the local model converged during the updating of the local model, it determines that the local model is frozen. If the local model did not converge during the updating of the local model, it determines that the local model is not frozen.
[0209] The examples of the apparatuses mentioned in the process 200-1 and the process 200-2 may refer to FIG. 4B, FIG. 5 and / or FIG. 6. The “model updating” mentioned in the process 200-1 and the process 200-2 corresponds to the step 406 in the Fig. 4A. The first apparatus in the process 200-1 and the process 200-2 can be the target type-1 PSF. The first one or more apparatuses in the in the process 200-1 and the process 200-2 can be the type-1 PSFs. For each iteration of the model evolution, a target type-1 PSF is selected among the type-1 PSFs. For example, the type-1 PSF 401 or the type-1 PSF 510 or the type-1 PSF A 610 is selected for performing the model updating, then it may be further referred to as the target type-1 PSF 401, 510 or 610, and is an example of the first apparatus 221. The type-2 PSF 403 or the first type-2 PSF 540 or the type-2 PSF B 640 may be an example of the second apparatus 220. The PSC (processing service controller) 419 or 560 may be an example of the third apparatus 230. The PCF 407 or 630 may be an example of the fourth apparatus 240. The NWC 411 or 520 may be an example of the fifth apparatus 250. The type-2 PSF 403 or the second type-2 PSF 550 may be an example of the sixth apparatus. The PMF 413 or 620 may be an example of the seventh apparatus. Some examples will be further described with reference to one or more FIGs of the present disclosure.
[0210] FIG. 4B illustrates an example architecture of a system 400-1 according to some embodiments of the present disclosure. As shown in FIG. 4B, the system may include a type-1 PSF 401, a core network (CN) and a type-2 PSF 403. The CN may include a control plane and a data plane. The control plane includes the following control plane functions: NEF 405, PCF 407, NSF 411a, NWC 411, and PMF 413. In some embodiments, the control plane of the CN further includes a control plane function AMF (the AMF is not shown in FIG. 4B) . The data plane includes a data plane function (DPF) 415. In some embodiments, the data plane is known as user plane. The type-1 PSF 401 connects with both the control plane and the data plane. The type-2 PSF 403 connects with the data plane. The type-1 PSF 401 has (is associated with) a model to train and has (is associated with) a private dataset that can be used by the type-1 PSF 401 to train the model. The model is the local model of the type-1 PSF 401. The type-2 PSF 403 does not have (is not associated with) a model to train, but can help train the model associated with the type-1 PSF 401. In some examples, the type-1 PSF 401 may be a wireless terminal device, such as a UE, or a server (e.g. an AS) , or a network function. In some examples of the type-1 PSF 401 being a wireless terminal device, the type-1 PSF 401 connects with the CN (including the control plane and the data plane) via a radio access network (RAN) 417. In this case, the system includes the RAN 417. In some examples of the type-1 PSF 401 being not a wireless terminal device, the type-1 PSF 401 may connect with the CN directly, and the RAN 417 is optional in the system 400-1.
[0211] The system may further include a processing service controller (PSC) 419. In some embodiments, the PSC 419 is a part of the control plane of the CN, and the type-2 PSF 403 is a part of the data plane of the CN. The PSC 419 may manage / control the type-2 PSF 403. The PSC 419 interacts with the control plane of the CN to manage, control or influence the control plane’s decision, e.g. decision on traffic routing. When interacting with the control plane, the PSC 419 may interact with relevant control plane functions, for example, the NSF 411a, the PCF 407, the NWC 411, and / or the PMF 413. In some embodiments, for example, the PSC 419 interacts directly with the relevant control plane functions. In some embodiments, the PSC 419 interacts indirectly with the relevant control plane functions, via the NEF 405. In some embodiments, the NWC 411 and the PSC 419 may be integrated as a same network entity.
[0212] Some network elements involved in the architecture of the system 400-1 may be further described as below. The AMF (access and mobility management function) may implement functionalities the same as or similar to those of the AMF in the 3GPP 5G system, e.g. registration management, connection management, reachability management, mobility Management, access authentication, access authorization, etc., as described in 3GPP TS 23.501 V18.2.0. In some embodiments, the AMF corresponds to the AMF in the 3GPP 5G system (i.e. 5G PCF in short) . NEF (network exposure function) 405 may implement the following functionalities, for example, but not limited to: exposure of network information, analytics, capabilities and events, secure provision of information from external application (e.g. AF) to the system, translation of internal-external information, etc. In some embodiments, the NEF 405 corresponds to the NEF in the 3GPP 5G system (i.e. 5G NEF in short) as described in 3GPP TS 23.501 V18.2.0. PCF (policy control function) 407 may implement the following functionalities, for example, but not limited: support of unified policy framework to govern network behavior, making policy decisions and providing resultant policy rules to control plane function (s) to enforce them, accessing subscription information and application data relevant for policy decisions in one or multiple network storage functions (e.g. NSF 411a) . In some embodiments, the PCF 407 corresponds to the PCF in the 3GPP 5G system (i.e. 5G PCF in short) as described in 3GPP TS 23.501 V18.2.0. NSF (network storage function) 411a may implement the following functionalities, for example, but not limited: storage and retrieval of subscription data, storage and retrieval of policy data, storage and retrieval of structured data for exposure, storage and retrieval of application data (including Packet Flow Descriptions (PFDs) for application detection, information associated with AF request, device group information for group management) , etc.
[0213] In some embodiments, the NSF 411a corresponds to the UDR in the 3GPP 5G system as described in 3GPP TS 23.501 V18.2.0. NWC (network controller) 411 makes traffic routing decisions and coordinates traffic routing, as described in the embodiments of the present disclosure. In some embodiments, the NWC 411 is integrated with the PMF 413. In some embodiments, the NWC 411 is integrated with the PCF 407. In some embodiments, the NWC 411 is a separate network function, different from the PMF 413 and the PCF 407. PMF (path management function) 413 may implement the following functionalities, for example, but not limited to: selection and control of DPF 415, maintaining / configuring data plane topology or paths, establishing and releasing data plan tunnels, configuring traffic forwarding at DPF 415 to apply local switching or packet forwarding, configures traffic steering at DPF 415 to route traffic to proper destination, etc. The PMF 413 configures a DPF 415 via the T4 interface. In some embodiments, the PMF 413 corresponds to the SMF in the 3GPP 5G system as described in 3GPP TS 23.501 V18.2.0. DPF (data plane function) 415 may implement the following functionalities, for example, but not limited: acting as a point of interconnect to the DN, acting as an anchor point for a device in the DP, routing / forwarding packets, enforcing policy rules (e.g. traffic gating, redirection, traffic steering) , performing traffic marking in the uplink and downlink, performing packet buffering and data notification triggering, performing packet inspection (e.g. application detection) . In some embodiments, the data plane corresponds to the user plane of the 3GPP 5G system, and the DPF 415 corresponds to the UPF in the 3GPP 5G system as described in 3GPP TS 23.501 V18.2.0.
[0214] The control plane functions and the DPF 415 are logical network functions, and each of them can be instantiated at one or multiple network locations, resulting in one or multiple instances. When a RAN 417 implements functionalities of a logical network function (any of the control plane functions and the DPF 415) , it is considered that the logical network function is instantiated at the RAN node. A network location may refer to a data center or a RAN node. There may be more than one instances of a logical network function at a same network location. The type-1 PSF 401connects with the DPF 415 (in face, an instance of the DPF 415) in the data plane of the CN, which is an anchor point for the type-1 PSF 401 in the data plane and can be referred to as DPA. The DPA in turn connects with the type-2 PSF 403. In some examples, it is possible that the type-1 PSF 401 connects with the DPA through one or multiple intermediate DPFs (i.e. other instances of the DPF 415) . The one or multiple intermediate DPFs are located between the type-1 PSF 401 and the DPA. The one or multiple intermediate DPFs (if any) , the DPA and communication tunnels in between constitute a data plane path connecting the type-1 PSF 401 and the type-2 PSF 403. The data plane path is considered associated to the type-1 PSF 401. When the type-1 PSF 401 is a wireless terminal device such as a UE, the type-1 PSF 401 is served by a RAN node in the RAN 417, i.e. having a wireless connection with the RAN node. In this case, the type-1 PSF 401 connects with the DPA through the RAN node. The RAN node is considered part of the data plane path. If the data plane path includes one or multiple intermediate DPFs, the one or multiple intermediate DPFs are located between the serving RAN node and the DPA.
[0215] A control plane function (CPF) , such as the AMF, the PMF 413 or the NWC 411, can interact with the type-1 PSF 401 via a first interface, shown as a dashed line in FIG. 6. In some embodiments, the interaction is through a relay function in the control plane. In this case, when the CPF interacts with the type-1 PSF 401 via the first interface, the CPF interacts with the relay function via a second interface, which in turn interacts with the type-1 PSF 401 with a third interface. For example, when the CPF sends a message to the type-1 PSF 401 via the first interface (e.g. when the NWC (i.e. the CPF) forwards / sends a notification message to the type-1 PSF 401 as described in the operation 503 or 517 in FIG. 5 or in the operation 609 (c) in FIG. 6) , the CPF sends the message to the rely function using the second interface, which then forwards the message to the type-1 PSF 401 using the third interface. When the type-1 PSF 401 sends a message to the CPF via the first interface (e.g. when the type-1 PSF 401 sends a notification message to the NWC 411 as described in the operation 513 in FIG. 5, or when the type-1 PSF 401 sends an ACK message to the NWC 411 as described in the operation 609 (c) in FIG. 6) , the type-1 PSF 401 sends the message to the relay function using the third interface, which then forwards the message to the CPF using the second interface. Before forward the message, the relay function may perform information mapping, wherein the control plane function changes or updates some original information in the message to a mapped information, such that the message includes the mapped information instead of the original information when being forwarded.
[0216] Hence, the first interface (between the CPF and the type-1 PSF) can be viewed as comprising or implemented through the second interface (between the control plane function and the relay function) and the third interface (the relay function and the type-1 PSF 401) , and the first interface is said through the relay function. In some embodiments, the first interface and the third interface are the same interface in the view of the type-1 PSF 401. In some embodiments, the first interface and the second interface are the same interface in the view of the CPF. In some embodiments, the first interface, the second interface or the third interface corresponds to a communication tunnel. In some embodiments, when the type-1 PSF 401 is a wireless terminal device, the first interface is through the RAN 417 (e.g. the serving RAN node of the type-1 PSF 401) and is a NAS (Non Access Stratum) interface. A message transmitted using the first interface is a NAS message, for example, when the NWC (i.e. the CPF) forwards / sends a notification message to a wireless terminal device (i.e. the type-1 PSF 401) via the RAN 417 as described in the operation 503 or 517 in FIG. 5 or in the operation 609 (c) in FIG. 6, the notification message is a NAS message. If the first interface is implemented through the second interface and the third interfaces, the third interface is through the RAN 417 and is also a NAS interface, and the first interface being through the RAN 417 is due to the third interface being through the RAN 417. When the relay function forwards the message, the relay function may include the message, or content of the message, in another message and send the other message using the third interface (if the message is from the CPF and targeting the type-1 PSF 401) or the second interface (if the message is from the type-1 PSF 401 and targeting the CPF) . The other message is a NAS message if the other message is sent using the third interface.
