Mobile telecommunication system data analytics for providing recommendations

Abstract

A method performed by a Data Analytics node configured to communicate with a consumer node associated with a mobile telecommunication system is disclosed. The method comprises determining, based on analytics input data and based on one or more predefined goals, a plurality of recommendations, each recommendation being associated with at least one action to take, wherein the plurality of recommendations differ from one another at least in the respective degrees to which they fulfil the one or more predefined goals. The method further comprises providing at least one of the determined plurality of recommendations to the consumer node. A method performed by the consumer node is also provided. The disclosure also relates to corresponding nodes, computer programs and carriers.

Description

[0001] MOBILE TELECOMMUNICATION SYSTEM DATA ANALYTICS FOR PROVIDING RECOMMENDATIONS

[0002] TECHNICAL FIELD

[0003] The present disclosure relates to a technique involving a Data Analytics node configured to communicate with a core network of a mobile telecommunication system and a consumer node.

[0004] BACKGROUND

[0005] Modern mobile telecommunication systems (e.g., according to the Third Generation Partnership, 3GPP, fifth generation, 5G, standard) comprise a (e.g., radio) access network, (R)AN, and a core network, CN.

[0006] Various Network Functions, NFs, may be implemented as part of the CN on one or more network nodes. These NFs may provide analytics data to one or more analytic entities, also referred to as Data Analytic entities herein, and may obtain results from these analytic entities, for example predictions, statistics and / or recommendations. For example, the Management Data Analytics Function, MDAF, can provide recommendations via a Management Data Analytics Service, MDAS. Management Data Analytics, MDA, according to the 3GPP 5G standard, is defined in 3GPP TS 28.104, see reference [2], and in 3GPP TS 28.809, see reference [3]. Another example of a Data Analytic entity is the Network Data Analytics Function, NWDAF. The NWDAF can provide an analytics report, which, according to the 3GPP 5G standard, is defined in 3GPP TS 23.288, see reference

[0010] .

[0007] In current solutions, a service consumer can either subscribe to MDAF or send a request and define the time window within which an initial response is required from the MDAF. The analytics performed by the MDAF currently focuses on providing recommendations as fast as possible to the service consumer (e.g., a consumer node configured as NF, communicating with a network node configured as MDAF). This approach may lead to recommendations by the MDAF that have drawbacks with regard to other parameters apart from response time.

[0008] SUMMARY

[0009] The present disclosure provides solutions to the above and other problems. Generally speaking, it is proposed to extend the capability of a Data Analytics, DA, node (e.g., configured as MDAF and / or MDAS) to:

[0010] 1. allow to generate multiple recommendations for given goal-defining parameters, for example not only parameters limited to prediction / statistics, thereby allowing exploration of various recommendations,

[0011] 2. allow for specifying an associated cost, for example in the form of a cost function, which could relate to energy consumption, energy costs, greenhouse gas etc.

[0012] The aforementioned cost function could (but is not limited to) specify one of or a mix of the following goals: time efficiency; resource efficiency; energy efficiency. If a request received by the DA node from a consumer node (e.g., a service subscriber) contains additional information with respect to combining goal-specific requirements, processing or transport need, Performance Measurements, PM, Configuration Management, CM, energy, or location information, the DA node implementing the MDAF may be requested to return multiple recommendations (e.g., prediction alternatives) and not a single recommendation (e.g., a single prediction), for example annotated with the combination of criteria required for the respective recommendation to be valid. Possible use cases related to this proposal are among others indicated in sections 5.5, 5.14, and 5.15 of TR 22.882, see reference [1].

[0013] For sake of clarity, it is noted that while some currently known solutions rely on the MDAF returning multiple objects for some analytics requests (e.g., identification of multiple cells and / or multiple User Plane Functions, UPFs), this is considered to be a single, albeit complex result. In other words, all such returned objects are connected to a single result prediction or recommendation and, as such, do not correspond to a plurality of recommendations as disclosed herein.

[0014] According to the present disclosure, an analytics request may be transmitted from the consumer node to the DA node for exploration of a plurality of recommendations, for example guided by a submitted cost- and / or constraints- and / or preferences- function. The DA node (e.g., MDAF) action recommendations may thus be extended with an exploration over possible parameters associated with multiple goals. This differs from other implementations in which only actions for a single goal are recommended to a consumer node (e.g. recommend actions to save energy in a part of the network). Additional information may be included in the request of the consumer node (e.g., constraints on resources, preferences, objective functions). This information may be used by the DA node to determine suitable recommendations. It is also possible to enable a guided exploration of the recommendations. In one example, a number of actions may be requested. In another example, prioritized actions may be requested. In a still further example, it may be requested to pause or stop exploring (e.g., determining additional recommendations and / or selecting appropriate recommendations from a plurality of already determined recommendations).

[0015] The consumer node may transmit a flexible subscription request to the DA node, which may differ from a typical request or subscription query. The flexible subscription request may register a policy with the DA node (e.g., configured as MDAF) that defines when and what action recommendation can be sent to the consumer node. This may reduce the number of required request messages. The flexible subscription request may instruct the DA node (e.g., configured as MDAF) to learn the pattern of access (e.g., learning a policy) of the consumer node. Generation of recommendations may be scheduled based on input time constraints (e.g., time relaxations), request parameters and / or network state. It is also envisaged that the DA node uses different analytic models with different energy profile. Furthermore, the collection of data to be analyzed by the DA node to determine the recommendations may be affected, for example in lower layers (e.g., through NWDAF). For example, additional counters may be collected or the data collection resolution may be temporarily increased while exploring (e.g., determining the recommendations), thereby also increasing the quality of the recommendations. The present disclosure also provides for validating actions during exploration. For example, the associated recommendation may not be returned to the consumer node until it has been validated with a digital twin or similar, e.g. to confirm that it matches with the cost and the goal-specific parameters.

