Information processing method, model generation and training method, electronic device, and medium

Abstract

The application discloses an information processing method, a model generation and training method, an electronic device and a medium. The information processing method comprises the following steps: acquiring a transmission parameter of service information in a network, wherein the transmission parameter represents transmission performance of the service information in the network; and performing transmission performance prediction on the transmission parameter by using a pre-trained network model, wherein the network model comprises a GCN model, and the GCN model is obtained according to transmission path information of the service information in the network. In the embodiment of the application, since the GCN model is obtained according to the transmission path information of the service information in the network, the GCN model can be consistent with the network topology for transmitting the service information, so that the transmission performance prediction of the service information in the network can be more accurately and effectively realized by processing the transmission parameter of the service information in the network by using the network model comprising the GCN model.

Description

Technical Field

[0001] The embodiments of the present invention relate to, but are not limited to, the field of network technology, and particularly to an information processing method, a method for generating network models, a method for training network models, an electronic device, and a computer-readable storage medium. Background Technology

[0002] With the rapid development of networks, network topologies are becoming increasingly complex. To ensure the stability and high quality of service transmission within the network, it is necessary to predict service transmission performance. However, current predictions of service transmission performance mainly rely on the experience of operations and maintenance personnel, resulting in a low level of intelligence. Although machine learning algorithms have been proposed to address service transmission performance prediction, the factors influencing this prediction include time, space, and social environment. These factors are numerous and complex, making it difficult for simple machine learning algorithms to effectively predict transmission performance influenced by these multidimensional factors. Summary of the Invention

[0003] The following is an overview of the subject matter described in detail herein. This overview is not intended to limit the scope of the claims.

[0004] This invention provides an information processing method, a network model generation method, a network model training method, an electronic device, and a computer-readable storage medium, which can effectively predict the transmission performance of service information in a network.

[0005] In a first aspect, embodiments of the present invention provide an information processing method, comprising:

[0006] Obtain transmission parameters of service information in the network, wherein the transmission parameters characterize the transmission performance of the service information in the network;

[0007] The transmission parameters are predicted using a pre-trained network model, wherein the network model includes a Graph Convolution Network (GCN) model, which is obtained based on the transmission path information of service information in the network.

[0008] Secondly, embodiments of the present invention also provide a method for generating a network model, comprising:

[0009] Obtain the transmission path information of service information in the network;

[0010] A GCN model is generated based on the transmission path information, wherein the GCN model includes an input node for inputting transmission parameters of the service information, the input node corresponding to the transmission path information, and the transmission parameters characterizing the transmission performance of the service information in the network.

[0011] Thirdly, embodiments of the present invention also provide a method for training a network model, comprising:

[0012] Obtain transmission parameters of service information in the network, wherein the transmission parameters characterize the transmission performance of the service information in the network;

[0013] The network model is trained using the transmission parameters, wherein the network model includes a GCN model, which is obtained based on the transmission path information of service information in the network.

[0014] Fourthly, embodiments of the present invention also provide an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the information processing method of the first aspect as described above.

[0015] Fifthly, embodiments of the present invention also provide an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the network model generation method of the second aspect as described above.

[0016] In a sixth aspect, embodiments of the present invention also provide an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the training method of the network model described in the third aspect above.

[0017] In a seventh aspect, embodiments of the present invention also provide a computer-readable storage medium storing computer-executable instructions, the computer-executable instructions being used to execute the information processing method of the first aspect as described above, or to execute the network model generation method of the second aspect as described above, or to execute the network model training method of the third aspect as described above.

[0018] This invention includes: acquiring transmission parameters of service information in a network, wherein the transmission parameters characterize the transmission performance of the service information in the network; and predicting the transmission performance of the transmission parameters using a pre-trained network model, wherein the network model includes a Graph Convolutional Network (GCN) model, which is obtained based on the transmission path information of the service information in the network. According to the solution provided by this invention, since the GCN model is obtained based on the transmission path information of the service information in the network, the GCN model can conform to the network topology transmitting the service information. Therefore, by using a network model including the GCN model to process the transmission parameters of the service information in the network, the prediction of the transmission performance of the service information in the network can be achieved more accurately and effectively.

[0019] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the description, claims and drawings. Attached Figure Description

[0020] The accompanying drawings are provided to further understand the technical solutions of the present invention and constitute a part of the specification. They are used together with the embodiments of the present invention to explain the technical solutions of the present invention, and do not constitute a limitation on the technical solutions of the present invention.

[0021] Figure 1 This is a schematic diagram of a network topology for performing an information processing method according to an embodiment of the present invention;

[0022] Figure 2 This is a flowchart of an information processing method provided in one embodiment of the present invention;

[0023] Figure 3 This is a schematic diagram of the structure of a GCN model provided in one embodiment of the present invention;

[0024] Figure 4 This is a flowchart of a transmission performance prediction method provided in another embodiment of the present invention;

[0025] Figure 5 This is a schematic diagram of a prediction model system for predicting transmission performance based on transmission parameters, provided as an example of the present invention.

[0026] Figure 6 This is a flowchart of a method for generating a network model according to an embodiment of the present invention;

[0027] Figure 7 This is a flowchart of the method for generating a GCN model in another embodiment of the present invention;

[0028] Figure 8 This is a schematic diagram of a service network virtual topology provided in a specific example of the present invention;

[0029] Figure 9 This is a schematic diagram of node linking processing provided in a specific example of the present invention;

[0030] Figure 10 This is a schematic diagram of a network topology provided in a specific example of the present invention;

[0031] Figure 11 This is another specific example provided by the present invention. Figure 10 A schematic diagram of the GCN model constructed from the network topology;

[0032] Figure 12 This is a flowchart of a network model training method provided in one embodiment of the present invention;

[0033] Figure 13 This is a flowchart of the training method for a network model provided in another embodiment of the present invention. Detailed Implementation

[0034] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0035] It should be noted that although a logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than that shown in the flowchart. The terms "first," "second," etc., in the specification, claims, and the aforementioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence.