[0217] If the CPF is the NWC 411, the relay function may be the PMF 413 or the AMF. Alternatively, if the CPF is the PMF 413, the relay function may be the AMF. It should be noted that the above description about the first interface between the CPF and the type-1 PSF 401, when the type-1 PSF 401 is a wireless terminal device, is recursive and can be applied to the third interface. That is, if the relay function is the PMF 413, the third interface may be through another relay function, e.g. the AMF. When the type-1 PSF 401 is not a wireless terminal device, in some examples, if the CPF is the NWC 411, the relay function may be the PMF 413 or the NEF 405. Alternatively, in some other examples, if the CPF is the PMF 413, the relay function may be the NEF 405. It should be noted that the above description about the first interface between the CPF and the type-1 PSF 401, when the type-1 PSF 401 is not a wireless terminal device, is recursive and can be applied to the third interface. That is, if the relay function is the PMF 413, the third interface may be through another relay function, e.g. the NEF 405.
[0218] If the type-1 PSF 401 is a wireless terminal device like UE, the PSC 419 informs the type-1 PSF 401 through a NAS message. The NAS message is transported to the type-1 PSF 401 via a RAN node. In some embodiment, the PSC 419 sends the NAS message to the RAN node and the RAN node forwards the NAS message to the type-1 PSF 401. In some embodiments, the NAS message is generated by a network entity, and the PSC 419 sends a message to network entity, and the network entity includes content of the message in the NAS message. In some embodiments, the NAS message is relayed through another network entity to the RAN node. When relaying the NAS message, the other network entity includes the NAS message in another NAS message, and sends the other NAS message to the type-1 PSF 401.
[0219] In some embodiments, the system architecture described above can map / correspond to the 3GPP 5G system architecture as described in 3GPP TS 23.501 V18.2.0. For example, the DPF 415 maps / corresponds to the 5G UPF; the AMF maps / corresponds to the 5G AMF; the PMF 413 maps / corresponds to the 5G SMF; the PCF 407 maps / corresponds to the 5G PCF; the NSF 411a maps / corresponds to the 5G UDR; the PSC 419 maps / corresponds to the AF; the type-2 PSF 403 maps / corresponds to an AS; the type-1 PSF 401 maps / corresponds to a UE; and the NWC 411 is integrated within, in other words, functionalities of the NWC 411 is implemented by the 5G SMF or the 5G PCF.
[0220] FIG. 5 illustrates an example procedure for an iteration of model evolution (e.g. operation in the phase 330 in Fig. 3) according to some other embodiments of the present disclosure. As shown in FIG. 5, in the procedure 500, at least one type-1 PSF (a type-1 PSF 510 is shown as an example) , an NWC 520, a data plane 530, at least one type-2 PSF (a first type-2 PSF 540 and a second type-2 PSF 550 are shown as an example) , and a PSC 560 are involved. In an example iteration process, the type-1 PSF 510 and the first type-2 PSF 540 may be involved. As an example, the type-1 PSF 510 is selected to be the target type-1 PSF. The iteration of model evolution involves at least one model, each of which being a local model associated / corresponding to a type-1 PSF. The at least one model includes the local model associated to the target type-1 PSF. For each of the at least one local model, the PSC 560 maintains status information about the local model (i.e. whether the local model is frozen or not) , which is received from the corresponding type-1 PSF. It is possible that the first type-2 PSF 540 has collected latest versions of the at least one model before the iteration of the model evolution starts, for example, when the all the type-1 PSFs send their local modes to the first type-2 PSF 540 during the model distribution phase. At the end of the iteration of the model evolution, the second type-2 PSF 550 may collect latest versions of the at least one model, e.g. as illustrated by the operation 519 in Fig. 5. In some embodiments, the first type-2 PSF 540 and the second type-2 PSF 550 are the same entity. In some embodiments, the first type-2 PSF 540 and the second type-2 PSF 550 are different entities. Among the at least one model, a model may be frozen and does evolve (in other words, stay unchanged) during the iteration of model evolution. In this case, model parameters of the model have same values at the end of the iteration of model evolution as before the iteration of model evolution starts. In some examples, the procedure 500 may be an example of interactions among some entities mentioned in process 200-1 (shown at FIG. 2A) and / or process 200-2 (shown at FIG. 2B) . In such examples, the type-1 PSF 510 may be an example of the first apparatus 221, and the first type-2 PSF 540 may be an example of the second apparatus 220.
[0221] In an operation 501, the type-1 PSF 510 is selected as the target type-1 PSF, and an application location, which corresponds to the first type-2 PSF 540, is selected from one or multiple potential applications locations corresponding to one or more type-2 PSFs, and the data plane 530 is accordingly configured so that data traffic in the operation 509 will be routed from the first type-2 PSF 540 to the type-1 PSF 510.
[0222] This operation 501 can be performed through a procedure described in association to the Fig. 6, wherein the type-1 PSF A 610 is the type-1 PSF 510 and the type-2 PSF B 640 is the first type-2 PSF 540. In this operation, the PSC 560 requests to influence traffic routing for the application. The request from the PSC 560 is transported (operation 603 in Fig. 6) to the PCF 630. The request includes information about traffic, information about type-1 PSFs and information about application location. The information about traffic identifies the data traffic in the operation 509. The information about type-1 PSFs identifies a group of type-1 PSFs that the at least one model is associated to. The information about application location identifies the one or multiple potential application locations corresponding to one or more type-2 PSFs.
[0223] If the first type-2 PSF 540 does not have the latest version of a model among the at least one model, in this operation 501, the PSC 560 may configure the first type-2 PSF 540 to obtain the latest version of the model. The latest version of the model is stored at a third type-2 PSF (which may be the same as or different from the second type-2 PSF 550) . The PSC 560 provides the first type-2 PSF 540 with model identifying information (e.g. an ID, a name or an index) , version information (e.g. a version number) and information about the third type-2 PSF (e.g. an ID or a network address) , e.g. when configuring the type-2 PSF B 640 in the operation 609 (b) or 619 in Fig. 6. The model identifying information identifies the model. The version information identifies the latest version of the model. According to the information provided from the PSC 560, the first type-2 PSF 540 may interact with the third type-2 PSF to obtain the latest version of the model.
[0224] In an operation 503, the PSC 560 instructs the type-1 PSF 510 to perform model update for the at least one model. In this operation 503, the PSC 560 may notify the type-1 PSF 510 whether one or multiple models among the at least one model are frozen or not. The one or multiple models may comprise all of the at least one model. The at least one model may be transported to the type-1 PSF 510 in the operation 509.
[0225] In some embodiments, the PSC 560 performs this operation 503 by sending a control command to the type-1 PSF 510. In some embodiments, the PSC 560 performs this operation 503 by sending a message (referred to as notification message) to the type-1 PSF 510. The notification message includes a list of model identifying information (s) , respective version information (s) and respective status indication (s) . In the list, each model identifying information (e.g. an ID, a name or an index) identifies a model among the one or multiple models, and the respective version information (e.g. a version number) indicates freshness of the model, and the respective status indication indicates whether the model is frozen or not. In some embodiments, the model identifying information (s) and / or the respective version information (s) may be generated by the PSC 560. In some embodiments, the notification message may be sent to the type-1 PSF 510 via a control plane function, e.g. the NWC 520 or the AMF. That is, the PSC 560 sends the notification message to the control plane function, and the control plane function then forwards / sends the notification message to the type-1 PSF 510. In some embodiments, this operation 503 is optional, for example, when the first type-2 PSF 540 is configured to send a message containing the information (i.e. the list of model identifying information (s) , respective version information (s) and respective status indication (s) ) to the type-1 PSF as part of the operation 509. In other words, the information sent in the operation 503 may be transmitted in the operations 505 and 509 instead. The control command or the message (or referred to as the notification message from the PSC 560 to the type-1 PSF 510 may be an example of the indication 235 above.
[0226] In an operation 505, the PSC 560 configures the first type-2 PSF 540, specifically, the PSC 560 instructs the first type-2 PSF 540 to perform model update for the at least one model. In this operation 505, the PSC 560 may inform the first type-2 PSF 540 whether the one or multiple models are frozen or not. The one or multiple models may comprise all of the at least one model, as described in the operation 503. In some embodiments, the PSC 560 performs this operation 505 by sending a message (an example of the fourth message 245 above) to the first type-2 PSF 540. The message includes the list of model identifying information (s) , respective version information (s) and respective status indication (s) , as described in the operation 503. For a model identified in the message, the message may include an indication indicating whether additional training data is needed. If additional training data is needed, the indication may further indicate where to obtain the additional training data and how to select the additional training data. The additional training data will be used to perform the operation 507.
[0227] In an operation 507, the first type-2 PSF 540 performs model evolution, specifically, the first type-2 PSF 540 updates one or more models among the at least one model. The one or more models updated in this operation 507 may be among those not frozen as informed by the PSC 560 in the operation 503. If a model among the at least one model is not frozen (e.g. as informed by the PSC 560 in the operation 505) , the first type-2 PSF 540 may update the model in this operation 507. When updating the model, some or all of the values of model parameters of the model are updated (changed) , and the version information associated to the model is updated / changed as a result (for example, the version number is increased) . If a model among the at least one model is frozen (e.g. as informed by the PSC in the operation 505) , the first type-2 PSF 540 will accordingly not update the model in this operation 507. Because the model is not updated, values of model parameters of the model remain unchanged, and the version information associated to the model may also remain unchanged. In some embodiments, even if the model is not updated, the version information associated to the model is updated / changed (for example, the version number is increased) .
[0228] In an operation 509, the first type-2 PSF 540 sends, to the type-1 PSF 510, updated models, i.e. information about the at least one model. The information is included in a data traffic routed to the type-1 PSF 510 through the data plane 530. A data traffic targets a type-1 PSF. For a model among the at least one model, the information may include a model identifying information (e.g. an ID, a name or an index) , a version information (e.g. a version number) , values of model parameters of the model, and a status indication. The model identifying information identifies the model. The version information is associated to the model and corresponds to the values of model parameters of the model, and it indicates the freshness of the model. The status indication indicates whether the model is frozen or not. If the model has been updated in the operation 507, the values of model parameters of the model comprise updated values, and the version information is the updated version information, and the status indication indicates that the model is frozen. If the model is not updated in the operation 507, the values of model parameters of the model and the version information are the same as those before this procedure. In some embodiments, the values of model parameters of the model are optional, for example, when they are pre-configured at the type-1 PSF 510, and can be identified by the type-1 PSF 510 in local configuration according to the model identifying information and possibly the version number information. The status indication is consistent with the status indication described in the operation 505 for the model. If the status indication for the model in the operation 505 indicates the model is frozen, the status indication for the model in this operation 507 also indicates the model is frozen. If the status indication for the model in the operation 505 indicates the model is not frozen, the status indication for the model in this operation 507 also indicates the model is not frozen. In some examples, the model identifying information, the version information, and the status indication may be in a message (an example of the first message) from the first type-2 PSF 540 to the type-1 PSF 510. In some examples, the operation 503 is optional, and if the operation 503 is not needed, an indication 225 in the first message instead of the indication 235 may be indicated to the type-1 PSF 510. The indication 225 includes the same information as the indication 235, but the indication 225 is transmitted from the first type-2 PSF 540 to the type-1 PSF 510, for example, via the message above (an example of the first message) .