[0016] Again, while current solutions do not address costs, in particular current energy analytics do not cover any aspect about energy cost or greenhouse gas, GHG, emissions of recommended actions, the proposed technique enables consideration of multiple goals by the DA node for providing suitable recommendations to a consumer node. Decisions to perform a task may be driven by the analysis and predictions of the network state coming from the DA node (e.g., configured as MDAF). By ensuring that the DA node does not determine the recommendations only to fulfil the goal of time efficiency (e.g., assumed urgency), but also in consideration of other goals, for example by introducing more information about both the task itself and availability of various energy sources, more flexibility is given and alternative solutions or actions for a given request can be provided. For example, assume that we have a request for prediction that will affect the decision to schedule a task that we need to assign to be executed on various units. Tasks and units may be very diverse - tasks may range from short to very long-time scales (e.g., in case of transmission of a software update), a unit may be a single antenna or entire sites. If one were to base the recommendations provided by the Data Analytics node only by optimizing for urgency, the opportunity to find other solutions that are more energy or resourceefficient but less time-efficient would be hindered.

[0017] These solutions can be rephrased as methods, nodes, computer programs, carriers and a system as set forth in the following aspects. It is thus to be understood that the following aspects may be combined with the features of the present technique as mentioned above.

[0018] According to a first aspect, a method performed by a DA node is provided. The DA node is configured to communicate with a consumer node associated with a mobile telecommunication system. The method comprises determining, based on analytics input data and based on one or more predefined goals, a plurality of recommendations, each recommendation being associated with at least one action to take, wherein the plurality of recommendations differ from one another at least in the respective degrees to which they fulfil the one or more predefined goals. The method further comprises providing at least one (e.g., two or more) of the determined plurality of recommendations to the consumer node.

[0019] The DA node may form part of an Operations Administrations and Management, 0AM, domain (e.g., of the mobile telecommunication system). The DA node may form part of a Management and Orchestration, domain (e.g., of the mobile telecommunication system). The DA node may be configured to implement one or more Data Analytic Functions, DAFs. The DA node may be referred to as a Data Analytics entity. The DA node may be configured as (e.g., configured to implement) a Data Analytics Function, DAF. The DA node may be configured as a MDAF or as NWDAF. The DA node may provide a MDAS.

[0020] The mobile telecommunication system as understood herein comprises a CN and optionally a (R)AN. In case "RAN" is mentioned herein, this abbreviation may refer to an access network, for example a radio access network. The RAN may be configured according to a 3GPP standard such as 5G or 6G. The DA node may be configured to communicate with the CN of the mobile telecommunication system (e.g., directly and / or via a predefined interface, or via the consumer node). A node being "associated with" the mobile telecommunication system as understood herein may mean one or more of: the node is configured to communicate with the CN (e.g., directly and / or via a predefined interface); the node is an integral part of the mobile telecommunication system; the node is part of a network domain of the mobile telecommunication system; the node is configured to implement one or more NFs of the CN of the mobile telecommunication system; the node is part of the CN of the mobile telecommunication system. These associations may differ between different associated nodes or be the same for each associated node. The consumer node may be configured as a NF of the CN of the mobile telecommunication system. The consumer node may be configured as NWDAF.

[0021] The analytics input data may be associated with the mobile telecommunication system, for example with the CN thereof. The analytics input data may be obtained from one or more NFs of the CN (e.g., NFs of the CN). The analytics input data may be related to network and / or service events and / or status (e.g., of the mobile telecommunication system, for example the CN and / or the (R)AN), including, e.g., performance measurements, Key Performance Indicators, KPIs, Trace reports, Minimization of Drive Tests-, MDT-, reports, Radio Link Failure-, RLF-, reports, Radio Resource Control Connection Establishment Failure-, RCEF-, reports, Quality of Experience-, QoE-, reports, alarms, configuration data, network analytics data, and service experience data from Application Functions, AFs, etc. The analytics input data may correspond to the data analyzed by the MDA according to the 3GPP standard, see for example reference [2] and / or [3], but is not limited thereto. For example, the analytics input data may correspond to the data analyzed by the NWDA(F) according to the 3GPP standard, see for example reference

[0010] and / or the other 3GPP documents referenced therein.

[0022] According to the present disclosure, one or more or each recommendation may relate to the mobile telecommunication system. One or more or each recommendation may be associated with one or more actions to take by the consumer node, or on behalf of instructions of the consumer node (e.g., to affect the mobile telecommunication system). One or more or each recommendation may be associated with a network and / or service operation (e.g., an operation of the mobile telecommunication system). The actions disclosed herein may comprise one or more of: an action for prevention and / or prediction of network and / or service demands; a remedy action (e.g. reconfigure or add cells, beams, antennas, etc.); an action to reconfigure (e.g., RAN) attributes including handover parameters, cell reselection parameters, beam configuration, computing resource and slice support in a cell; an action which may suggest stopping paging the UE for Daily-Out-Of-Coverage- Duration at Daily-Out-Of-Coverage-Location; a recommended action to solve an E2E latency issue; a recommended action to remedy or prevent a network slice load issue; a recommended action for service recovery (e.g., update of one or more NFs, change configuration of CN NF etc.); a policy and configuration action to guarantee a network performance and end user service experience; a recommended action for optimal handover parameters. The recommendations and / or actions disclosed herein may comprise or correspond to those disclosed in TS 28.104, see reference [2], but are not limited thereto.

[0023] According to the present disclosure, providing information (e.g., a recommendation, an indication, a parameter, a goal or else) to another entity as understood herein may comprise at least one of {i} transmitting this information to the other entity directly (e.g., contained in one or more messages) or indirectly (e.g., by transmitting an indication of said information to the other entity) and {ii} keeping this information available to be grabbed by the other entity via a data retrieval interface such as an API. Not all information need to be provided in the same manner, different information may be provided in a different manner.

[0024] According to the present disclosure, obtaining information (e.g., data, a recommendation, an indication, a parameter, a goal or else) from another entity as understood herein may comprise at least one of {i} receiving this information from the other entity directly (e.g., contained in one or more messages) or indirectly (e.g., by receiving an indication of said information from the other entity) and {ii} grabbing this information from the other entity via a data retrieval interface thereof, such as an API. Not all information need to be obtained in the same manner, different information may be obtained in a different manner.