[0036] This invention provides an information processing method, a network model generation method, a network model training method, an electronic device, and a computer-readable storage medium. The information processing method includes: acquiring transmission parameters of service information in a network, wherein the transmission parameters characterize the transmission performance of the service information in the network; and predicting the transmission performance of the transmission parameters using a pre-trained network model, wherein the network model includes a Generative Network Path (GCN) model, which is obtained based on the transmission path information of the service information in the network. Since the GCN model is obtained based on the transmission path information of the service information in the network, it conforms to the network topology transmitting the service information. Therefore, by using a network model including the GCN model to process the transmission parameters of the service information in the network, the prediction of the transmission performance of the service information in the network can be achieved more accurately and effectively.

[0037] The embodiments of the present invention will be further described below with reference to the accompanying drawings.

[0038] like Figure 1 As shown, Figure 1 This is a schematic diagram of a network topology for performing an information processing method according to an embodiment of the present invention. Figure 1 In the example, the network topology includes a network controller (not shown in the figure) and multiple nodes (such as...). Figure 1 Nodes A, B, ..., S in the diagram are connected by links to form a structure like... Figure 1 The network topology is shown, and the network controller is connected to each node.

[0039] The network controller can acquire node information and service information reported by each node, thereby forming a network topology model. Based on this model, it can plan and configure transmission paths for service information. Once the network controller has completed the planning and configuration of these paths and obtained a forwarding policy, it distributes the policy to the corresponding nodes, enabling them to transmit service information according to the policy. For example, in... Figure 1 In the network topology shown, there are the following service transmission paths: the first service transmission path from node A to node L, the second service transmission path from node A to node P, the third service transmission path from node A to node Q, the fourth service transmission path from node B to node P, the fifth service transmission path from node B to node L, the sixth service transmission path from node B to node H, the seventh service transmission path from node L to node Q, the eighth service transmission path from node P to node Q, and the ninth service transmission path from node H to node Q.

[0040] In addition, when planning and configuring the transmission path of service information, the network controller can also predict the transmission performance of service information based on the node information and service information uploaded by each node, and can perform corresponding optimization processing on the prediction results.

[0041] Furthermore, in a distributed network system, each node can obtain node information announced by other nodes in the network, thereby obtaining network topology information based on this node information. For the first node sending service information, it can form a corresponding network topology model based on this network topology information, and plan and configure the transmission path of the service information according to the transmission performance requirements of the service information to be transmitted, and obtain a forwarding strategy that meets the requirements, and then forward the service information according to the forwarding strategy.

[0042] In addition, when planning and configuring the transmission path of the service information, the first node that sends the service information can also predict the transmission performance of the service information and perform corresponding optimization processing on the prediction results.

[0043] The network topologies and application scenarios described in the embodiments of this invention are intended to more clearly illustrate the technical solutions of the embodiments of this invention, and do not constitute a limitation on the technical solutions provided in the embodiments of this invention. As those skilled in the art will know, with the evolution of network topologies and the emergence of new application scenarios, the technical solutions provided in the embodiments of this invention are also applicable to similar technical problems.

[0044] It will be understood by those skilled in the art that Figure 1 The network topology shown does not constitute a limitation on the embodiments of the present invention. It may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0045] exist Figure 1 In the network topology shown, the network controller or the first node sending service information can call its stored information processing program to execute the information processing method.

[0046] Based on the above network topology, various embodiments of the present invention are proposed.

[0047] like Figure 2 As shown, Figure 2 This is a flowchart of an information processing method provided in an embodiment of the present invention. The information processing method is applied to a network controller or the first node that sends service information. The information processing method includes, but is not limited to, steps S110 and S120.

[0048] Step S110: Obtain the transmission parameters of service information in the network, wherein the transmission parameters characterize the transmission performance of service information in the network.

[0049] In one embodiment, transmission parameters characterize the transmission performance of service information in the network. These parameters include, but are not limited to, service traffic, optical signal-to-noise ratio (OSNR), bit error rate, and output optical power. It is understood that the transmission parameters of the service information can be obtained based on the node capability information of the first node, the node capability information of the last node, and the node capability information of the intermediate nodes traversed from the first node to the last node. Furthermore, the transmission parameters of the service information can be predicted or obtained from real-time detection in the network; this embodiment does not impose specific limitations on either. For example, when the transmission parameters of the service information are obtained through prediction processing, a prediction model constructed using neural networks or similar methods can be used to predict the node capability information of the first node, the node capability information of the last node, and the node capability information of the intermediate nodes traversed from the first node to the last node, thereby obtaining the transmission parameters of the service information.

[0050] Step S120: Use a pre-trained network model to predict the transmission performance of the transmission parameters. The network model includes a GCN model, which is obtained based on the transmission path information of the service information in the network.

[0051] It is understandable that, since the GCN model is derived from the transmission path information of service information in the network, the GCN model can conform to the network topology that transmits the service information. Therefore, by using a pre-trained network model that includes the GCN model to process the transmission parameters of service information, it is possible to accurately and effectively predict the transmission performance of service information in the network.