[0229] In an operation 511, the type-1 PSF 510 performs model evolution, specifically, updates one or more models among the at least one model. The one or more models updated in this operation 511 may comprise some or all of those not frozen as informed by the PSC 560 in the operation 505 or 507. In some embodiments, the one or more models include the model associated to the type-1 PSF 510, and the model associated to the type-1 PSF 510 may be indicated frozen in the operation 505 or 507. If a model among the at least one model is not frozen (i.e. is unfrozen) as informed by the PSC 560 in the operation 505, the type-1 PSF 510 may update the model in this operation 511. When updating the model, some or all of the values of model parameters of the model are updated (changed) , and the version information associated to the model is updated / changed as a result (for example, the version number is increased) . If a model among the at least one model is frozen as informed by the PSC 560 in the operation 505 (and in some embodiments if the model does not correspond to the type-1 PSF 510) , the type-1 PSF 510 will accordingly not update the model in this operation 511. Because the model is not updated, the values of model parameters of the model remain unchanged, and the version information associated to the model may also remain unchanged. In some embodiments, even if the model is not updated, the version information associated to the model is updated / changed (for example, the version number is increased) .
[0230] In an operation 513, the type-1 PSF 510 notifies the PSC 560 about the status (i.e. frozen or not frozen) of a model among the at least one model, by sending a message (referred to as notification message, may be an example of the first notification 255) to the PSC 560. In some embodiments, the notification message may be sent to the PSC 560 via a control plane function, e.g. the NWC 520 or the AMF. That is, the type-1 PSF 510 sends the notification message to the control plane function, and the control plane function then forwards / sends the message to the PSC 560. The notification message may include a status indication, indicating whether the model is frozen or not. The notification message may further include a model identifying information (e.g. an ID, index or name) , which identifies the model. In various embodiments, the model is associated to the type-1 PSF 510, and the status of the model is determined by the type-1 PSF 510.
[0231] In some embodiments, the status of the model indicated in this operation is not consistent with (or the same as) that indicated in the operation 505 or 507. For example, the model was indicated frozen in the operation 505 or 507 and not frozen in this operation 513, or the model was indicated not frozen in the operation 505 or 507 and frozen in this operation 513. In some other embodiments, the status of the model indicated in this operation 513 is consistent with (or the same as) the status of the model indicated in the operation 505 or 507. In some embodiments, the notification message sent to the type-1 PSF 510 further indicates that the model update for the at least one model is finished.
[0232] In an operation 515, an application location, which corresponds to the second type-2 PSF 550, is selected from one or multiple application locations, and the data plane 530 is accordingly configured so that data traffic in the operation 519 will be routed to the second type-2 PSF 550.
[0233] This operation 515 can be performed through a procedure described in association to the Fig. 6, wherein the type-1 PSF A 610 is the type-1 PSF 510 and the type-2 PSF B 640 is the second type-2 PSF 550. In this operation 515, the PSC 560 requests to influence traffic routing for the application. The request from the PSC 560 is transported (operation 603 in Fig. 6) to the PCF 630. The request includes information about traffic, information about type-1 PSFs and information about application location. The information about traffic identifies the data traffic in the operation 519. The information about type-1 PSFs identifies the type-1 PSF A 610. The information about application location identifies the one or multiple potential application locations.
[0234] In this operation 515, the PSC 560 may inform the second type-2 PSF 550 which models, among the at least one model, that the second type-2 PSF 550 can expect to receive in the data traffic, e.g. by providing a list of model ID (s) , name (s) or indices that identify the model (s) . For a model, among the at least one model, that the second type-2 PSF 550 will not receive in the data traffic, if the second type-2 PSF 550 does not have the latest version of the model, in this operation 515, the PSC 560 may configure the second type-2 PSF 550 to obtain the latest version of the model. The latest version of the model is stored at a fourth type-2 PSF (which may be the same as or different from the second type-2 PSF 550 or the third type-2 PSF) . The PSC 560 provides the second type-2 PSF 550 with model identifying information (e.g. an ID, a name or an index) , version information (e.g. a version number) and information about the forth type-2 PSF (e.g. an ID or a network address) , e.g. when configuring the type-2 PSF B 640 in the operation 609 (b) or 619 in FIG. 6. The model identifying information identifies the model, and the version information identifies the latest version of the model. According to the information provided from the PSC 560, the second type-2 PSF 550 may interact with the fourth type-2 PSF to obtain the latest version of the model. In some examples, the first type-2 PSF 540 and the second type-2 PSF 550 may be the same type-2 PSF, and the PSC 560 may transmit a notification (an example of the second notification above) to the first type-2 PSF 540 to notify that the first type-2 PSF 540 is selected for receiving at least one latest version of one or more models from the type-1 PSF 510, in which the at least one latest version is obtained through the model updating performed by the type-1 PSF 510.
[0235] In an operation 517, the PSC 560 transmit a control command to the type-1 PSF 510, specifically, the PSC 560 notifies the type-1 PSF 510 to send latest version of the at least one model, by sending a message (referred to as a notification message, an example of the third message above) to the type-1 PSF 510.
[0236] In some embodiments, the notification message may be integrated with a response to the operation 513. If the notification message from the type-1 PSF 510 to the PSC 560 in the operation 513 is sent via a control plane function (e.g. the NWC or the AMF) , the notification message in this operation 517 is sent to the type-1 PSF 510 via the control plane function. That is, the PSC 560 sends the notification message to the control plane function, and the control plane function then forwards / sends the notification message to the type-1 PSF 510. In some embodiments, the PSC 560 notifies the type-1 PSF 510 to send latest version of some of the at least one model, and the notification message may include information (e.g. a list of mode ID (s) , names (s) or indices) identifying the some of the at least one model.
[0237] In an operation 519, the type-1 PSF 510 sends updated models to the second type-2 PSF 550, specifically, the type-1 PSF includes the latest version of the at least one model, or the some of the at least one model as identified in the operation 519, in a data traffic, and sends the traffic to the second type-2 PSF 550 through the data plane 530. A data traffic targets a type-2 PSF.
[0238] For each of these model (s) , the type-1 PSF 510 includes the following information in the traffic: a model identifying information (e.g. an ID, a name or an index) , a version information (e.g. a version number) , and values of model parameters of the model. The model identifying information identifies the model. The version information is associated to the model and corresponds to the values of model parameters of the model, and it indicates the freshness of the model. The type-1 PSF 510 may include the latest version of its own local model in the traffic. The type-1 PSF 510 may further includes a status indication for its own local model in the traffic, indicating whether its own local model is frozen or not (. e. frozen or unfrozen) . In some embodiments, the type-1 PSF 510 is configured to include the indication in the traffic only when the instruction in the operation 503 is configured to be integrated with the operations 505 and 509. The data plane 530 transports the traffic to the second type-2 PSF 550 such that the second type-2 PSF 550 receives the latest version of the at least one model or the some of the at least one model.
[0239] In an operation 521, the second type-2 PSF 550 sends a notification to the PSC 560. The notification may indicate that the PSC 560 has received the models as indicated in the operation 515. The notification may further include a status indication for certain model, indicating that the certain model is frozen or not. The certain model is the local model of the type-1 PSF 510, and the indication is consistent with (or the same as) the indication described in the operation 519.
[0240] FIG. 6 illustrates an example procedure of request for influencing traffic routing according to some embodiments of the present disclosure. As shown in FIG. 6, in the procedure 600, at least one type-1 PSF (a type-1 PSF A 610 is shown as an example) , a NWC 520, a PMF 620, a PCF 630, a data plane 530, at least one type-2 PSF (a type-2 PSF B 640 is shown as an example) , and a PSC 560 are involved. The PSC 560 may request to influence traffic routing for the application, using a procedure illustrated in the FIG. 6. The traffic routing is influenced so that traffic of the application, e.g. the traffic in the operation 509 or 519 of FIG. 5, is transported between the type-1 PSF A 610 (e.g. the type-1 PSF 510 in Fig. 5) and the type-2 PSF B 640 (e.g. the first type-2 PSF 540 or the second type-2 PSF 550 in FIG. 5) along a data plane path. The type-1 PSF A 610, the type-2 PSF B 640 and the data plane path are dynamically selected according to information in the request of PSC 560. The type-1 PSF A 610 is selected from a group of type-1 PSFs identified in the request. The type-2 PSF B 640 corresponds to an application location and is selected as a result of selecting the application location from one or multiple potential application locations identified in the request. The data plane path includes one or multiple DPFs and connects the type-1 PSF A 610 and the type-2 PSF B 640. The dynamic selection described above can take into account network status (e.g. throughput, congestion, and delay) and PSF status, i.e. status of the type-1 PSF A 610 and the status of the type-2 PSF (such as loading, remaining energy, computing resource availability or sufficiency) to balance load and optimize overall performance.
[0241] The request from the PSC 560 is transported to the PCF 630, as illustrated by the operation 603 in FIG. 6. In some embodiments, the request from the PSC 560 is sent to the PCF 630, directly or via the NEF. In some embodiments, the request from the PSC 560 is sent to a NSF (the NSF is not shown in FIG. 6) , directly or via the NEF. The NSF stores the request as application data. In this case, the PCF 630 subscribes to receive application data related to the application from the NSF, and the NSF provides the request to the PCF 630 according to the subscription. The PCF 630 may perform the subscription when the PCF 630 receives request (e.g. from the NWC 520) for policies related to the application.
[0242] When the NEF is involved as described above during the transport of the request from the PSC 560 to the PCF 630 (i.e. the operation 603 in FIG. 6) , the NEF receives the request from the PSC 560 and sends the request to the next network entity (i.e. the PCF 630 or the NSF) . The NEF processes the request before sending the request to the next network entity. When processing the request, the NEF may perform information mapping, wherein the NEF replaces / updates some information in the request with mapped information so that the next network entity receives the request with the mapped information. For example, the NEF may map a service identifier in the request to a DNN or a combination of a DNN and network slice information, and / or an external ID to an internal ID.