[0025] Each recommendation may be optimized for a different goal or for a different weighting of the one or more predefined goals. The goal(s) and / or weighting(s) for which the recommendations are optimized may be obtained from the consumer node. For example, each recommendation may be optimized based on a cost function obtained from the consumer node. The cost function may indicate the goal(s), the weighting(s) and / or one or more parameters defining said goal(s). In one example, the one or more predefined goals include(s) at least one of: an energy efficiency; a resource efficiency; a time efficiency; a Quality of Service, QoS. The energy efficiency may be defined by at least one energy efficiency parameter selected from: an energy consumption; an amount of greenhouse gas emissions; a percentage of renewable energy; a selection or number of computing resources using renewable energy and / or emitting a predefined maximum amount of greenhouse gases; a use of energy in a time period at which renewable energy is available. The resource efficiency may be defined by at least one resource efficiency parameter selected from: a computing power; a data storage amount; an amount of redistribution of computing resources; a usage of one or more network functions. The time efficiency may be defined by at least one time efficiency parameter selected from: a required time; a delay. The QoS may be defined by at least one QoS parameter selected from: a minimum data transmission rate, a maximum communication latency, a minimum communication security.

[0026] The respective degrees to which the recommendations fulfil the one or more goals may be indicative of or correspond to the degrees to which the associated actions fulfil the one or more goals and / or the parameter(s) defining said one or more goals. The method may further comprise selecting at least one of a plurality of analytic models, each being associated with a different degree to which the recommendations determined by said model fulfil the one or more goals and / or the parameter(s) defining said one or more goals. The plurality of recommendations may be determined based on the selected at least one analytic model.

[0027] The respective degrees to which the recommendations fulfil the one or more goals may be indicative of or correspond to the degrees to which the determination of the respective recommendations by the DA node fulfils the one or more goals and / or the parameter(s) defining said one or more goals. The method may further comprise selecting at least one of a plurality of analytic models, each being associated with a different degree to which the determination by said model fulfils the one or more goals and / or the parameter(s) defining said one or more goals. The plurality of recommendations may be determined based on the selected at least one analytic model.

[0028] The method may further comprise obtaining an indication of the one or more predefined goals, wherein the plurality of recommendations differ from one another in the respective degrees to which they fulfil these indicated one or more predefined goals. The indication of the one or more predefined goals may include an indication of the at least one parameter defining the one or more predefined goals. The indication of the one or more predefined goals may include a (e.g., the) cost function.

[0029] The method may further comprise obtaining an indication of a prioritization or selection of potential recommendations. The at least one recommendation may be provided to the consumer node based on the indication of the prioritization or selection.

[0030] The method may further comprise obtaining an indication to pause or stop the determination of recommendations. The method may comprise pausing or stopping the determination of the plurality of recommendations in response to obtaining said indication to pause or stop the determination of recommendations. The at least one recommendation provided to the consumer node may be part of the recommendation(s) determined before pausing or stopping the determination of the plurality of recommendations.

[0031] The method may further comprise obtaining an indication of at least one trigger condition for providing the at least one recommendation to the consumer node. The method may comprise determining whether the at least one trigger condition is met. the method may comprise in response to determining that the at least one trigger condition is met, providing the at least one recommendation to the consumer node. The at least one trigger condition may relate to a state of the mobile telecommunication system, for example a state of the CN of the mobile telecommunication system and / or to a state of the RAN of the mobile telecommunication system.

[0032] One or more or all of the aforementioned indication(s) may be defined by and / or obtained from a policy associated with the consumer node and known to the DA node.

[0033] The method may further comprise: based on the indication(s) that were obtained, deriving and / or learning the policy associated with the consumer node for a subsequent provision of recommendations.

[0034] The method may further comprise receiving one or more messages from the consumer node. The one or more messages may {i} request to provide at least one recommendation, {ii} contain the indication(s), {iii} contain the policy and / or {iv} request associating the policy with the consumer node.

[0035] The method may further comprise requesting one or more network nodes associated with the mobile telecommunication system to collect the analytics input data such that at least one data collection condition is fulfilled. The at least one data collection condition may comprise a sampling rate, a number of parameters to be monitored and / or a data collection time. The one or more network nodes may comprise a network node configured as a Network Function, NF, of the mobile telecommunication system, for example a NF of the CN of the mobile telecommunication system.

[0036] The method may further comprise validating one or more of the determined recommendations. The at least one recommendation provided to the consumer node may be a validated recommendation. Validating a recommendation may comprise applying the at least one action associated with said recommendation to a digital twin network.

[0037] The at least one action to take may comprise or consists of at least one action to take by the consumer node or by another network node configured as a NF of the mobile telecommunication system, for example a NF of the CN of the of the mobile telecommunication system.

[0038] According to a second aspect, a DA node configured to communicate with a consumer node associated with a mobile telecommunication system is provided. The DA node is configured to determine, based on analytics input data and based on one or more predefined goals, a plurality of recommendations, each recommendation being associated with at least one action to take, wherein the plurality of recommendations differ from one another at least in the respective degrees to which they fulfil the one or more predefined goals; and provide at least one of the determined plurality of recommendations to the consumer node. The DA node may be configured to perform the method according to the first aspect.

[0039] According to a third aspect, a method performed by a consumer node is provided. The consumer node is associated with a mobile telecommunication system and configured to communicate with a DA node. The method comprises obtaining, from the DA node, at least one of a plurality of recommendations, the plurality of recommendations being determined by the DA node based on one or more predefined goals, each recommendation being associated with at least one action to take, wherein the plurality of recommendations differ from one another at least in the respective degrees to which they fulfil the one or more predefined goals.

[0040] It is to be understood that the method according to the aspect may be a counterpart to the method according to the first aspect and vice versa. Accordingly, all examples, explanations, features and definitions provided for the first aspect may similarly apply to the third aspect, mutatis mutandis, and vice versa. A repetition thereof will be avoided for brevity.

[0041] Each recommendation may be optimized for a different goal or for a different weighting of the one or more predefined goals. Each recommendation may be optimized with respect to a cost function (e.g., obtained from the consumer node).

[0042] The one or more predefined goals may include(s) at least one of: an energy efficiency; a resource efficiency; a time efficiency; a Quality of Service, QoS. The energy efficiency may be defined by at least one energy efficiency parameter selected from: an energy consumption; an amount of greenhouse gas emissions; a percentage of renewable energy; a selection or number of computing resources using renewable energy and / or emitting a predefined maximum amount of greenhouse gases; a use of energy in a time period at which renewable energy is available. The resource efficiency may be defined by at least one resource efficiency parameter selected from: a computing power; a data storage amount; an amount of redistribution of computing resources; a usage of one or more network functions.

[0043] The time efficiency may be defined by at least one time efficiency parameter selected from: a required time; a delay. The QoS may be defined by at least one QoS parameter selected from: a minimum data transmission rate, a maximum communication latency, a minimum communication security.