[0052] It is understandable that the network model can be pre-trained and stored in the network controller or in the first node sending service information. Therefore, after obtaining the transmission parameters of service information in the network, the network model can be used to predict the transmission performance of the service information. Since the network model predicts the transmission performance of service information based on its transmission parameters, the training data used to train the network model is also the transmission parameters of the service information. It is understood that there are multiple ways to obtain the training data used to train the network model, and this embodiment does not impose any specific limitations on this. For example, it can be obtained by acquiring the transmission parameters of the service information actually transmitted in the network, or it can be obtained through artificial simulation. When obtaining training data through artificial simulation, a virtual network topology can be constructed based on the actual network topology, and then the transmission of service information can be simulated in the virtual network topology to obtain the training data.

[0053] By employing the information processing method including the above steps S110 and S120, the transmission performance of the service information transmission parameters is predicted using a pre-trained network model including the GCN model. Since the GCN model is obtained based on the transmission path information of the service information in the network, the GCN model can conform to the network topology that transmits the service information. Therefore, by using the network model including the GCN model to process the transmission parameters of the service information in the network, the transmission performance prediction of the service information in the network can be achieved more accurately and effectively.

[0054] In one embodiment, since the GCN model is derived based on the transmission path information of service information in the network, therefore, as Figure 3 As shown, the GCN model can include input nodes for inputting transmission parameters. These input nodes correspond to transmission path information, and the input nodes corresponding to transmission path information that have the same source or same destination relationship are connected.

[0055] exist Figure 3 In the example, the GCN model includes input nodes AL, AP, AQ, BP, BL, BH, LQ, PQ, and HQ. Input node AL corresponds to... Figure 1 The first service transmission path in the network topology shown has an input node AP corresponding to, as follows: Figure 1 The second service transmission path in the network topology shown has an input node AQ corresponding to, for example: Figure 1 The third service transmission path in the network topology shown has an input node BP corresponding to, as follows: Figure 1 The fourth service transmission path in the network topology shown has an input node BL corresponding to, for example: Figure 1 The fifth service transmission path in the network topology shown has an input node BH corresponding to, for example: Figure 1 The sixth service transmission path in the network topology shown has an input node LQ corresponding to, as follows: Figure 1 The seventh service transmission path in the network topology shown has an input node PQ corresponding to, as follows: Figure 1 The eighth service transmission path in the network topology shown has an input node HQ corresponding to, as follows: Figure 1The ninth service transmission path in the network topology shown. The first, second, and third service transmission paths share the same source node (node ​​A), meaning they have a common-source relationship. Therefore, input nodes AL, AP, and AQ are connected in pairs. Similarly, input nodes BP, BL, and BH are connected in pairs; input nodes AQ, LQ, PQ, and HQ are connected in pairs; input nodes AL, BL, and LQ are connected in pairs; input nodes AP, BP, and PQ are connected in pairs; and input nodes BH and HQ are connected.

[0056] It is understandable that, since transmission parameters characterize the transmission performance of service information in the network, that is, the transmission parameters characterize the transmission performance of service information when it is transmitted in the corresponding transmission path in the network, by using the GCN model obtained above based on the transmission path information of the service information to process the transmission parameters of the service information, the transmission performance attributes (i.e., transmission parameters) based on the service information can be converted into parameters that meet the needs of the GCN model to be input to each input node, thereby enabling the network model including the GCN model to predict the transmission performance of service information in the network.

[0057] In another embodiment, the network model may further include a deep neural network (DNN) model, the input of which is connected to the output of the GCN model. In this case, such as... Figure 4 As shown, step S120, which uses a pre-trained network model to predict transmission performance of transmission parameters, may include, but is not limited to, the following steps:

[0058] Step S121: Process the transmission parameters using the GCN model to obtain transmission characteristic parameters;

[0059] Step S122: Process the transmission feature parameters using a DNN model to obtain the transmission performance prediction results.

[0060] In one embodiment, transmission parameters of service information at different times can be obtained. Then, the GCN model is used to extract features from these transmission parameters according to the time and space dimensions to obtain transmission feature parameters. Next, these transmission feature parameters are used as input to the DNN model and transmitted to the DNN model so that the DNN model can process these transmission feature parameters to obtain transmission performance prediction results for future times of service information.

[0061] To illustrate with an example, in such cases... Figure 5In the prediction model system shown, the transmission parameters of the service information at each time point are first obtained. For example, the transmission parameters of the service information corresponding to the n input nodes at time t and time t+1 are obtained, i.e. Figure 5 The input parameters for the n input nodes in the input section shown are then input into, for example,... Figure 5 In the GCN model processing section shown, the GCN model can process these input parameters to obtain the transmission characteristic parameters corresponding to the next time step. Then, these transmission characteristic parameters are input into a system such as... Figure 5 In the DNN model processing section shown, the DNN model can process the transmission feature parameters to obtain the transmission performance prediction result of the service information at the next time step. Therefore, the transmission performance prediction result can be used as guidance information for optimization processing to optimize the transmission performance of the service information and improve the transmission performance of the service information.

[0062] It is worth noting that using the GCN model to extract features from transmission parameters along the time dimension refers to using the GCN model to extract features from transmission parameters at different times. Using the GCN model to extract features from transmission parameters along the spatial dimension refers to using the GCN model to perform convolution processing on the transmission parameters to obtain spatially characteristic feature parameters.