[0243] The PCF 630 transforms the request to one or multiple policies (in other words, generates one or multiple policies based on the request) , as illustrated by the operation 605 in FIG. 6, and sends the one or multiple policies to the NWC 520, as illustrated by the operation 607 in FIG. 6. According to the one or multiple policies, the NWC 520 makes traffic routing decisions (e.g. the operation 609 in FIG. 6) and interacts (e.g. the operations 611, 615 in FIG. 6) with the PMF 620 to implement the traffic routing decisions. When interacting with the PMF 620, the NWC 520 provides traffic routing information to the PMF 620, the traffic routing information specifying the traffic routing decisions. In some embodiments, the NWC 520 and the PMF 620 are integrated, as a single network entity. In some embodiments, the NWC 520 and the PCF 630 are integrated, as a single network entity.
[0244] The traffic routing decisions made by the NWC 520 includes a type-1 PSF selection decision (i.e. selecting the type-1 PSF A 610) and / or an application location selection decision (i.e. selecting the application location corresponding to the type-2 PSF B 640) . In some embodiments, the traffic routine decisions made by the NWC 520 include a data plane path (re) selection decision. In some embodiments, the traffic routine decisions made by the NWC 520 does not include the data plane path (re) selection decision, and the data plane path (re) selection decision is instead made by the PMF 620 according to the traffic routing decisions (i.e. the type-1 PSF selection decision and / or the type-2 PSF selection decision) made by the NWC 520. When implementing the traffic routing decisions, the PMF 620 configures (e.g. the operation 613 in FIG. 6) the data plane path so that traffic is transported between the type-1 PSF A 610 and the type-2 PSF B 640 along the data plane path. More details of the procedure 600 may refer to operations 601 to 621 below.
[0245] In an operation 601, the PSC 560 selects one or multiple potential locations of the application (in other words, potential application locations) , each of which corresponds to a type-2 PSF. The type-2 PSF (s) corresponding to one or multiple potential application locations can be used to support the application. When selecting the one or multiple potential application locations, the PSC 560 may consider loading of the corresponding type-2 PSF (s) .
[0246] In an operation 603, The PSC 560 requests to influence traffic routing for the application. The request from the PSC 560 is transported to the PCF 630. In some examples, the request is transmitted in a way of message, for example, the fifth message mentioned in the process 200-1 or 200-2 above. In some examples, the request may include information about traffic (i.e. an example of the information of data traffic identifying data traffic of the at least one application (which in some embodiments may be associated with the second one or more apparatuses) , as comprised in the fifth message in the process 200-1 or 200-2) , information about type-1 PSFs (i.e. an example of the information of the first one or more apparatuses, as comprised in the fifth message in the process 200-1 or 200-2) , information about application location (i.e. an example of the information of at least one location of the at least one application (the at least one application in some embodiments may be associated with the second apparatus) , as comprised in the fifth message in the process 200-1 or 200-2) , and information about traffic routing requirements (i.e. an example of the information of one or more requirements associated with the traffic routing, as comprised in the fifth message in the process 200-1 or 200-2) .
[0247] In some examples, the information about traffic identifies the traffic of the application. The traffic identified in this information may include traffic to be routed to (or targeting) a type-1 PSF (referred to as downlink traffic) and traffic to be routed to (or targeting) a type-2 PSF (referred to as uplink traffic) . In some examples, this information may include a service identifier, or a DNN, or a combination of a DNN and slicing information (e.g. S-NSSAI) . The service identifier corresponds to the DNN or the combination of the DNN and the slicing information. Additionally, this information may further include traffic filtering information or, when the traffic filtering information is pre-configured (e.g. at DPFs) , an application identifier corresponding to the traffic filtering information. The application identifier may also identify the application. The traffic filtering information may be expressed using any combination of the following: a destination address, a source address, a source port, a destination port, a transport protocol (e.g. a protocol ID or name) and will be used by a DPF to detect the traffic. In some examples, the information about type-1 PSFs may comprise information identifying the group of type-1 PSFs, e.g. a group ID or a list of IDs or network address, each identifying a type-1 PSF. The information about application location may describe the potential application location (s) , which are selected in the operation 601. In this information, each of the potential application location (s) can be expressed or identified by an ID or a network address, and in some embodiments, the ID or the network address is associated to a type-2 PSF corresponding to the potential application location. The information about traffic routing requirements describes traffic routing requirements. This information may indicate that an application location should be selected from the potential application location (s) . This information may indicate that a type-1 PSF should be selected from the group of type-1 PSFs. This information may further indicate that traffic routing should be performed between the application location and the type-1 PSF for the traffic as identified in the information about traffic, wherein uplink traffic (if any) is routed to the application location and downlink traffic (if any) is routed to the type-1 PSF.
[0248] In an operation 605, the PCF 630 generates one or multiple policies based on the request received in the operation 601. The one or multiple policies include the information about traffic, the information about application location and the information about traffic routing requirements. The one or multiple policies may further include the information about type-1 PSFs. The details of the information comprised in the one or multiple policies may refer to the same information comprised in the request described at the operation 603.
[0249] In an operation 607, the PCF 630 sends the one or multiple policies (an example of the at least one policy 275) to the NWC 520, which makes traffic routing decisions for the traffic of the application. In some embodiments, the NWC 520 is associated to the service identifier or to the DNN or the combination of the DNN and the slicing information and is selected using the service identifier or using the service DNN or the combination of the DNN and the slicing information. In some embodiments, the NWC 520 is associated to the application and is selected using the application identifier.
[0250] In an operation 609, the NWC 520 makes traffic routing decisions, wherein the type-1 PSF A 610 and the type-2 PSF B 640 are selected, according to the one or multiple policies. The type-2 PSF B 640 is selected by selecting an application location corresponding to the type-2 PSF B 640. This operation 609 may include the following sub operations 609 (a) -609 (c) .
[0251] In an operation 609 (a) , according to the one or multiple policies, the NWC 520 selects the application location, which corresponds to the type-2 PSF B 640, and one or multiple type-1 PSF (s) . The one or multiple type-1 PSF (s) includes the type-1 PSF A 610. In some examples, the application location may be selected from the potential application location (s) according to the traffic routing requirements. The one or multiple type-1 PSFs are selected, with respected to the application location, from the group of type-1 PSFs according to the traffic routing requirements. In some embodiments, when selecting the application location and the type-1 PSF (s) , the NWC 520 may consider to optimize data plane 530 efficiency.
[0252] In an operation 609 (b) , the NWC 520 sends a notification (e.g. in the form of a message) to the PSC 560. The notification (i.e. the notification message) may be about a data plane management event and considered as an early notification (i.e. a notification sent before the data plane management event happens) . The data plane management event may be about (re) selection of an application location (or application location change) . In this case, the notification includes information identifying the application location, e.g. an ID or a network address, and indicates that the application location is selected. The data plane management event may also be about (re) selection of type-1 PSF (s) . In this case, the notification further includes information identifying the one or multiple type-1 PSFs, e.g. a list of ID (s) or network address (s) . In some embodiments, the PSC 560 may send an acknowledgement (ACK) , e.g. in the form of a message, to the NWC 520, acknowledging the recipient of the notification. The ACK may be a positive ACK for confirming (agreeing on) the selection of the application location, or a negative ACK for rejecting (disagreeing on) the selection of the application location. The ACK includes information indicating whether the acknowledgement is a positive ACK or a negative ACK. When the ACK is a positive ACK, before sending the ACK, the PSC 560 may configure the type-2 PSF B 640 corresponding to the application location, for example, obtain the latest version of one or multiple models as described in the operation 501 in Fig. 5. In some examples, the NWC 520 may be integrated in the PCF 630, then a message (an example of the sixth message 285 above) may be transmitted from the PCF 630 to the PSC 560, in which the sixth message 285 indicates the type-2 PSF B 240 (an example of the second apparatus 220) and the one or multiple type-1 PSFs (an example of the at least one candidate apparatus, specifically, for example, the sixth message 285 may be about (re) selection of an application location (or application location change) , and about (re) selection of type-1 PSF (s) , by indicating the data plane management event.
[0253] In this operation 609 (b) , when multiple type-1 PSFs are identified in the notification, the PSC 560 may make a further selection (i.e. a down select) , wherein a single type-1 PSF (i.e. the type-1 PSF A 610) is selected from the multiple type-1 PSFs. In this case, the positive ACK (i.e. the ACK message) sent to the NWC 520 includes information identifying the type-1 PSF A 610, e.g. an ID or a network address. If the NWC 520 receives a negative ACK in this operation, the NWC 520 may repeat the operation 609 so that a different type-2 PSF can be selected and / or that different one or multiple type-1 PSFs can be selected.
[0254] In an operation 609 (c) , if the ACK received from the PSC 560 is a positive ACK, the NWC 520 notifies the type-1 PSF A 610 that the type-1 PSF A 610 is selected by sending a message (referred to as notification message, an example of the second message above) to the type-1 PSF A 610. The notification message includes information identifying the application. According to the notification, the type-1 PSF A 610 knows that it will receive data traffic related to the application, and is to perform the model updating. In some examples, the type-1 PSF A 610 may send an acknowledgement (ACK) , e.g. in the form of a message, to the NWC. The ACK may be a positive ACK for confirming (agreeing on) the selection of the type-1 PSF A 610, or a negative ACK for rejecting (disagreeing on) the selection of the type-1 PSF A. In some examples, the ACK (i.e. the ACK message) may include information indicating whether the ACK is a positive ACK or a negative ACK. If the NWC 520 receives a negative ACK in this operation 609 (c) , the NWC 520 may repeat the operation 609 so that a different type-1 PSF can be selected.
[0255] In some embodiments, the type-1 PSF A 610 may be an AS, the NWC 520 may send the notification (i.e. the notification message) to an AF (not shown in FIG. 6) that subscribes to the notification. The notification may be transported to the AF via the NEF (not shown in FIG. 6) . The NEF may refer to FIG. 4B. After receiving the notification, the AF may in turn notify the AS so that the AS receives the information in the notification sent from the NWC 520.
[0256] In an operation 611, the NWC 520 provides traffic routing information to the PMF 620 for the application. The traffic routing information includes information identifying the application location (e.g. an ID or a network address) and information identifying the type-1 PSF A 610 (e.g. an ID or a network address) . The traffic routing information may further include the information about traffic, which was received by the NWC 520 from the PCF 630 in the operation 607. The traffic routing information indicates that traffic routing should be performed between the application location and the type-1 PSF for the traffic as identified in the information about traffic, wherein uplink traffic (if any) is routed to the application location and downlink traffic (if any) is routed to the type-1 PSF.
[0257] In some embodiments, the NWC 520 performs a data plane path (re) selection for the type-1 PSF A 610 (e.g. to improve data plane 530 efficiency) , wherein the NWC 520 selects or reselects a data plane path according to the one or multiple policies received from the PCF 630. The data plane path may include one or multiple DPFs and connects the type-1 PSF A 610 and the application location. The one or multiple DPFs may be (re) selected as part of the data plane path with respect to the application location. The NWC 520 may include information about data plane path in the traffic routing information sent to the PMF 620. The information about data plane path specifies the data plane path, e.g. by comprising information identifying the one or multiple DPFs (e.g. a list of ID (s) or network address (es) ) . When the traffic routing information includes the information about data plane path (i.e. specifies the data plane path) , it implies that the traffic routing should be performed along the data plane path as described in the information about data plane path. In some embodiments, the NWC 520 does not perform the data plane path (re) selection, and the PMF 620 is configured to perform the data plane path (re) selection as described in the operation 613. In this case, the traffic routing information sent to the PMF 620 does not include the information about data plane path.