[0044] The respective degrees to which the recommendations fulfil the one or more goals may be indicative of or correspond to the degrees to which the associated actions fulfil the one or more goals and / or the parameter(s) defining said one or more goals. The plurality of recommendations may be determined by the DA node based on at least one analytic model selected from a plurality of analytic models, each being associated with a different degree to which the recommendations determined by said model fulfil the one or more goals and / or the parameter(s) defining said one or more goals. The respective degrees to which the recommendations fulfil the one or more goals may be indicative of or correspond to the degrees to which the determination of the respective recommendations by the DA node fulfils the one or more goals and / or the parameter(s) defining said one or more goals. The plurality of recommendations may be determined by the DA node based on at least one analytic model selected from a plurality of analytic models, each being associated with a different degree to which the determination by said model fulfils the one or more goals and / or the parameter(s) defining said one or more goals.

[0045] The method may further comprise providing, to the DA node, an indication of the one or more predefined goals, wherein the plurality of recommendations differ from one another in the respective degrees to which they fulfil these indicated one or more predefined goals. The indication of the one or more predefined goals may include an indication of the at least one parameter defining the one or more predefined goals. The indication of the one or more predefined goals includes a cost function.

[0046] The method may further comprise providing, to the DA node, an indication of a prioritization or selection of potential recommendations. The at least one recommendation obtained from the DA node may be based on the indication of the prioritization or selection.

[0047] The method may further comprise providing, to the DA node, an indication to pause or stop the determination of recommendations, said indication causing the DA node to pause or stop the determination of the plurality of recommendations, wherein the at least one recommendation obtained from the DA node is part of the recommendation(s) determined by the DA node before the determination of the plurality of recommendations is paused or stopped.

[0048] The method may further comprise providing, to the DA node, an indication of at least one trigger condition for providing the at least one recommendation to the consumer node, said indication causing the DA node to determine whether the at least one trigger condition is met and to provide the at least one recommendation to the consumer node in response to determining that the at least one trigger condition is met. The at least one trigger condition may relate to a state of the mobile telecommunication system, for example a state of the CN and / or a state of the RAN of the mobile telecommunication system. The indication(s) may be defined by and / or provided via a policy associated with the consumer node and known to the DA node. The method may further comprise requesting the DA node to derive and / or learn the policy associated with the consumer node based on the provided indication(s) for a subsequent provision of recommendations.

[0049] The method may further comprise transmitting one or more messages to the DA node. The one or more messages may {i} request to provide at least one recommendation, {ii} contain the indication(s), {iii} contain the policy and / or {iv} request associating the policy with the consumer node.

[0050] The method may further comprise requesting the DA node to request one or more network nodes associated with the mobile telecommunication system to collect the analytics input data such that at least one data collection condition is fulfilled. The at least one data collection condition may comprise a sampling rate, a number of parameters to be monitored and / or a data collection time. The one or more network nodes may comprise a network node configured as a NF of the mobile telecommunication system, for example a NF of the CN of the mobile telecommunication system.

[0051] The method may further comprise requesting the DA node to validate one or more of the determined recommendations. The at least one recommendation obtained from the DA node may be a validated recommendation. Validating a recommendation may comprise applying the at least one action associated with said recommendation to a digital twin network.

[0052] The at least one action to take may comprise or consists of at least one action to take by the consumer node or by another network node configured as a NF of the (e.g., CN of the) mobile telecommunication system.

[0053] The method may comprise performing the at least one action to take or instructing the another network node to perform the at least one action to take, in particular the at least one action to take that is associated with the at least one recommendation obtained from the DA node.

[0054] The DA node and the consumer node may be configured as explained above for the first aspect. For example, the DA node may be configured as NWDAF or MDAF. The consumer node may be configured as a NF of the CN. According to a fourth aspect, a consumer node is provided. The consumer node is configured to communicate with a DA node of a CN of a mobile telecommunication system. The consumer node is configured to obtain, from the DA node, at least one of a plurality of recommendations, the plurality of recommendations being determined by the DA node based on one or more predefined goals, each recommendation being associated with at least one action to take, wherein the plurality of recommendations differ from one another at least in the respective degrees to which they fulfil the one or more predefined goals. The consumer node may be configured to perform the method according to the third aspect.

[0055] According to a fifth aspect, a system comprising the DA node of the second aspect and comprising the consumer node of the fourth aspect is provided. The system may be configured such that the DA node performs the method of the first aspect and the consumer node performs the method of the third aspect (e.g., in collaboration). A corresponding method performed by a system comprising a DA node and a consumer node is also provided as a sixth aspect.

[0056] According to a seventh aspect, a computer program is provided. The computer program comprises instructions that, when executed by processing circuitry (e.g., of the DA node of the second aspect or of the consumer node of the fourth aspect), cause the processing circuitry to carry out the method according to the first aspect or according to the third aspect.

[0057] According to an eight aspect, a carrier is provided. The carrier contains the computer program of the seventh aspect. The carrier may be one of an electronic signal, optical signal, radio signal, or a (e.g., non-transitory) computer-readable medium. The carrier may be part of the DA node of the second aspect or of the consumer node of the fourth aspect.

[0058] BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Fig. 1 i Illustrates an exemplary mobile telecommunication system in accordance with the present disclosure;

[0060] Fig. 2 illustrates an exemplary node in accordance with the present disclosure; Fig. 3 illustrates an exemplary mobile telecommunication system in accordance with the present disclosure and the 3GPP network architecture;

[0061] Fig. 4 illustrates a flow diagram of a DA node method in accordance with the present disclosure;

[0062] Fig. 5 illustrates a flow diagram of a consumer node method in accordance with the present disclosure;

[0063] Fig. 6 illustrates a flow diagram of a first exemplary scheme in accordance with the present disclosure; and

[0064] Fig. 7 illustrates a flow diagram of a second exemplary scheme in accordance with the present disclosure.

[0065] DETAILED DESCRIPTION

[0066] The technique disclosed herein will now be explained with reference to the drawings.

[0067] Fig. 1 shows an example of a mobile telecommunication system 100 in accordance with the present disclosure.