[0063] In another embodiment, the GCN model can be characterized by the formula y = Θ * gx = σ(α(L)x), where y is the transmission feature parameter, that is, y is the output parameter of the GCN model after predicting the transmission parameter; x is the transmission parameter, that is, x is the input parameter to the GCN model; Θ is the convolution kernel function; *g is the graph convolution operator using the spectral method, representing the product of x and Θ; α is the convolution kernel parameter; L is the normalized graph Laplacian operator; and σ is the activation function, such as the ReLU activation function.

[0064] It is understandable that after obtaining the transmission parameters of service information in the network, these transmission parameters are used as input parameters of the GCN model and input into the formula y=Θ*gx=σ(α(L)x), then we can obtain:

[0065]

[0066] Where, y∈R n , representing the transmission feature parameters with temporal and spatial characteristics obtained by graph convolution of transmission parameters in the GCN model, where n is the number of service information transmissions in the network (i.e., the number of transmission paths for transmitting service information); x∈R M×n×kα represents the n transmission parameters input to the GCN model at M time points, where each transmission parameter is a k-dimensional transmission performance attribute; jt ∈R n , represents the convolution kernel parameter of the j-th dimension of the transmission performance attribute of n transmission parameters at time t; x jt ∈R n , represents the j-th dimension of the transmission performance attribute of the n transmission parameters input to the GCN model at time t.

[0067] In another embodiment, after step S122, the information processing method may further include, but is not limited to, the following steps:

[0068] The transmission performance prediction results are used to optimize the transmission performance of service information.

[0069] It is understandable that, since the GCN model is derived from the transmission path information of service information in the network, it can conform to the network topology for transmitting service information. Therefore, after using a network model including the GCN model to predict the transmission performance of service information in the network, and obtaining the transmission performance prediction result, this prediction result can be used as guidance for optimization processing to improve the transmission performance of service information. Different processing methods can be used to optimize the transmission performance of service information using the transmission performance prediction result. This embodiment does not impose specific limitations on this. For example, the performance parameters of service information in the corresponding transmission path can be adjusted according to the transmission performance prediction result to achieve the purpose of optimizing transmission performance, or the transmission path of service information in the network can be changed according to the transmission performance prediction result to achieve the purpose of optimizing transmission performance.

[0070] In addition, such as Figure 6 As shown, Figure 6 This is a flowchart of a network model generation method provided in another embodiment of the present invention. This network model generation method is applied to a network controller or the first node sending service information, and includes, but is not limited to, the following steps:

[0071] Step S210: Obtain the transmission path information of service information in the network;

[0072] Step S220: Generate a GCN model based on the transmission path information. The GCN model includes input nodes for inputting transmission parameters of service information. The input nodes correspond to the transmission path information, and the transmission parameters characterize the transmission performance of service information in the network.

[0073] It is understood that the GCN model generated in this embodiment based on the transmission path information is as follows: Figure 2 The GCN model shown in the embodiment is used to predict the transmission performance of transmission parameters for service information in the network.

[0074] It is understandable that the transmission path information of service information in a network can include head node information and tail node information. The head node information includes the address information of the head node that sends the service information, and the tail node information includes the address information of the tail node that receives the service information. When the head node needs to send service information to the tail node, it can obtain the address information of the head node and the tail node from the head node, thereby obtaining the transmission path information of the service information.

[0075] In one embodiment, transmission parameters characterize the transmission performance of service information in the network. These parameters include, but are not limited to, service traffic, OSNR, bit error rate, and output optical power. It is understood that the transmission parameters of the service information can be obtained based on the node capability information of the first node, the node capability information of the last node, and the node capability information of the intermediate nodes traversed from the first node to the last node. Furthermore, the transmission parameters of the service information can be predicted or obtained from real-time detection in the network; this embodiment does not impose specific limitations on either. For example, when the transmission parameters of the service information are obtained through prediction processing, a prediction model constructed using neural networks or similar methods can be used to predict the node capability information of the first node, the node capability information of the last node, and the node capability information of the intermediate nodes traversed from the first node to the last node, thereby obtaining the transmission parameters of the service information.

[0076] Once the transmission path information of service information in the network is obtained, a GCN model can be generated based on this transmission path information. This model includes input nodes containing transmission parameters for inputting service information. The input nodes of the GCN model correspond to the transmission path information. For example, ... Figure 3 The input node AL shown is as follows Figure 1 The first service transmission path shown corresponds to this. Since transmission parameters characterize the transmission performance of service information in the network, and the GCN model is derived from the transmission path information of service information in the network, using transmission parameters as input parameters for the input nodes of the GCN model can more accurately and effectively predict the transmission performance of service information in the network.

[0077] Additionally, in one embodiment, such as Figure 7 As shown, the step S220 of generating the GCN model based on the transmission path information may include, but is not limited to, the following steps:

[0078] Step S221: Establish the input node based on the first node information and the last node information;

[0079] Step S222: Connect the input nodes corresponding to the transmission path information that have the same source relationship or the same destination relationship to obtain the GCN model.

[0080] It is understandable that by establishing the input nodes of the GCN model based on the first and last node information of the transmitted service information, and connecting the input nodes corresponding to the transmission path information that have the same source or same destination relationship, the resulting GCN model can predict the transmission parameters of the service information in the corresponding transmission path, and can more accurately and effectively predict the transmission performance of the service information in the network.

[0081] The generation of GCN models will be explained in detail below with specific examples.