[0258] In an operation 613, the PMF 620 performs a data plane path (re) selection for the type-1 PSF A 610 (e.g. to improve data plane 530 efficiency) , wherein the PMF 620 selects or reselects a data plane path according to the traffic routing information received from the NWC 520. The data plane path includes one or multiple DPFs and connects the type-1 PSF A 610 and the application location. The one or multiple DPFs may be (re) selected as part of the data plane path with respect to an application location. The application location is identified in the traffic routing information.
[0259] In some embodiments, this operation 613 is optional, for example, when the NWC 520 is configured to perform the data plane path (re) selection or when the traffic routing information, as described in the operation 611, specifies the data plane path. When the data plane path (re) selection is performed by the NWC 520, the traffic routing information includes the information about data plane path, and the information about data plane path specifies the data plane path.
[0260] In an operation 615, the PMF 620 configures the data plane path to support or implement the traffic routing. The data plane path may be (re) selected by the NWC 520 or by the PMF 620, as described above, and includes one or multiple DPFs. When configuring the data plane path, the PMF 620 configures the one or multiple DPFs to establish communication tunnel (s) between them, and the PMF 620 may further provide one or more DPFs, among the one or multiple DPFs, with traffic handling rules. The traffic handling rules are generated by the PMF 620 according to the traffic routing requirements received from the NWC 520 in the operation 611. The one or more DPFs will according to the traffic handling rules to detect and route the traffic. When routing the traffic, the one or multiple DPFs will route the traffic along the data plane path through the communication tunnel (s) .
[0261] In an operation 617, the PMF 620 notifies the NWC 520 that the traffic routing has been configured in the data plane 530, in other words, the data plane 530 has been configured to support or implement the traffic routing. This operation 617 may be performed when the PMF 620 responds to the NWC 520 to acknowledge the recipient of the traffic routing information, that is, the notification may be integrated with the response. The notification may be an example of the third notification above.
[0262] In an operation 619, the NWC 520 sends a notification to the PSC 560. The notification is about a data plane management event and considered as a late notification (i.e. a notification sent after the data plane management event happens) . In some examples, the data plane management event is related to (re) selection of an application location (or application location change) . In this case, the notification includes information identifying the application location, e.g. an ID or a network address, and indicates that the application location is selected. In some examples, the data plane management event may also be related to (re) selection of type-1 PSF (s) . In this case, the notification may further include information identifying the type-1 PSF A 610, e.g. an ID or a network address.
[0263] The PSC 560 may send an acknowledgement (ACK) , e.g. in the form of a message, to the NWC 520, acknowledging the recipient of the notification. The ACK may be a positive ACK for confirming (agreeing on) the selection of the application location, or a negative ACK for rejecting (disagreeing on) the selection of the application location. The ACK may include information indicating whether the acknowledgement is a positive ACK or a negative ACK. When the ACK is a positive ACK, before sending the ACK, the PSC 560 may configure the type-2 PSF B 640 corresponding to the application location, for example, obtain the latest version of one or multiple models as described in the operation 501 in Fig. 5.
[0264] After the operation 619, in an operation 621, the traffic of the application is transported between the type-1 PSF A 610 and the type-2 PSF B 640 via the data plane, more specifically, the data plane path configured in the operation 613.
[0265] FIG. 7 illustrates a flowchart of an example method 700 implemented at a first apparatus according to some embodiments of the present disclosure. As shown in FIG. 7, at block 710, the first apparatus may receive, an indication for indicating the first apparatus to perform model updating of one or more models associated with first one or more apparatuses. The first apparatus is one of the first one or more apparatuses. At block 720, the first apparatus may transmit a first notification for notifying that the model updating of the one or more models is completed. In some embodiments, the first apparatus for performing the method 700 may be the first apparatus 221 in the process 200-1 and / or the first apparatus in the process 200-2. The operations in the method 700 performed by the first apparatus may further refer to the embodiments as mentioned in the process 200-1 and / or the process 200-2 above.
[0266] FIG. 8 illustrates a flowchart of an example method 800 implemented at a second apparatus (e.g. the second apparatus 220 above) according to some embodiments of the present disclosure. As shown in FIG. 8, at block 810, the second apparatus may receive, from a third apparatus, a fourth message for indicating whether additional training data should be used by the second apparatus to perform model updating of one or more models associated with first one or more apparatuses. The second apparatus is one of second one or more apparatuses, and the additional training data is different from training data for performing model updating at the first one or more apparatuses. At block 820, the second apparatus may update the one or more models based on receiving the fourth message. In some embodiments, the second apparatus for performing the method 800 may be the second apparatus 220 in the process 200-1 and / or the second apparatus in the process 200-2. The operations performed by the second apparatus in the method 800 may further refer to the embodiments as mentioned in the process 200-1 and / or the process 200-2 above.
[0267] FIG. 9 illustrates a flowchart of an example method 900 implemented at a third apparatus (e.g. the third apparatus 230 or 231 above) according to some embodiments of the present disclosure. As shown in FIG. 9, at block 910, the third apparatus may transmit, to a second apparatus, a fourth message for indicating whether additional training data should be used by the second apparatus to perform model updating of one or more models associated with first one or more apparatuses, in which the second apparatus is one of second one or more apparatuses, and the additional training data is different from training data for performing model updating at the first one or more apparatuses.
[0268] FIG. 10 illustrates a flowchart of an example method 1000 implemented at a third apparatus (e.g. the third apparatus 230 or 231 above) according to some other embodiments of the present disclosure. As shown in FIG. 10, at block 1010, the third apparatus may transmit, to a fourth apparatus, a fifth message for requesting to influence traffic routing. Influencing the traffic routing is for selecting a second apparatus among second one or more apparatuses and at least one candidate apparatus among first one or more apparatuses. The second apparatus and the at least one candidate apparatus jointly perform model updating of one or more models associated with the first one or more apparatuses. The fourth apparatus is configured to perform a policy control function. At block 1020, the third apparatus may receive, from a fifth apparatus, a sixth message for indicating the second apparatus and the at least one candidate apparatus, wherein the fifth apparatus is configured to be a network controller.
[0269] In some embodiments, the third apparatus for performing the method 900 and / or the method 1000 may be the third apparatus 230 in the process 200-1 and / or the third apparatus 231 in the process 200-2. The operations performed by the third apparatus in the method 900 and / or the method 1000 may further refer to the embodiments as mentioned in the process 200-1 and / or the process 200-2 above.
[0270] FIG. 11 illustrates a flowchart of an example method 1100 implemented at a fourth apparatus (e.g. the fourth apparatus 240 above) according to some embodiments of the present disclosure. As shown in FIG. 11, at block 1110, the fourth apparatus may receive, from a third apparatus, a fifth message for requesting to influence traffic routing. Influencing the traffic routing is for selecting a second apparatus among second one or more apparatuses and at least one candidate apparatus among first one or more apparatuses. The second apparatus and the at least one candidate apparatus jointly perform model updating of one or more models associated with the first one or more apparatuses. The fourth apparatus is configured to perform a policy control function. At block 1120, the fourth apparatus may generate, based on the fifth message, at least one policy on influencing the traffic routing. In some embodiments, the fourth apparatus for performing the method 1100 may be the fourth apparatus in the process 200-1 and / or the fourth apparatus 240 in the process 200-2. The operations performed by the fourth apparatus in the method 1100 may further refer to the embodiments as mentioned in the process 200-1 and / or the process 200-2 above.
[0271] FIG. 12 illustrates a flowchart of an example method 1200 implemented at a fifth apparatus (e.g. the fifth apparatus above) according to some embodiments of the present disclosure. As shown in FIG. 12, at block 1210, the fifth apparatus may receive, from a fourth apparatus, at least one policy on influencing traffic routing. Influencing the traffic routing is for selecting a second apparatus among second one or more apparatuses and at least one candidate apparatus among first one or more apparatuses. The second apparatus and the at least one candidate apparatus jointly perform model updating of one or more models associated with the first one or more apparatuses. The fourth apparatus is configured to perform a policy control function, and the fifth apparatus is configured to be a network controller. At block 1220, the fifth apparatus may select, based on the at least one policy, the second apparatus and the at least one candidate apparatus. At block 1230, the fifth apparatus may transmit, to a third apparatus, a sixth message for indicating the second apparatus and the at least one candidate apparatus. The third apparatus is configured to control the second one or more apparatuses. In some embodiments, the fifth apparatus for performing the method 1200 may be the fifth apparatus in the process 200-1 and / or the fifth apparatus 250 in the process 200-2. The operations performed by the fifth apparatus in the method 1200 may further refer to the embodiments as mentioned in the process 200-1 and / or the process 200-2 above.
[0272] FIG. 13 illustrates a flowchart of an example method 1300 implemented at a system according to some embodiments of the present disclosure. The example method 1300 may comprise one or more operations of the first apparatus, the second apparatus, the third apparatus, the fourth apparatus and the fifth apparatus as mentioned above. For example, at block 1310, the third apparatus may transmit a fourth message for indicating whether additional training data should be used by a second apparatus to perform model updating of one or more models associated with first one or more apparatuses, wherein the second apparatus is one of second one or more apparatuses, and the additional training data is different from training data for performing model updating at the first one or more apparatuses. At block 1320, the second apparatus may receive, from the third apparatus, the fourth message, and updating the one or more models based on receiving the fourth message. At block 1330, the first apparatus may receive, from the second apparatus or the third apparatus, an indication for indicating the first apparatus to perform the model updating, and transmitting a first notification for notifying that the model updating of the one or more models is completed, wherein the first apparatus is one of the first one or more apparatuses. At block 1340, the third apparatus may control the second one or more apparatuses, and transmit, to a fourth apparatus, a fifth message for requesting to influence traffic routing, wherein influencing the traffic routing is for selecting the second apparatus and at least one candidate apparatus among the first one or more apparatuses, and wherein the second apparatus and the at least one candidate apparatus jointly perform the model updating. At block 1350, the fourth apparatus may receive, the fifth message from the third apparatus, to generate, based on the fifth message, at least one policy on influencing the traffic routing, and transmitting the at least one policy to a fifth apparatus. At block 1360, the fifth apparatus may receive, the at least one policy from the fourth apparatus, and the fifth apparatus is further configured to be a network controller and configured to select, based on the at least one policy, the second apparatus and the at least one candidate apparatus, and transmit to the third apparatus, a sixth message for indicating the second apparatus and the at least one candidate apparatus.