[0068] In the example, the mobile communication system 100 includes an access network 300, such as a radio access network, RAN, and a core network 200, which includes one or more core network nodes 202, 204, 206, each of which may be configured as one or more NFs of the CN 200. The mobile telecommunication system 100 may also be referred to as mobile telecommunication network. The access network 300 includes one or more access network nodes, such as network nodes 302, 304, or any other similar 3GPP access nodes or non-3GPP access points. Moreover, as will be appreciated by those of skill in the art, a network node is not necessarily limited to an implementation in which a radio portion and a baseband portion are supplied and integrated by a single vendor. Thus, it will be understood that network nodes include disaggregated implementations or portions thereof. For example, in some embodiments, the telecommunication system 100 includes one or more Open-RAN, ORAN, network nodes. An ORAN network node is a node in the telecommunication system 100 that supports an ORAN specification (e.g., a specification published by the O-RAN Alliance, or any similar organization) and may operate alone or together with other nodes to implement one or more functionalities of any node in the telecommunication system 100, including one or more access network nodes 302, 304 and / or core network nodes 202, 204, 206.

[0069] The network nodes 302, 304 facilitate direct or indirect connection of user equipment, UE, such as by connecting UEs 402, 404 to the core network 200 over one or more wireless connections. Example wireless communications over a wireless connection include transmitting and / or receiving wireless signals using electromagnetic waves, radio waves, infrared waves, and / or other types of signals suitable for conveying information without the use of wires, cables, or other material conductors. Moreover, in different embodiments, the communication system 100 may include any number of wired or wireless networks, network nodes, UEs, and / or any other components or systems that may facilitate or participate in the communication of data and / or signals whether via wired or wireless connections. The communication system 100 may include and / or interface with any type of communication, telecommunication, data, cellular, radio network, and / or other similar type of system.

[0070] The UEs 402, 404 may be any of a wide variety of communication devices, including wireless devices arranged, configured, and / or operable to communicate wirelessly with the access network nodes 302, 304 and other communication devices. Similarly, the access network nodes 302, 304 are arranged, capable, configured, and / or operable to communicate directly or indirectly with the UEs 402, 404 and / or with other network nodes or equipment in the telecommunication system 100 to enable and / or provide network access, such as wireless network access, and / or to perform other functions, such as administration in the telecommunication system 100.

[0071] As a whole, the communication system 100 of Figure 1 enables connectivity between the UEs, network nodes, and hosts. In that sense, the communication system may be configured to operate according to predefined rules or procedures, such as specific standards that include, but are not limited to: Global System for Mobile Communications, GSM; Universal Mobile Telecommunications System, UMTS; Long Term Evolution, LTE, and / or other suitable 2G, 3G, 4G, 5G standards, or any applicable future generation standard (e.g., 6G); wireless local area network, WLAN, standards, such as the Institute of Electrical and Electronics Engineers, IEEE, 802.11 standards, WiFi; and / or any other appropriate wireless communication standard, such as the Worldwide Interoperability for Microwave Access, WiMax, Bluetooth, Z- Wave, Near Field Communication, NFC, ZigBee, LiFi, and / or any low-power wide-area network, LPWAN, standards such as LoRa and Sigfox. In some examples, the telecommunication system 100 is a cellular network that implements 3GPP standardized features. Accordingly, the telecommunications network 100 may support network slicing to provide different logical networks to different devices that are connected to the telecommunication system 100. For example, the telecommunications network 100 may provide Ultra Reliable Low Latency Communication, URLLC, services to some UEs, while providing Enhanced Mobile Broadband, eMBB, services to other UEs, and / or Massive Machine Type Communication, mMTC / Massive loT services to yet further UEs.

[0072] In some examples, the UEs 402, 404 are configured to transmit and / or receive information without direct human interaction. For instance, a UE may be designed to transmit information to the access network 300 on a predetermined schedule, when triggered by an internal or external event, or in response to requests from the access network 300. Additionally, a UE may be configured for operating in single- or multi- RAT or multi-standard mode. For example, a UE may operate with any one or combination of Wi-Fi, New Radio, NR, and LTE, i.e. being configured for multi-radio dual connectivity, MR-DC, such as Evolved-UMTS Terrestrial Radio Access Network, E-UTRAN, New Radio - Dual Connectivity, EN-DC.

[0073] In the illustrated example, network nodes 502, 504 are shown, which are configured to communicate with the mobile telecommunication system 100, in particular with the CN 200. The nodes 502, 504 may also be configured to communication directly with one another. These nodes may be associated with an OAM domain and configured to perform one or more DAFs, so these nodes may also be referred to as DA nodes. For example, node 502 may be configured as MDAF and node 504 may be configured as NWDAF.

[0074] Although a certain number of network nodes is illustrated in Fig. 1, it is to be understood that a larger or lower number of network nodes may be foreseen for implementing the technique disclosed herein. Each network node may be configured to implement one or more Functions (e.g., NFs and / or DAFs). A network node may be implemented as a physical entity, a distributed entity (e.g., jointly provided by multiple physical devices) or a virtual entity (e.g., provided by a cloud computing platform).

[0075] Fig. 2 illustrates an exemplary node 1000 in accordance with the present disclosure.

[0076] It is to be understood that one or more or each of the nodes 202, 204, 206, 302, 304, 502, 504 may be configured in this manner. As shown, the mobile network node 1000 may include network interface circuitry 1100 (also referred to as a network interface) configured to provide communications with other nodes of the core network and / or the network (e.g., the overall mobile communication system). The mobile network node 1000 may also include a processing circuitry 1200 (also referred to as a processor) coupled to the network interface circuitry 1100, and memory circuitry 1300 (also referred to as memory) coupled to the processing circuitry 1200. The memory circuitry 1300 may include computer readable program code that when executed by the processing circuitry 1200 causes the processing circuitry 1200 to perform the method(s) or scheme(s) disclosed herein. According to other embodiments, processing circuitry 1200 may be defined to include memory so that a separate memory circuitry is not required. As discussed herein, operations of the mobile network node 1000 may be performed by processing circuitry 1200 and / or network interface circuitry 1100. For example, processing circuitry 1200 may control network interface circuitry 1100 to transmit communications through network interface circuitry 1100 to one or more other network nodes and / or to receive communications through network interface circuitry 1100 from one or more other network nodes. Moreover, modules may be stored in memory 1300, and these modules may provide instructions so that when instructions of a module are executed by processing circuitry 1200, processing circuitry 1200 performs respective operations (e.g., in accordance with the methods and schemes discussed herein).