[0082] In such Figure 1 In the network topology shown, for the case of a single link between adjacent nodes, based on the first and last nodes of each service path, ignoring intermediate nodes, it is abstracted as follows: Figure 8 The illustrated service network virtual topology includes only the first and last nodes of each service path, with the first and last nodes directly connected to form a virtual service link for each path. This virtual topology will be used in conjunction with other virtual topologies. Figure 1 The network topology shown is abstracted as follows: Figure 8 After showing the virtual topology of the service network, according to... Figure 8 The out-degree of each node in the virtual topology of the business network shown, as well as the pairwise dependencies (such as source or destination relationships) of business virtual links sharing the same end node, can be abstracted and transformed for each node into... There are several virtual paths, where z represents the out-degree of the node. Figure 8 Taking node L as an example, it includes service virtual link AL, service virtual link BL, and service virtual link LQ sharing node L as the end node. Therefore, referring to... Figure 9 The diagram illustrates node linking, where node L can be abstracted and transformed into three virtual paths. Furthermore, in the case of... Figure 1 The network topology shown is abstracted as follows: Figure 8 After illustrating the virtual topology of the service network, each virtual service link in the virtual topology is abstracted and transformed into an input node of the GCN model. This is done by establishing input nodes based on the first and last node information; for example, the virtual service link from node A to node L is abstracted and transformed into input node AL. Then, as shown... Figure 8 After all service virtual links in the shown service network virtual topology are abstracted and converted into input nodes, according to... Figure 9The linking process shown connects the input nodes corresponding to virtual links of services that have the same origin or destination relationship, thus obtaining the following... Figure 3 The diagram shows a GCN model that uses a virtual node (i.e., the input node) to represent a service path. Since the GCN model is derived from the transmission path information of service information in the network, it conforms to the network topology for transmitting service information. Therefore, by using the GCN model to perform relevant prediction processing on the transmission parameters of service information in the network, it is possible to accurately and effectively predict the transmission performance of service information in the network.

[0083] It is worth noting that, for cases with multiple links between adjacent nodes, the network topology of multiple links between adjacent nodes can be transformed into a network topology of a single link between adjacent nodes by adding virtual dummy nodes, thus enabling the construction of the GCN model as described above. For example, suppose in a scenario like... Figure 10 In the network topology shown, there are three virtual service links between node A and node M. These three virtual service links can be abstracted and transformed into input nodes AM1, AM2, and AM3 by adding virtual dummy nodes. These input nodes are then combined with the input node AQ (abstracted from the virtual service link between node A and node Q), the input node QR (abstracted from the virtual service link between node Q and node R), and the input node MR (abstracted from the virtual service link between node M and node R), to construct a network topology as shown. Figure 11 The GCN model shown.

[0084] Furthermore, it is worth noting that if a business path needs to pass through intermediate nodes as electrical relays, the entire business path can be divided into multiple business path segments based on the first node, the last node, and the electrical relay nodes, and each business path segment is considered an independent business path and is virtualized.

[0085] In one embodiment, the GCN model generated based on the transmission path information of service information in the network can be characterized by the formula y = Θ * gx = σ(α(L)x), where y is the output parameter of the GCN model; x is the input parameter to the GCN model; Θ is the convolution kernel function; *g is the graph convolution operator using the spectral method, representing the product of x and Θ; α is the convolution kernel parameter; L is the normalized graph Laplacian operator; and σ is the activation function, such as the ReLU activation function.

[0086] It is understandable that once the transmission parameters of service information in the network are obtained, these transmission parameters can be used as input parameters to the GCN model and input into the formula y=Θ*gx=σ(α(L)x), then we can obtain:

[0087]

[0088] Where, y∈R n , representing the transmission feature parameters with temporal and spatial characteristics obtained by graph convolution of transmission parameters in the GCN model, where n is the number of service information transmissions in the network (i.e., the number of transmission paths for transmitting service information); x∈R M×n×k α represents the n transmission parameters input to the GCN model at M time points, where each transmission parameter is a k-dimensional transmission performance attribute; jt ∈R n , represents the convolution kernel parameter of the j-th dimension of the transmission performance attribute of n transmission parameters at time t; x jt ∈R n , represents the j-th dimension of the transmission performance attribute of the n transmission parameters input to the GCN model at time t.

[0089] In addition, such as Figure 12 As shown, Figure 12 This is a flowchart of a network model training method provided in another embodiment of the present invention. This network model training method is applied to a network controller or the first node sending service information, and includes, but is not limited to, the following steps:

[0090] Step S310: Obtain the transmission parameters of service information in the network, wherein the transmission parameters characterize the transmission performance of service information in the network;

[0091] Step S320: Train the network model using transmission parameters. The network model includes a GCN model, which is obtained based on the transmission path information of service information in the network.

[0092] It is understandable that the GCN model trained using the transmission parameters of service information in the network in this embodiment is as follows: Figure 2 The pre-trained GCN model in the illustrated embodiment is also as follows: Figure 6 The GCN model shown in the embodiment is generated based on transmission path information. By utilizing, for example... Figure 6 The generation method in the illustrated embodiment yields a GCN model, which is then trained using the training method described in this embodiment. Once training is complete, the resulting model is as follows: Figure 2 The pre-trained GCN model in the illustrated embodiment.

[0093] It is understandable that the transmission path information of service information in a network can include head node information and tail node information. The head node information includes the address information of the head node that sends the service information, and the tail node information includes the address information of the tail node that receives the service information. When the head node needs to send service information to the tail node, it can obtain the address information of the head node and the tail node from the head node, thereby obtaining the transmission path information of the service information.