[0273] FIG. 14 is a simplified block diagram of a device 1400 that is suitable for implementing some embodiments of the present disclosure. The device 1400 can be considered as a further example embodiment of the first apparatus, the second apparatus, the third apparatus, the fourth apparatus or the fifth apparatus in the above embodiments. Accordingly, the device 1400 can be implemented at or as at least a part of the above apparatuses.
[0274] As shown, the device 1400 includes a processor 1410, a memory 1420 coupled to the processor 1410, a suitable transmitter (TX) and receiver (RX) 1440 coupled to the processor 1410, and a communication interface coupled to the TX / RX 1440. The memory 1420 stores at least a part of a program 1430. The TX / RX 1440 is for bidirectional communications. The TX / RX 1440 has at least one antenna to facilitate communication, though in practice an Access Node mentioned in this application may have several ones. The communication interface may represent any interface that is necessary for communication with other network elements, such as X2 interface for bidirectional communications between gNBs or eNBs, S1 interface for communication between a Mobility Management Entity (MME) / Serving Gateway (S-GW) and the gNB or eNB, Un interface for communication between the gNB or eNB and a relay node (RN) , or Uu interface for communication between the gNB or eNB and a terminal device.
[0275] The program 1430 is assumed to include program instructions that, when executed by the associated processor 1410, enable the device 1400 to operate in accordance with the embodiments of the present disclosure, as discussed herein with reference to FIGS. 1A-13. The embodiments herein may be implemented by computer software executable by the processor 1410 of the device 1400, or by hardware, or by a combination of software and hardware. The processor 1410 may be configured to implement various embodiments of the present disclosure. Furthermore, a combination of the processor 1410 and memory 1420 may form processing means 1450 adapted to implement various embodiments of the present disclosure.
[0276] The memory 1420 may be of any type suitable to the local technical network and may be implemented using any suitable data storage technology, such as a non-transitory computer readable storage medium, semiconductor based memory devices, magnetic memory devices and systems, optical memory devices and systems, fixed memory and removable memory, as non-limiting examples. While one memory 1420 is shown in the device 1400, there may be several physically distinct memory modules in the device 1400. The processor 1410 may be of any type suitable to the local technical network, and may include one or more of general purpose computers, special purpose computers, microprocessors, digital signal processors (DSPs) and processors based on multicore processor architecture, as non-limiting examples. The device 1400 may have multiple processors, such as an application specific integrated circuit chip that is slaved in time to a clock which synchronizes the main processor.
[0277] In some embodiments, a first apparatus comprises circuitry configured to perform method 700. In some embodiments, a second apparatus comprises circuitry configured to perform method 800. In some embodiments, a third apparatus comprises circuitry configured to perform method 900 or 1000. In some embodiments, a fourth apparatus comprises circuitry configured to perform method 1100. In some embodiments, a fifth apparatus comprises circuitry configured to perform method 1200.
[0278] In some embodiments, an example system in the present disclosure comprises the first apparatus, the second apparatus, the third apparatus, the fourth apparatus and the fifth apparatus as mentioned above. For example, in the example system, the third apparatus is configured to transmit a fourth message for indicating whether additional training data should be used by the second apparatus to perform model updating of one or more models associated with first one or more apparatuses, wherein the second apparatus is one of second one or more apparatuses, and the additional training data is different from training data for performing model updating at the first one or more apparatuses; the second apparatus is configured to receive, from the third apparatus, the fourth message, and update the one or more models based on receiving the fourth message; the first apparatus is configured to receive, from the second apparatus or the third apparatus, an indication for indicating the first apparatus to perform the model updating, and transmit a first notification for notifying that the model updating of the one or more models is completed, wherein the first apparatus is one of the first one or more apparatuses; the third apparatus is further configured to control the second one or more apparatuses, and transmit, to the fourth apparatus, a fifth message for requesting to influence traffic routing, wherein influencing the traffic routing is for selecting the second apparatus and at least one candidate apparatus among the first one or more apparatuses, and wherein the second apparatus and the at least one candidate apparatus jointly perform the model updating; the fourth apparatus is configured to receive the fifth message from the third apparatus, to generate, based on the fifth message, at least one policy on influencing the traffic routing, and transmit the at least one policy to the fifth apparatus; and the fifth apparatus is configured to be a network controller and to receive the at least one policy from the fourth apparatus, and the fifth apparatus is further configured to select, based on the at least one policy, the second apparatus and the at least one candidate apparatus, and transmit to the third apparatus, a sixth message for indicating the second apparatus and the at least one candidate apparatus.
[0279] The components included in the apparatuses and / or devices of some embodiments of the present disclosure may be implemented in various manners, including software, hardware, firmware, or any combination thereof. In one embodiment, one or more units may be implemented using software and / or firmware, for example, machine-executable instructions stored on the storage medium. In addition to or instead of machine-executable instructions, parts or all of the units in the apparatuses and / or devices may be implemented, at least in part, by one or more hardware logic components. For example, and without limitation, illustrative types of hardware logic components that can be used include Field-programmable Gate Arrays (FPGAs) , Application-specific Integrated Circuits (ASICs) , Application-specific Standard Products (ASSPs) , System-on-a-chip systems (SOCs) , Complex Programmable Logic Devices (CPLDs) , and the like.
[0280] Generally, various embodiments of the present disclosure may be implemented in hardware or special purpose circuits, software, logic or any combination thereof. Some aspects may be implemented in hardware, while other aspects may be implemented in firmware or software which may be executed by a controller, microprocessor or other computing device. While various aspects of embodiments of the present disclosure are illustrated and described as block diagrams, flowcharts, or using some other pictorial representation, it will be appreciated that the blocks, apparatus, systems, technique terminal devices or methods described herein may be implemented in, as non-limiting examples, hardware, software, firmware, special purpose circuits or logic, general purpose hardware or controller or other computing devices, or some combination thereof.
[0281] Some embodiments of the present disclosure also provide at least one computer program product tangibly stored on a non-transitory computer readable storage medium. The computer program product includes computer-executable instructions, such as those included in program modules, being executed in a device on a target real or virtual processor, to carry out the process or method as described above with reference to any of FIGS. 3 to 12. Generally, program modules include routines, programs, libraries, objects, classes, components, data structures, or the like that perform particular tasks or implement particular abstract data types. The functionality of the program modules may be combined or split between program modules as desired in various embodiments. Machine-executable instructions for program modules may be executed within a local or distributed device. In a distributed device, program modules may be located in both local and remote storage media.
[0282] Program code for carrying out methods of some embodiments of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing apparatus, such that the program codes, when executed by the processor or controller, cause the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may execute entirely on a machine, partly on the machine, as a stand-alone software package, partly on the machine and partly on a remote machine or entirely on the remote machine or server.
[0283] The above program code may be embodied on a machine readable medium, which may be any tangible medium that may contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. The machine readable medium may be a machine readable signal medium or a machine readable storage medium. A machine readable medium may include but not limited to an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of the machine readable storage medium would include an electrical connection having one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM) , a read-only memory (ROM) , an erasable programmable read-only memory (EPROM or Flash memory) , an optical fiber, a portable compact disc read-only memory (CD-ROM) , an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
[0284] Further, while operations are depicted in a particular order, this should not be understood as requiring that such operations be performed in the particular order shown or in sequential order, or that all illustrated operations be performed, to achieve desirable results. In certain circumstances, multitasking and parallel processing may be advantageous. Likewise, while several specific embodiment details are contained in the above discussions, these should not be construed as limitations on the scope of the present disclosure, but rather as descriptions of features that may be specific to particular embodiments. Certain features that are described in the context of separate embodiments may also be implemented in combination in a single embodiment. Conversely, various features that are described in the context of a single embodiment may also be implemented in multiple embodiments separately or in any suitable sub-combination.
[0285] Although some embodiments of the present disclosure have been described in language specific to structural features and / or methodological acts, it is to be understood that embodiments of the present disclosure defined in the appended claims is not necessarily limited to the specific features or acts described above. Rather, the specific features and acts described above are disclosed as example forms of implementing the claims.
[0286] When the functions are implemented in the form of a software functional unit and sold or used as an independent product, the functions may be stored in a computer-readable storage medium. Based on such an understanding, the technical solutions of this application essentially, or the part contributing to the prior art, or some of the technical solutions may be implemented in a form of a software product. The software product is stored in a storage medium, and includes several instructions for instructing a computer device (which may be a personal computer, a server, or a network device) to perform all or some of the steps (or operations) of the methods described in the embodiments of this application. The foregoing storage medium includes: any medium that can store program code, such as a USB flash drive, a removable hard disk, a read-only memory (Read-Only Memory, ROM) , a random access memory (Random Access Memory, RAM) , a magnetic disk, or an optical disc.
[0287] The foregoing descriptions are specific implementations of this application, but are not intended to limit the protection scope of this application. Any variation or replacement readily figured out by a person skilled in the art within the technical scope disclosed in this application shall fall within the protection scope of this application. Therefore, the protection scope of this application shall be subject to the protection scope of the claims.