[0077] An exemplary architecture of the CN 200 of a mobile telecommunication system according to the 3GPP's 5G standard is illustrated in Fig. 3. The RAN 300, UE 402, NWDAF and MDAF are also indicated, as well as the Data Network, DN. As can be seen, UE 402 is connected to the Access Management Function, AMF via an N1 interface or via RAN 300 using an N2 interface. The CN comprises various NFs in addition to said AMF, for example an Application Function, AF, and a Policy Control Function, PCF. For details on these and the other illustrated NFs of the 5G CN, it is referred to 3GPP TS 23.501, see reference

[0011] , and 3GPP TS 23.502, see reference

[0012] .

[0078] In the CN (e.g., 200) of a mobile telecommunication system according to the 3GPP's 5G standard, artificial intelligence-, AI-, based NFs and / or Management Services may be of interest that can be used by service consumers (e.g., consumer nodes) to get statistics and predictions for different types of reports. For instance, the NWDAF may give analytics reports about user equipment, UE, data congestion or UE mobility which will predict whether there will be congestion in the network or how a UE trajectory will look like in the future. MDAS's may be in place to predict different events related to management of the network, such as failure in NFs, energy efficiency, etc. Although in both cases a service consumer may send a request to the respective NF with the possibility to define complementary information to make the request more granular, there currently is no way to put any constraints on the results provided by the CA implementing said NF. Moreover, these NFs may not provide any recommendation for any given value such as energy optimization parameters, QoS, policies, etc.

[0079] In current solutions, a service consumer (e.g., a NF of CN 200) may collect all necessary information from different sources (e.g., a DA node configured as NWDAF or MDAS / MDAF), and internally calculate the best values of interest before triggering any relevant action (e.g., network reconfiguration). The problems that might arise from this approach are set out below:

[0080] • Signaling in the network can be increased since instead of a centric analysis, each service consumer collects necessary data and analytics reports to calculate the best values (e.g., for network configuration parameters).

[0081] • Inconsistency can happen while different service consumers might use different logics which can lead to triggering actions that might increase the risk of conflict in the overall system configuration.

[0082] • Several service consumers might work on same network parameters which may make resource utilization inefficient.

[0083] • Security concerns might make this approach impossible if some of the inputs are inaccessible or protected.

[0084] In view of these issues, the present disclosure ensures that a plurality of recommendations are (e.g., centrally) determined by a DA node, and at least one of these recommendations is then provided to a service consumer. The recommendations are determined in consideration of a plurality of goals, so different recommendations provide different tradeoffs between these goals. The consumer node may provide a cost function to the DA node to influence the determination of the recommendations accordingly.

[0085] Fig. 4 illustrates a flowchart of an exemplary method in accordance with the present disclosure. This method is performed by the DA node (e.g., 202), for example being configured as MDAF. At S2, the DA node obtains an indication of one or more predefined goals. The indication of the one or more predefined goals may include an indication of at least one parameter defining the one or more predefined goals and / or a cost function (e.g., for these parameters).

[0086] At S4, the DA node obtains an indication of a prioritization or selection of potential recommendations.

[0087] At S6, the DA node obtains an indication of at least one trigger condition for providing at least one recommendation to the consumer node. The at least one trigger condition may relate to a state of the CN 200 and / or a state of RAN 300.

[0088] At S8, the DA node determines that the at least one trigger condition is met, and then proceeds to step S10.

[0089] At S10, the DA node requests one or more network nodes (e.g., NFs of the CN 200) to collect analytics input data. This request may be triggered by a corresponding trigger message received by the DA node from the consumer node. The request may be aimed to fulfil at least one data collection condition such as a sampling rate, a number of parameters to be monitored and / or a data collection time. An indication of said data collection condition(s) may be provided by the consumer node to the DA node.

[0090] At S12, the DA node determines, based on the analytics input data and based on the one or more predefined goals, a plurality of recommendations. Each recommendation is associated with at least one action to take, wherein the plurality of recommendations differ from one another at least in the respective degrees to which they fulfil the one or more predefined goals. The at least one action to take may comprise or consists of at least one action to take by a consumer node (e.g., 204 or 206) or by another network node configured as a NF of the CN 200 (e.g., in response to corresponding instructions by the consumer node).

[0091] Each recommendation may be optimized for a different goal or for a different weighting of the one or more predefined goals, for example an energy efficiency, a resource efficiency, or a time efficiency. Each of these goals can be defined by one or more parameters. The respective degrees to which the recommendations fulfil the one or more goals may be indicative of or correspond to the degrees to which the associated actions fulfil the one or more goals and / or the parameter(s) defining said one or more goals. In this case, the DA node may, in sub-step S121, select at least one of a plurality of analytic models, each being associated with a different degree to which the recommendations determined by said model fulfil the one or more goals and / or the parameter(s) defining said one or more goals. In this case, the plurality of recommendations are determined based on the selected at least one analytic model.

[0092] The respective degrees to which the recommendations fulfil the one or more goals may be indicative of or correspond to the degrees to which the determination of the respective recommendations by the DA node fulfils the one or more goals and / or the parameter(s) defining said one or more goals. In this case, the DA node may, in substep S121, select at least one of a plurality of analytic models, each being associated with a different degree to which the determination by said model fulfils the one or more goals and / or the parameter(s) defining said one or more goals. Also in this case, the plurality of recommendations are determined based on the selected at least one analytic model.

[0093] At S14, the DA node obtains an indication to pause or stop the determination of recommendations, and stops said determination accordingly.

[0094] At S16, the DA node validates one or more of the determined recommendations, e.g., by applying the at least one action associated with said recommendation to a digital twin network.

[0095] At S18, the DA node selects, sorts and / or filters the plurality of recommendations (e.g., in accordance with the indications obtained previously).

[0096] At S20, the DA node provides at least one of the recommendations (e.g., multiple recommendations) to the consumer node. The consumer node may then perform the action(s) associated with the at least one recommendation, optionally after having selected one of the recommendations, or may instruct other NFs to perform these actions accordingly.

[0097] Fig. 5 illustrates a flowchart of an exemplary method in accordance with the present disclosure. This method is performed by the consumer node, for example being configured as NF of the CN 200. The method of Fig. 5 works in conjunction with that of Fig. 4 and vice versa.