[0094] In one embodiment, transmission parameters characterize the transmission performance of service information in the network. These parameters include, but are not limited to, service traffic, OSNR, bit error rate, and output optical power. It is understood that the transmission parameters of the service information can be obtained based on the node capability information of the first node, the node capability information of the last node, and the node capability information of the intermediate nodes traversed from the first node to the last node. Furthermore, the transmission parameters of the service information can be predicted or obtained from real-time detection in the network; this embodiment does not impose specific limitations on either. For example, when the transmission parameters of the service information are obtained through prediction processing, a prediction model constructed using neural networks or similar methods can be used to predict the node capability information of the first node, the node capability information of the last node, and the node capability information of the intermediate nodes traversed from the first node to the last node, thereby obtaining the transmission parameters of the service information.

[0095] Since the network model including the GCN model predicts transmission performance based on the transmission parameters of service information, the training data used to train the network model also consists of the transmission parameters of service information. It is understood that the training data used to train the network model can be obtained in various ways, and this embodiment does not impose any specific limitations on this. For example, it can be obtained by acquiring the transmission parameters of the service information actually transmitted in the network, or it can be obtained through artificial simulation. When obtaining training data through artificial simulation, a virtual network topology can be constructed based on the actual network topology, and then the transmission of service information can be simulated in the virtual network topology to obtain the training data.

[0096] By employing a training method including steps S310 and S320, the network model including the GCN model is trained using the transmission parameters of service information in the network. Since the GCN model is obtained based on the transmission path information of service information in the network, the GCN model can conform to the network topology that transmits the service information. Since the transmission parameters characterize the transmission performance of service information when it is transmitted in the corresponding transmission path in the network, by training the network model including the GCN model using the transmission parameters of service information, the trained network model including the GCN model can more accurately and effectively predict the transmission performance of service information in the network.

[0097] In one embodiment, the GCN model may include input nodes for inputting transmission parameters, which correspond to transmission path information and are connected to each other for transmission path information that has a common source or common destination relationship.

[0098] It should be noted that, since the GCN model in this embodiment is as follows... Figure 2 The GCN model shown in the embodiment is described below. Therefore, for a detailed explanation of the specific structure of the GCN model in this embodiment, please refer to the following: Figure 3 The detailed descriptions of the examples shown are not repeated here.

[0099] In another embodiment, the network model may further include a DNN model, the input of which is connected to the output of the GCN model. In this case, such as... Figure 13 As shown, training the network model using transmission parameters in step S320 may include, but is not limited to, the following steps:

[0100] Step S321: Input the transmission parameters to the input node to train the GCN model and obtain transmission feature samples;

[0101] Step S322: Input the transmitted feature samples into the DNN model to train the DNN model.

[0102] In one embodiment, transmission parameters of service information at different times can be obtained. Then, these transmission parameters are input to the input node to train the GCN model, enabling the GCN model to extract features from these transmission parameters according to the time and space dimensions to obtain transmission feature samples. Next, these transmission feature samples are input to the DNN model to train the DNN model, so that the trained GCN model and DNN model can accurately and effectively predict the transmission performance of service information in the network.

[0103] An example illustrates the process of training a network model using transmission parameters, such as... Figure 5 As shown, firstly, the transmission parameters of the service information at different times are obtained as initial training sample data. For example, the transmission parameters of the service information corresponding to n input nodes at times t and t+1 are obtained as training sample data. Then, these training sample data are input into a program such as... Figure 5 In the GCN model processing section shown, these training sample data are used to train the GCN model to obtain transmission feature samples. Then, these transmission feature samples are input into... Figure 5 In the DNN model processing section shown, the transmission feature samples are used to train the DNN model, so that the network model including the GCN model after training can more accurately and effectively predict the transmission performance of service information in the network.

[0104] It is worth noting that the GCN model extracts features from the transmission parameters along the time dimension, meaning that the GCN model performs feature extraction on the transmission parameters at different times. The GCN model also extracts features from the transmission parameters along the spatial dimension, meaning that the GCN model performs convolution processing on the transmission parameters to obtain spatially characteristic feature parameters.

[0105] In another embodiment, step S321, which involves inputting the transmission parameters to the input node to train the GCN model, may include, but is not limited to, the following steps:

[0106] The transmission parameters are input to the input node, and the GCN model is trained using the backpropagation algorithm.

[0107] It is understandable that training the GCN model using the backpropagation algorithm enables supervised learning, allowing the trained GCN model to more accurately predict the transmission performance of service information in the network. This, in turn, enables subsequent information processing operations to more accurately and effectively predict the transmission performance of service information in the network.

[0108] In another embodiment, step S322 may include, but is not limited to, the following steps:

[0109] The transmitted feature samples are input into the DNN model, and the DNN model is trained using the backpropagation algorithm.

[0110] It is understandable that using the backpropagation algorithm to train the DNN model enables supervised learning, allowing the trained DNN model to more accurately predict the transmission performance of service information in the network. This, in turn, enables subsequent information processing operations to more accurately and effectively predict the transmission performance of service information in the network.

[0111] In one embodiment, the GCN model obtained based on the transmission path information of service information in the network can be characterized by the formula y = Θ * gx = σ(α(L)x), where y is the output parameter of the GCN model; x is the input parameter to the GCN model; Θ is the convolution kernel function; *g is the graph convolution operator using the spectral method, representing the product of x and Θ; α is the convolution kernel parameter; L is the normalized graph Laplacian operator; and σ is the activation function, such as the ReLU activation function.