Claims
1.A method comprising:receiving, at a first apparatus, an indication for indicating the first apparatus to perform model updating of one or more models associated with first one or more apparatuses, wherein the first apparatus is one of the first one or more apparatuses; andtransmitting, at the first apparatus, a first notification for notifying that the model updating of the one or more models is completed.2.The method of claim 1, wherein the indication comprises at least one of the following:a list of model identifying information for identifying the one or more models;version information of the one or more models, wherein each of the one or more models has corresponding version information; orone or more status indications, wherein each of the one or more status indications indicates a status of a model among the one or more models, and the status of the model comprises frozen or unfrozen.3.The method of claim 1 or 2, wherein the indication is carried on a first message, the first message is from a second apparatus, and the first message further comprises values of model parameters of the one or more models.4.The method of claim 1 or 2, wherein the indication is carried on a first message, the first message is from a second apparatus, and the method further comprises:obtaining, at the first apparatus, values of model parameters of the one or more models, wherein the values of model parameters is pre-configured at the first apparatus.5.The method of claim 3 or 4, wherein:the second apparatus is a part of a data plane of a core network, orthe second apparatus is connected to a data plane of a core network.6.The method of any of claims 3-5, wherein:the indication is transmitted from the second apparatus and via a data plane function to the first apparatus.7.The method of claim 1 or 2, wherein the indication is from a third apparatus, and the method further comprises:receiving, at the first apparatus and from a second apparatus, values of model parameters of the one or more models.8.The method of claim 7, wherein the third apparatus is a part of a control plane of the core network.9.The method of claim 7 or 8, wherein at least one of the following:the indication is transmitted from the third apparatus to the first apparatus; orthe indication is transmitted from the third apparatus and via a control plane function to the first apparatus.10.The method of any of claims 1-9, further comprising:updating, at the first apparatus, at least one model among the one or more models.11.The method of any of claims 1-10, wherein the first notification further comprises at least one of the following:a list of model identifying information for identifying the one or more models; orone or more status indications, wherein each of the one or more status indications indicates a status of a model among the one or more models, and the status of the model comprises frozen or unfrozen.12.The method of claim 2 or 10, wherein:a frozen model is a converged model; oran unfrozen model is an unconverged model.13.The method of any of claims 1-12, wherein at least one of the following:the first notification is transmitted from the first apparatus to a third apparatus; orthe first notification is transmitted from the first apparatus and via a control plane function to the third apparatus.14.The method of any of claims 1-13, further comprising:receiving, at the first apparatus, a second message for notifying that the first apparatus is selected for performing the model updating.15.The method of claim 14, further comprising:transmitting, at the first apparatus, a positive acknowledgement (ACK) indicating confirmation of the selection.16.The method of claim 15, wherein:the second message is from a fourth apparatus, and the positive ACK is transmitted to the fourth apparatus, wherein the fourth apparatus is configured to perform a policy control function; orthe second message is from a fifth apparatus, and the positive ACK is transmitted to the fifth apparatus, wherein the fifth apparatus is a network controller.17.The method of any of claims 1-16, further comprising:receiving, at the first apparatus and from a third apparatus, a third message for notifying the first apparatus to transmit at least one latest version of the one or more models obtained through the model updating; andtransmitting, at the first apparatus, the at least one latest version.18.The method of claim 17, wherein at least one of the following:the third message is received from the third apparatus via a control plane function; orthe at least one latest version is transmitted via a data plane function, wherein the data plane function is configured to transport the at least one latest version to a sixth apparatus among second one or more apparatuses.19.The method of claim 18, wherein the sixth apparatus is the second apparatus.20.The method of any of claims 9, 13 and 18, wherein the control plane function comprises at least one of the following: a network exposure function (NEF) , a policy control function (PCF) , a network storage function (NSF) , a network controller (NWC) , a path management function (PMF) or an access and mobility management function (AMF) .21.The method of any of claims 17-19, wherein the third message further comprises:model identifying information for identifying at least one model associated with the at least one latest version.22.The method of any of claims 1-21, wherein the first apparatus is one of a terminal apparatus, a sever or a network function.23.A method comprising:receiving, at a second apparatus and from a third apparatus, a fourth message for indicating whether additional training data should be used by the second apparatus to perform model updating of one or more models associated with first one or more apparatuses, wherein the second apparatus is one of second one or more apparatuses, and the additional training data is different from training data for performing model updating at the first one or more apparatuses; andupdating, at the second apparatus, the one or more models based on receiving the fourth message.24.The method of claim 23, wherein the fourth message comprises the following:a list of model identifying information for identifying the one or more models;version information of the one or more models, wherein each of the one or more models has corresponding version information; andone or more status indications, wherein each of the one or more status indications indicates a status of a model among the one or more models, and the status of the model comprises frozen or unfrozen.25.The method of claim 24, wherein the fourth message further indicates where to obtain the additional training data and / or how to select the additional training data in the event that the additional training data should be used.26.The method of any of claims 23-25, wherein updating the one or more models comprises:in the event of the additional training data should be used by the second apparatus to perform the model updating, using the additional training data to update at least one model with status indication indicating unfrozen among the one or more models.27.The method of any of claims 23-26, further comprising:transmitting, at the second apparatus and to a first apparatus among the first one or more apparatuses, an indication for indicating the first apparatus to perform the model updating of the one or more models.28.The method of claim 27, wherein the indication is carried on a first message, and the first message further comprises values of model parameters of the one or more models.29.The method of claim 27 or 28, wherein the indication comprises the following:a list of model identifying information for identifying the one or more models;version information of the one or more models, wherein each of the one or more models has corresponding version information; andone or more status indications, wherein each of the one or more status indications indicates a status of a model among the one or more models, and the status of the model comprises frozen or unfrozen.30.The method of claim 24 or 29, wherein:a frozen model is a converged model; oran unfrozen model is an unconverged model.31.The method of claim 27, further comprising:receiving, at the second apparatus and from the third apparatus, a second notification that the second apparatus is selected for receiving at least one latest version of the one or more models from the first apparatus, wherein the at least one latest version is obtained through the model updating performed by the first apparatus.32.The method of claim 31, wherein the second notification further comprises:model identifying information for identifying at least one model associated with the at least one latest version.33.The method of claim 31 or 32, further comprising:receiving, at the second apparatus and from the third apparatus, a configuration for the second apparatus to obtain one or more latest versions from a further apparatus among the second one or more apparatuses, wherein the one or more latest versions are among a plurality of versions of the one or more models excluding the at least one latest version.34.The method of claim 27 or 31, wherein:the first apparatus is one of a terminal apparatus, a sever or a network function.35.The method of any of claims 23-34, wherein:the second apparatus is a part of a data plane of a core network; orthe second apparatus is connected to a data plane of a core network.36.The method of any of claims 23-35, wherein:the third apparatus is a part of a control plane of the core network.37.A method comprising:transmitting, at a third apparatus and to a second apparatus, a fourth message for indicating whether additional training data should be used by the second apparatus to perform model updating of one or more models associated with first one or more apparatuses, wherein the second apparatus is one of second one or more apparatuses, and the additional training data is different from training data for performing model updating at the first one or more apparatuses.38.The method of claim 37, further comprising:determining, at the third apparatus, one of the first one or more apparatuses as a first apparatus for performing the model updating.39.The method of claim 37 or 38, further comprising:determining, at the third apparatus, the second apparatus by selecting an application location associated with the second apparatus among at least one application location.40.The method of any of claims 37-39, further comprising:transmitting, at the third apparatus and to the second apparatus, a configuration for the second apparatus to obtain version information of the one or more models from a further apparatus among the second one or more apparatuses.41.The method of any of claims 37-40, further comprising:transmitting, at the third apparatus and to a first apparatus among the first one or more apparatuses, an indication for indicating the first apparatus to perform the model updating.42.The method of claim 41, wherein the indication comprises at least one of the following:a list of model identifying information for identifying the one or more models;version information of the one or more models, wherein each of the one or more models has corresponding version information; orone or more status indications, wherein each of the one or more status indications indicates a status of a model among the one or more models, and the status of the model comprises frozen or unfrozen.43.The method of claim 41 or 42, wherein:the indication is transmitted from the third apparatus to the first apparatus, orthe indication is transmitted from the third apparatus and via a control plane function to the first apparatus44.The method of any of claims 37-43, wherein the fourth message comprises at least one of the following:a list of model identifying information for identifying the one or more models;version information of the one or more models, wherein each of the one or more models has corresponding version information; orone or more status indications, wherein each of the one or more status indications indicates a status of a model among the one or more models, and the status of the model comprises frozen or unfrozen.45.The method of claim 42 or 44, wherein:a frozen model is a converged model; oran unfrozen model is an unconverged model.46.The method of claim 44, wherein the fourth message further indicates where to obtain the additional training data and / or how to select the additional training data in the event that the additional training data should be used.47.The method of any of claims 37-46, further comprising:receiving, at the third apparatus and from a first apparatus among the first one or more apparatuses, a first notification for notifying that the model updating of one or more models performed by the first apparatus is completed; anddetermining, based on the first notification, at the third apparatus and from the second one or more apparatuses, a sixth apparatus for receiving at least one latest version of the one or more models, wherein the at least one latest version is obtained through model updating performed by the first apparatus, and a latest version corresponds to an updated model.48.The method of claim 47, further comprising:transmitting, at the third apparatus and to the sixth apparatus, a second notification that the sixth apparatus is selected for receiving the at least one latest version from the first apparatus.49.The method of claim 48, wherein the second notification further comprises:model identifying information for identifying at least one model associated with the at least one latest version.50.The method of claim 48 or 49, further comprising:transmitting, at the third apparatus and to the sixth apparatus, a configuration for the sixth apparatus to obtain one or more latest versions from a further apparatus among the second one or more apparatuses, wherein the one or more latest versions are among a plurality of versions of the one or more models excluding the at least one latest version.51.The method of claim 44, further comprising:transmitting, at the third apparatus and to a first apparatus among the first one or more apparatuses, a third message for notifying the first apparatus to transmit at least one latest version of the one or more models obtained through the model updating to a sixth apparatus among the second one or more apparatuses.51.The method of any of claims 47-51, wherein the sixth apparatus is the second apparatus.52.The method of any of claims 38, 41-43, 47-48 and 51, wherein:the first apparatus is one of a terminal apparatus, a sever or a network function.53.The method of any of claims 37-52, wherein:the second apparatus is a part of a data plane of a core network; orthe second apparatus is connected to a data plane of a core network.54.The method of any of claims 37-53, wherein:the third apparatus is a part of a control plane of the core network.55.A method comprising:transmitting, at a third apparatus and to a fourth apparatus, a fifth message for requesting to influence traffic routing, wherein influencing the traffic routing is for selecting a second apparatus among second one or more apparatuses and at least one candidate apparatus among first one or more apparatuses, the second apparatus and the at least one candidate apparatus jointly perform model updating of one or more models associated with the first one or more apparatuses, and the fourth apparatus is configured to perform a policy control function; andreceiving, at the third apparatus and from a fifth apparatus, a sixth message for indicating the second apparatus and the at least one candidate apparatus, wherein the fifth apparatus is configured to be a network controller.56.The method of claim 55, further comprising:determining, at the third apparatus, the second apparatus by selecting an application location associated with the second apparatus among at least one application location.57.The method of claim 56, wherein the fifth message comprises at least one of the following:information of data traffic identifying data traffic of the at least one application associated with the second one or more apparatuses, wherein the data traffic is to be routed during the traffic routing;information of the first one or more apparatuses;information of at least one location of the at least one application;information of one or more requirements associated with the traffic routing; orinformation of traffic filtering associated with the data traffic.58.The method of