[0098] At S42, the consumer node provides, to the DA node, the indication of the one or more predefined goals. As said, the plurality of recommendations may differ from one another in the respective degrees to which they fulfil these indicated one or more predefined goals. The indication of the one or more predefined goals may include an indication of the at least one parameter defining the one or more predefined goals. The indication of the one or more predefined goals may include a cost function.

[0099] At S44, the consumer node provides, to the DA node, the indication of the prioritization or selection of potential recommendations.

[0100] At S46, the consumer node provides, to the DA node, the indication of at least one trigger condition for providing the at least one recommendation to the consumer node, said indication causing the DA node to determine whether the at least one trigger condition is met and to provide the at least one recommendation to the consumer node in response to determining that the at least one trigger condition is met

[0101] At S48, the consumer node provides, to the DA node, an indication or a request (e.g., a triggering message) for the DA node to request one or more network nodes (e.g., configured as NFs of the CN 200) to collect the analytics input data. The consumer node may provide, to the DA, an indication of node data collection condition(s).

[0102] At S50, the consumer node provides, to the DA node, the indication to pause or stop the determination of recommendations, said indication causing the DA node to pause or stop the determination of the plurality of recommendations.

[0103] At S52, the consumer node obtains, from the DA node, at least one of the plurality of recommendations. As explained above, the plurality of recommendations are determined by the DA node based on one or more predefined goals, each recommendation being associated with at least one action to take, wherein the plurality of recommendations differ from one another at least in the respective degrees to which they fulfil the one or more predefined goals. At 54, the consumer node performed the at least one action to take that is associated with at least one of the obtained recommendation(s), for example a recommendation selected by the consumer node. It is also possible for the consumer node to instruct another network node such as the AF or PCF to perform the at least one action to take.

[0104] In the context of Figs. 4 and 5, it is to be understood that the indications are not essential and may be omitted. It is possible to obtain one or more indications as part of a single message or in separate (e.g., request and / or subscription) messages from the consumer node. The indications may be obtained from a policy associated with the consumer node. In this case, no messages including the indications need to be received by the DA node. Said policy may be learned and / or adapted by the DA node over time, for example upon request by the consumer node. It is also noted that one or more of S2-S10, S121, S14-S18 and / or one or more of S42-S50, S54 may be omitted. The sequences of S2-S20 and / or S42-S54 may be changed.

[0105] As a specific example of how the present technique can be put into practice, it is referred to clause 8.4.4 of TS 28.104, see reference [2], according to which an MDA can provide energy saving analysis upon request from a service consumer. The input to MDA to receive energy consumption analytics data is demonstrated in Table 1 of this document, see reference [2]. In this case, 0AM may be the service consumer of this analytics data as well as Observed Service Experience from NWDAF. Then based on the recommended action by incorporating the cost function, the 0AM will decide what action to trigger (e.g., re-configuring a NF, re-selection of a NF service provider, etc.). Said Table 1 lists input information that is sent to the MDA from a service consumer (e.g., consumer node) to receive analytics data. The last row in the table reproduced below shows a cost function which could be added to put the present technique into practice, thereby enhancing the request with constraints over the recommended actions.

[0106] Table 1: Enabling data for energy saving analysis

[0107] Two schemes will now be described that are applicable not only but also to this particular example. In this particular example, the OAM subscribes to the MDA to receive a prediction about a traffic load trend for a UE as well as a recommendation to reduce energy consumption. The OAM uses such information to update configuration and management parameters to improve energy footprint in the network. The MDA may thereby support the following recommendations in response to the energy saving data analytics request: rANenergySavingRecommendations:

[0108] • For energy saving, ES, on New Radio, NR, cells. It may contain a set of: o Recommended NR Cell, ES-Cell, to enter energysaving state. o Recommended candidate cells with precedence for taking over the traffic of the ES-Cell. oThe time to enter and terminate the energy saving state. oThe load threshold to enter and terminate the energy saving state for the ES-Cell.

[0109] • This may only exist in case RAN energy saving is supported. cNenergySavingRecommendations:

[0110] • For ES on UPFs. It contains a set of: o Recommended UPF, ES-UPF, to conduct energy saving. o Recommended candidate UPFs with precedence for taking over the traffic of the ES-UPF. oThe time to conduct energy saving for the ES-UPF.

[0111] • This may only exists in case CN energy saving is supported.

[0112] When applying this particular example to the known MDA functionality, the MDA will return only a single prediction about traffic load trend of the UE, for example the following recommendation:

[0113] • Cells A and B are the best candidates to take over the traffic due to technology, low resource utilization and better coverage.

[0114] • UPF A shall enter energy saving mode at time tl and the UE can be served by UPF B at t2. The time difference |t2 - tl| is neglectable which is the transition delay when switching between UPFs happen. The gain is to reduce energy consumption.

[0115] In difference to this current solution, Fig. 6 illustrates a flow diagram of a first exemplary scheme according to the present disclosure. The first exemplary scheme may be referred to as exploration scheme. In step S202, the consumer node, in this case configured as a NF, transmits a prediction request to the MDAF (e.g., of the same CN 200, for example implemented by CN node 206).

[0116] In step S204, the MDAF analyzes the received request.

[0117] In step S206, the MDAF collects data items, also referred to as analytics data herein. The MDAF may obtain the analytics data from other entities of the communication network 100, for example from a NWDAF.

[0118] In step S208, the MDAF analyzes the collected data items to generate solutions for predictions and / or recommendations. This is also referred to as determining a plurality of recommendations herein.

[0119] The MDAF may update the consumer node periodically or once a new recommendation has been determined, as indicated by step S210.

[0120] The consumer node may transmit, in step S212, an indication to the MDAF that the exploration shall be stopped. In other words, the consumer node may instruct the MDAF to stop determining additional recommendations. It is also possible for the MDAF to only continue determining additional recommendations if requested explicitly with further request messages from the consumer node.

[0121] In step S214, the MDAF may reset the data collection and / or search context (e.g., goal-parameters considered when determining the recommendations).

[0122] In step S216, the MDAF may inform the consumer node that the reset is finished.

[0123] In step S218, the consumer node may collect additional information that is required of helpful for execution of the at least one action associated with a recommendation provided by the MDAF. For example, if the MDAF recommends to reconfigure the network to optimize energy usage, but this recommendation is based on a particular network load, the consumer node may collect information on the current network load from the NWDAF and then perform the reconfiguration accordingly.