[0112] Understandably, once the transmission parameters of service information in the network are obtained, these parameters can be used as training sample data and input into the formula y = Θ * gx = σ(α(L)x) to train the GCN model. It should be noted that when the transmission parameters are input as training sample data into the formula y = Θ * gx = σ(α(L)x), the following can be obtained:

[0113]

[0114] Where, y∈R n , representing the transmission feature parameters with temporal and spatial characteristics obtained by graph convolution of transmission parameters in the GCN model, where n is the number of service information transmissions in the network (i.e., the number of transmission paths for transmitting service information); x∈R M×n×k α represents the n transmission parameters input to the GCN model at M time points, where each transmission parameter is a k-dimensional transmission performance attribute; jt ∈R n , represents the convolution kernel parameter of the j-th dimension of the transmission performance attribute of n transmission parameters at time t; x jt ∈R n , represents the j-th dimension of the transmission performance attribute of the n transmission parameters input to the GCN model at time t.

[0115] In addition, one embodiment of the present invention provides an electronic device, which includes: a memory, a processor, and a computer program stored in the memory and executable on the processor.

[0116] The processor and memory can be connected via a bus or other means.

[0117] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. Furthermore, memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory may optionally include memory remotely located relative to the processor, and these remote memories can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0118] It should be noted that the electronic device in this embodiment can be applied as follows: Figure 1 The network controller or the first node for sending service information in the illustrated embodiment, the electronic device in this embodiment, and such Figure 1 The network controllers or the first nodes used to send service information in the illustrated embodiments belong to the same inventive concept. Therefore, these embodiments have the same implementation principle and technical effect, which will not be described in detail here.

[0119] The non-transitory software program and instructions required to implement the information processing method of the above embodiments are stored in memory. When executed by a processor, the information processing method of the above embodiments is executed, for example, the method described above. Figure 2 Method steps S110 to S120, Figure 4 Method steps S121 to S122.

[0120] In addition, one embodiment of the present invention provides an electronic device, which includes: a memory, a processor, and a computer program stored in the memory and executable on the processor.

[0121] The processor and memory can be connected via a bus or other means.

[0122] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. Furthermore, memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory may optionally include memory remotely located relative to the processor, and these remote memories can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0123] It should be noted that the electronic device in this embodiment can be applied as follows: Figure 1 The network controller or the first node for sending service information in the illustrated embodiment, the electronic device in this embodiment, and such Figure 1 The network controllers or the first nodes used to send service information in the illustrated embodiments belong to the same inventive concept. Therefore, these embodiments have the same implementation principle and technical effect, which will not be described in detail here.

[0124] The non-transient software program and instructions required to implement the network model generation method of the above embodiments are stored in memory. When executed by a processor, the network model generation method of the above embodiments is executed, for example, the method described above is executed. Figure 6 Method steps S210 to S220, Figure 7 Method steps S221 to S222.

[0125] In addition, one embodiment of the present invention provides an electronic device, which includes: a memory, a processor, and a computer program stored in the memory and executable on the processor.

[0126] The processor and memory can be connected via a bus or other means.

[0127] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. Furthermore, memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory may optionally include memory remotely located relative to the processor, and these remote memories can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0128] It should be noted that the electronic device in this embodiment can be applied as follows: Figure 1 The network controller or the first node for sending service information in the illustrated embodiment, the electronic device in this embodiment, and such Figure 1 The network controllers or the first nodes used to send service information in the illustrated embodiments belong to the same inventive concept. Therefore, these embodiments have the same implementation principle and technical effect, which will not be described in detail here.

[0129] The non-transient software program and instructions required to implement the network model training method of the above embodiments are stored in memory. When executed by the processor, the network model training method of the above embodiments is executed, for example, the method described above. Figure 12 Method steps S310 to S320 in the text Figure 13 Method steps S321 to S322.

[0130] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0131] Furthermore, one embodiment of the present invention provides a computer-readable storage medium storing computer-executable instructions that are executed by a processor or controller, for example, by a processor in the above-described electronic device embodiment, causing the processor to perform the information processing methods described above, such as executing the methods described above. Figure 2 Method steps S110 to S120, Figure 4 Method steps S121 to S122; or, executed by a processor in the above-described electronic device embodiment, causing the processor to execute the network model generation method in the above-described embodiment, for example, executing the method described above. Figure 6 Method steps S210 to S220, Figure 7 Method steps S221 to S222; or, executed by a processor in the above-described electronic device embodiment, causing the processor to execute the network model training method in the above-described embodiment, for example, executing the above-described... Figure 12 Method steps S310 to S320 in the text Figure 13 Method steps S321 to S322.

[0132] It will be understood by those skilled in the art that all or some of the steps and systems in the methods disclosed above can be implemented as software, firmware, hardware, and suitable combinations thereof. Some or all of the physical components can be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit. Such software can be distributed on a computer-readable medium, which can include computer storage media (or non-transitory media) and communication media (or transient media). As is known to those skilled in the art, the term computer storage media includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disc (DVD) or other optical disc storage, magnetic cartridges, magnetic tape, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and is accessible to a computer. Furthermore, as is known to those skilled in the art, communication media typically contain computer-readable instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

[0133] The above is a detailed description of the preferred embodiments of the present invention. However, the present invention is not limited to the above embodiments. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of the present invention. All such equivalent modifications or substitutions are included within the scope defined by the claims of the present invention.