claim 56, further comprising:transmitting, at the third apparatus and to the fifth apparatus, feedback for the sixth message, wherein the feedback indicates a positive acknowledgement (ACK) or a negative ACK to the selection for the application location associated with the second apparatus.59.The method of claim 58, further comprising:determining, at the third apparatus, a candidate apparatus as a first apparatus among the at least one candidate apparatus, wherein information identifying the first apparatus is included in the positive ACK.60.The method of claim 59, wherein:the first apparatus is one of a terminal apparatus, a sever or a network function.61.The method of any of claims 55-60, wherein:the second apparatus is a part of a data plane of a core network; orthe second apparatus is connected to a data plane of a core network.62.The method of any of claims 55-61, wherein:the third apparatus is a part of a control plane of the core network.63.A method comprising:receiving, at a fourth apparatus and from a third apparatus, a fifth message for requesting to influence traffic routing, wherein influencing the traffic routing is for selecting a second apparatus among second one or more apparatuses and at least one candidate apparatus among first one or more apparatuses, the second apparatus and the at least one candidate apparatus jointly perform model updating of one or more models associated with the first one or more apparatuses, and the fourth apparatus is configured to perform a policy control function; andgenerating, at the fourth apparatus and based on the fifth message, at least one policy on influencing the traffic routing.64.The method of claim 63, further comprising at least one of the following:transmitting, at the fourth apparatus, the at least one policy to a fifth apparatus configured to be a network controller for determining a second apparatus and the at least one candidate apparatus; ordetermining, at the fourth apparatus and based on the at least one policy, the second apparatus and the at least one candidate apparatus.65.The method of claim 64, wherein the at least one policy comprises at least one of the following:information of data traffic to be routed in the traffic routing;information of at least one location of at least one application, wherein one of the at least one application is associated with the second apparatus;information of one or more requirements associated with the traffic routing; orinformation of the first one or more apparatuses.66.The method of claim 63, further comprising:transmitting, at the fourth apparatus and to the third apparatus, a sixth message for indicating a second apparatus and the at least one candidate apparatus.67.The method of claim 66, wherein the sixth message indicates a data plane management event associated with a selection of an application location associated with the second apparatus among at least one application location.68.The method of claim 67, further comprising:receiving, at the fourth apparatus and from the third apparatus, a feedback for the sixth message, wherein the feedback indicates a positive acknowledgement (ACK) for confirmation of the selection or a negative ACK for rejection of the selection.69.The method of claim 68, wherein the feedback is the positive ACK, and the positive ACK further comprises:information identifying a candidate apparatus as a first apparatus among the at least one candidate apparatus.70.The method of claim 69, further comprising:transmitting, at the fourth apparatus and to the first apparatus, a second message for notifying that the first apparatus is selected to perform the model updating.71.The method of claim 70, further comprising:receiving, at the fourth apparatus and from the first apparatus, a positive ACK for confirmation of performing the model updating or a negative ACK for rejection of performing the model updating.72.The method of any of claims 66-70, further comprising:transmitting, at the fourth apparatus and to a seventh apparatus for selecting a data plane path for the traffic routing, traffic routing information indicating the traffic routing.73.The method of claim 70, further comprising at least one of the following:receiving, at the fourth apparatus and from the seventh apparatus, a third notification that a data plane path has been configured for the traffic routing; orreceiving, at the fourth apparatus and from the seventh apparatus, a response to the transmission of the traffic routing information, wherein the response comprises the third notification that the data plane path has been configured for the traffic routing.74.The method of claim 72 or 73, wherein the traffic routing information comprises the at least one of the following:information identifying an application location associated with the second apparatus among at least one application location;information identifying a first apparatus, wherein the first apparatus is determined among the at least one candidate apparatus;information associated with the at least one policy; orinformation of a data plane path for the traffic routing.75.The method of any of claims 69-71 and 74, wherein:the first apparatus is one of a terminal apparatus, a sever or a network function.76.The method of any of claims 63-75, wherein:one of the second one or more apparatuses is a part of a data plane of a core network; orone of the second one or more apparatuses is connected to a data plane of a core network.77.The method of any of claims 63-76, wherein:the third apparatus is a part of a control plane of the core network.78.A method comprising:receiving, at a fifth apparatus and from a fourth apparatus, at least one policy on influencing traffic routing, wherein influencing the traffic routing is for selecting a second apparatus among second one or more apparatuses and at least one candidate apparatus among first one or more apparatuses, the second apparatus and the at least one candidate apparatus jointly perform model updating of one or more models associated with the first one or more apparatuses, the fourth apparatus is configured to perform a policy control function, and the fifth apparatus is configured to be a network controller;selecting, at the fifth apparatus and based on the at least one policy, the second apparatus and the at least one candidate apparatus; andtransmitting, at the fifth apparatus and to a third apparatus, a sixth message for indicating the second apparatus and the at least one candidate apparatus, wherein the third apparatus is configured to control the second one or more apparatuses.79.The method of claim 78, wherein the at least one policy comprises at least one of the following:information of data traffic to be routed in the traffic routing;information of at least one location of at least one application, wherein one of the at least one application is associated with the second apparatus;information of one or more requirements associated with the traffic routing; orinformation of the first one or more apparatuses.80.The method of claim 78 or 79, wherein the sixth message indicates a data plane management event associated with a selection of an application location associated with the second apparatus among at least one application location.81.The method of claim 80, further comprising:receiving, at a fifth apparatus and from the third apparatus, feedback for the sixth message, wherein the feedback indicates a positive acknowledgement (ACK) for confirmation of the selection or a negative ACK for rejection of the selection.82.The method of claim 79, wherein the feedback is the positive ACK, and the positive ACK further comprises:information identifying a candidate apparatus as a first apparatus among the at least one candidate apparatus.83.The method of claim 82, further comprising:transmitting, at a fifth apparatus and to the first apparatus, a second message for notifying that the first apparatus is selected to perform the model updating.84.The method of claim 83, further comprising:receiving, at a fifth apparatus and from the first apparatus, a positive ACK for confirmation of performing the model updating or a negative ACK for rejection of performing the model updating.85.The method of any of claims 78-84, further comprising:transmitting, at a fifth apparatus and to a seventh apparatus for selecting a data plane path for the traffic routing, traffic routing information indicating the traffic routing.86.The method of claim 84, wherein the traffic routing information comprises the at least one of the following:information identifying an application location associated with the second apparatus among at least one application location;information identifying a first apparatus, wherein the first apparatus is determined among the at least one candidate apparatus;information associated with the at least one policy;information of a data plane path for the traffic routing.87.The method of claim 85 or 86, further comprising at least one of the following:receiving, at a fifth apparatus and from the seventh apparatus, a third notification that a data plane path has been configured for the traffic routing; orreceiving, at a fifth apparatus and from the seventh apparatus, a response to the transmission of the traffic routing information, wherein the response comprises the third notification that the data plane path has been configured for the traffic routing.88.The method of any of claims 78-87, wherein the fifth apparatus and the fourth apparatus are integrated in a single network entity.89.The method of any of claims 82-84 and 86, wherein:first apparatus is one of a terminal apparatus, a sever or a network function.90.The method of any of claims 78-89, wherein:one of the second one or more apparatuses is a part of a data plane of a core network; orone of the second one or more apparatuses is connected to a data plane of a core network.91.The method of any of claims 78-90, wherein:the third apparatus is a part of a control plane of the core network.92.A first apparatus comprising:a transceiver; anda processor communicatively coupled with the transceiver, wherein the processor is configured to perform the method of any of claims 1-22.93.A second apparatus comprising:a transceiver; anda processor communicatively coupled with the transceiver, wherein the processor is configured to perform the method of any of claims 23-36.94.A third apparatus comprising:a transceiver; anda processor communicatively coupled with the transceiver, wherein the processor is configured to perform the method of any of claims 37-54.95.A third apparatus comprising:a transceiver; anda processor communicatively coupled with the transceiver, wherein the processor is configured to perform the method of any of claims 55-62.96.A fourth apparatus comprising:a transceiver; anda processor communicatively coupled with the transceiver, wherein the processor is configured to perform the method of any of claims 63-77.97.A fifth apparatus comprising:a transceiver; anda processor communicatively coupled with the transceiver, wherein the processor is configured to perform the method of any of claims 78-91.98.A system comprising a first apparatus, a second apparatus, a third apparatus, a fourth apparatus and a fifth apparatus, wherein:the third apparatus is configured to transmit a fourth message for indicating whether additional training data should be used by the second apparatus to perform model updating of one or more models associated with first one or more apparatuses, wherein the second apparatus is one of second one or more apparatuses, and the additional training data is different from training data for performing model updating at the first one or more apparatuses;the second apparatus is configured to receive, from the third apparatus, the fourth message, and update the one or more models based on receiving the fourth message;the first apparatus is configured to receive, from the second apparatus or the third apparatus, an indication for indicating the first apparatus to perform the model updating, and transmit a first notification for notifying that the model updating of the one or more models is completed, wherein the first apparatus is one of the first one or more apparatuses;the third apparatus is further configured to control the second one or more apparatuses, and transmit, to the fourth apparatus, a fifth message for requesting to influence traffic routing, wherein influencing the traffic routing is for selecting the second apparatus and at least one candidate apparatus among the first one or more apparatuses, and wherein the second apparatus and the at least one candidate apparatus jointly perform the model updating;the fourth apparatus is configured to receive the fifth message from the third apparatus, to generate, based on the fifth message, at least one policy on influencing the traffic routing, and transmit the at least one policy to the fifth apparatus; andthe fifth apparatus is configured to be a network controller and to receive the at least one policy from the fourth apparatus, and the fifth apparatus is further configured to select, based on the at least one policy, the second apparatus and the at least one candidate apparatus, and transmit to the third apparatus, a sixth message for indicating the second apparatus and the at least one candidate apparatus.99.A method comprising:transmitting, at a third apparatus, a fourth message for indicating whether additional training data should be used by a second apparatus to perform model updating of one or more models associated with first one or more apparatuses, wherein the second apparatus is one of second one or more apparatuses, and the additional training data is different from training data for performing model updating at the first one or more apparatuses;receiving, at the second apparatus and from the third apparatus, the fourth message, and updating the one or more models based on receiving the fourth message;receiving, at a first apparatus and from the second apparatus or the third apparatus, an indication for indicating the first apparatus to perform the model updating, and transmitting a first notification for notifying that the model updating of the one or more models is completed, wherein the first apparatus is one of the first one or more apparatuses;controlling, at the third apparatus, the second one or more apparatuses, and transmitting, to a fourth apparatus, a fifth message for requesting to influence traffic routing, wherein influencing the traffic routing is for selecting the second apparatus and at least one candidate apparatus among the first one or more apparatuses, and wherein the second apparatus and the at least one candidate apparatus jointly perform the model updating;receiving at the fourth apparatus, the fifth message from the third apparatus, to generate, based on the fifth message, at least one policy on influencing the traffic routing, and transmitting the at least one policy to a fifth apparatus; andreceiving, at the fifth apparatus, the at least one policy from the fourth apparatus, and the fifth apparatus is further configured to be a network controller and configured to select, based on the at least one policy, the second apparatus and the at least one candidate apparatus, and transmit to the third apparatus, a sixth message for indicating the second apparatus and the at least one candidate apparatus.100.A system comprising at least one of the following: the first apparatus of claim 92, the second apparatus of claim 93, the third apparatus of claim 94 or 95, the fourth apparatus of claim 96, or the fifth apparatus of claim 97.101.A non-transitory computer readable medium comprising computer program stored thereon, the computer program, when executed on at least one processor, causing the at least one processor to perform the method of any of claims 1-91.102.A chip comprising at least one processing circuit configured to perform the method of any of claims 1-91.103.A computer program product tangibly stored on a computer-readable medium and comprising computer-executable instructions which, when executed, cause an apparatus to perform the method of any of claims 1-91.
Citation Information
Patent Citations
Model processing method and device, electronic equipment and storage medium
CN116432773A
Trusted and decentralized aggregation for joint learning
CN117242463A
System and method for facilitating logo-recognition training of a recognition model
US20180285696A1
Machine learning model for allocation of resources
US20210398016A1