Claims

1. An information processing method, comprising: Obtain transmission parameters of service information in the network, wherein the transmission parameters characterize the transmission performance of the service information in the network; The transmission performance of the transmission parameters is predicted using a pre-trained network model, wherein the network model includes a graph convolutional network (GCN) model, which is obtained based on the transmission path information of the service information in the network, and the GCN model conforms to the network topology of the network transmitting the service information. The GCN model includes multiple input nodes for inputting the transmission parameters. These input nodes correspond to the transmission path information, and each of the multiple input nodes corresponds to one of the multiple service transmission paths in the network topology. The network topology includes multiple network nodes, and a service transmission path is formed between two network nodes. The input nodes corresponding to transmission path information with the same source relationship or the same destination relationship are connected. The same source relationship means that at least two of the multiple service transmission paths have the same source node, and the same destination relationship means that at least two of the multiple service transmission paths have the same destination node.

2. The method according to claim 1, characterized in that, The network model further includes a deep neural network (DNN) model, and the step of using the pre-trained network model to predict the transmission performance of the transmission parameters includes: The transmission parameters are processed using the GCN model to obtain transmission characteristic parameters; The transmission feature parameters are processed using the DNN model to obtain transmission performance prediction results.

3. The method according to claim 2, characterized in that, The process of using the GCN model to process the transmission parameters to obtain transmission characteristic parameters includes: The transmission parameters are processed using the following formula to obtain the transmission characteristic parameters: Among them, the The transmission characteristic parameters; The transmission parameters; The convolution kernel function; For the graph convolution operator using the spectral method, the following is represented: With the The product; the The convolution kernel parameters; The normalized graph Laplace operator; This is the activation function.

4. The method according to claim 2, characterized in that, After obtaining the transmission performance prediction results, the method further includes: The transmission performance prediction results are used to optimize the transmission performance of the service information.

5. The method according to any one of claims 1 to 4, characterized in that, The transmission parameters include at least one of the following: service traffic, optical signal-to-noise ratio (OSNR), bit error rate, and output optical power.

6. A method for generating a network model, comprising: Obtain transmission path information for service information in the network, wherein the transmission path information includes first node information and last node information; An input node is established based on the first node information and the last node information; The input nodes corresponding to the transmission path information that have the same source relationship or the same destination relationship are connected to obtain the GCN model, so that the GCN model conforms to the network topology of the transmission of the service information; The GCN model includes multiple input nodes for inputting transmission parameters of the service information. The multiple input nodes correspond to the transmission path information. The transmission parameters characterize the transmission performance of the service information in the network. Each of the multiple input nodes corresponds to one of the multiple service transmission paths in the network topology. The network topology includes multiple network nodes, and a service transmission path is formed between two network nodes. The same-source relationship indicates that at least two of the multiple service transmission paths have the same source node, and the same-destination relationship indicates that at least two of the multiple service transmission paths have the same destination node.

7. The method according to claim 6, characterized in that, The GCN model is represented by the following formula: Among them, the The output parameters of the GCN model; The input parameters are those input to the GCN model; The convolution kernel function; For the graph convolution operator using the spectral method, the following is represented: With the The product; the The convolution kernel parameters; The normalized graph Laplace operator; This is the activation function.

8. The method according to claim 6, characterized in that, The transmission parameters include at least one of the following: service traffic, OSNR, bit error rate, and output optical power.

9. A method for training a network model, comprising: Obtain transmission parameters of service information in the network, wherein the transmission parameters characterize the transmission performance of the service information in the network; The network model is trained using the transmission parameters, wherein the network model includes a GCN model, which is obtained based on the transmission path information of service information in the network, and the GCN model conforms to the network topology that transmits the service information; The GCN model includes multiple input nodes for inputting the transmission parameters. These input nodes correspond to the transmission path information, and each of the multiple input nodes corresponds to one of the multiple service transmission paths in the network topology. The network topology includes multiple network nodes, and a service transmission path is formed between two network nodes. The input nodes corresponding to transmission path information with the same source relationship or the same destination relationship are connected. The same source relationship means that at least two of the multiple service transmission paths have the same source node, and the same destination relationship means that at least two of the multiple service transmission paths have the same destination node.

10. The method according to claim 9, characterized in that, The network model further includes a DNN model, and training the network model using the transmission parameters includes: The transmission parameters are input to the input node to train the GCN model and obtain transmission feature samples; The transmission feature samples are input into the DNN model to train the DNN model.

11. The method according to claim 10, characterized in that, The step of inputting the transmission parameters to the input node to train the GCN model includes: The transmission parameters are input to the input node, and the GCN model is trained using the backpropagation algorithm.

12. The method according to claim 10, characterized in that, The step of inputting the transmitted feature samples into the DNN model to train the DNN model includes: The transmitted feature samples are input into the DNN model, and the DNN model is trained using the backpropagation algorithm.

13. The method according to any one of claims 9 to 12, characterized in that, The GCN model is represented by the following formula: Among them, the The output parameters of the GCN model; The input parameters are those input to the GCN model; The convolution kernel function; For the graph convolution operator using the spectral method, the following is represented: With the The product; the The convolution kernel parameters; The normalized graph Laplace operator; This is the activation function.

14. The method according to any one of claims 9 to 10, characterized in that, The transmission parameters include at least one of the following: service traffic, OSNR, bit error rate, and output optical power.

15. An electronic device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, when the processor executes the computer program, it implements the information processing method as described in any one of claims 1 to 5.

16. An electronic device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor, when executing the computer program, implements the generation method as described in any one of claims 6 to 8.

17. An electronic device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor, when executing the computer program, implements the training method as described in any one of claims 9 to 14.

18. A computer-readable storage medium storing computer-executable instructions, the computer-executable instructions being configured to perform the information processing method according to any one of claims 1 to 5, or to perform the generation method according to any one of claims 6 to 8, or to perform the training method according to any one of claims 9 to 14.

Images