[0124] An exemplary pseudo-code in accordance with the first exemplary scheme is given below. - J -

[0125] @startuml actor Operator participant NF participant MDAF participant Network

[0126] NF -> MDAF : prediction request with exploration loop exploration

[0127] MDAF -> MDAF : analyze the request information group search

[0128] MDAF -> Network : collect data items return

[0129] MDAF -> MDAF : generate solutions for predictions / recommendations end

[0130] MDAF -> NF : return next solution

[0131] NF -> MDAF : stop exploring

[0132] MDAF -> MDAF : reset data collection and search context

[0133] MDAF —> NF : done

[0134] NF -> Network : collect additional information for the execution context return

[0135] NF -> Network : execute proposed solution end @enduml

[0136] In difference to current solutions in which the MDAF only returns a single prediction, when asked to perform exploration of the prediction / recommendation space according to the first exemplary scheme, the MDAF determines multiple recommendations (e.g., predictions about traffic load trend of the UE accompanied by recommendations), for example:

[0137] 1. Recommendation with Cells A and B as best candidates, UPF A->B switch (same as in the legacy)

[0138] 2. (if requested to continue exploring) Recommendation with only Cell B, UPF B->A switch as the 2ndbest candidate 3. (if requested to continue exploring) Recommendation with only Cell A, optional UPF A->C switch as the 3rdbest candidate.

[0139] Fig. 7 illustrates a flow diagram of a second exemplary scheme according to the present disclosure. The second exemplary scheme may be referred to as guided exploration scheme. Also in this case, the consumer node may be configured as NF.

[0140] In step S302, an operator transmits configuration and domain knowledge to the MDAF.

[0141] In step S304, consumer node implementing the NF transmits, to the MDAF, a prediction request with an additional cost function, for example indicating goalspecific parameters to be met by the determining of the MDAF and / or the recommendations determined by the MDAF. For example, the cost function may express deadlines, energy requirements and resource constraints.

[0142] In step S306, the MDAF analyzes the request and the additional request information.

[0143] In step S308, the MDAF collects data items, also referred to as analytics data herein. The MDAF may obtain the analytics data from other entities of the communication network 100, for example from NWDAF.

[0144] In step S310, the MDAF analyzes the collected data items to generate solutions for predictions and / or recommendations.

[0145] In step S312, the MDAF reevaluates the determined predictions and / or recommendations against the additional information indicated in the request, in particular the cost function.

[0146] In step S314, valid solutions are filtered out of the previously determined plurality of solutions (e.g., recommendations and / or predictions). This filtering may be based on the additional information indicated in the request. The filtering may also include a validation of the solutions by checking the same using a digital twin.

[0147] In step S316, a message indicating one or more valid solutions are transmitted to the consumer node. In other words, the consumer node is provided with at least one of the plurality of recommendations determined by the MDAF. In step S318, the consumer node transmits a request to the MDAF to instruct the MDAF to stop exploring. Accordingly, the MDAF stops determining additional recommendations, reevaluating and filtering the same, but instead rests its data collection and optionally the search context (e.g., the constraints used for finding suitable solutions, including the cost function).

[0148] In step S322, the MDAF may inform the consumer node that the data collection has been reset.

[0149] In step S324, the consumer node may execute one or more of the proposed solutions, in particular perform the action(s) associated with the at least one recommendation provided by the MDAF in step S316.

[0150] An exemplary pseudo-code in accordance with the first exemplary scheme is given below.

[0151] @startuml actor Operator participant NF participant MDAF participant Network

[0152] Operator -> MDAF : configuration and domain knowledge NF -> MDAF : prediction request with additional cost function (expressing deadlines, energy, resources) loop exploration

[0153] MDAF -> MDAF : analyze the additional request information group search

[0154] MDAF -> Network : collect data items return

[0155] MDAF -> MDAF : generate solutions for predictions / recommendations MDAF -> MDAF : reevaluates solution against the internal metrics and specified cost function '

[0156] MDAF -> MDAF : filter valid solutions end

[0157] MDAF -> NF: return one or more valid soiution(s) NF -> MDAF : stop exploring

[0158] MDAF -> MDAF : reset data collection and search context

[0159] MDAF —> NF : done

[0160] NF -> Network : execute proposed solutions end @enduml

[0161] According to the second scheme, the consumer node (e.g., configured as 0AM) may submit a cost function that describes the latency in the communication, since the UE 402 has URLLC constraints. The constraint may be defined as hard and specify that the communication delay should not be greater than T. With respect to the cost function, the MDAF may return the following recommendation for energy saving:

[0162] 1. Only Cell A is a viable candidate to serve the UE because of low traffic and ability to provide high bandwidth.

[0163] 2. UPF A shall enter energy saving mode at time tl only if UPF C is available. UPF B shall not be used because transition delay may exceed time T. If UPF C is not available, UPF A should continue serving the UE and never enter the energy saving mode.

[0164] The exploration scheme can be considered to be a subset of the guided exploration scheme. To perform the guided exploration, the MDAF generates the prediction and / or recommendation with a relaxed requirement on time urgency to return a first solution. This allows for additional solutions which would otherwise not be accepted. The additional information provided to the MDAF may guide the selection of the methods to use for the prediction, e.g. based on their energy footprint or the time it takes to generate a solution. The underlying models may return a list of solutions ordered by some metric (e.g. prediction accuracy for the requested KPI). Both of these may then be taken into account when evaluating the cost function over the solution and exploring the solutions.

[0165] The example above is an energy efficiency use case that could be further extended with information about the quality of the supplied energy such as greenhouse gas emissions, types of energy sources (e.g., local solar panels), their availability now and in the future etc. Resource constraints may reflect resource availability including predictions for the future availability and optionally related costs. Another example is working with different time scales, where tasks have varying urgency and need to be completed in varying timescales from very short (e.g., seconds) to very long (e.g., weeks). Taking this into account, a different solution strategy may be taken for such varying tasks and can more easily take advantage of other constraints like energy and resources. The returned solutions in this case may reflect different outlined options and return solutions which are best for just one of the plurality of goals, for instance most energy efficient vs most time efficient.

[0166] To recap, in known solutions, the MDAF determines only a single recommendation, so the consumer node needs to perform additional processing in order to refine it if there are additional constraints to consider. This incurs an overhead in querying this additional information and might not even be possible or desirable from the interfaces available to the consumer node. With the proposal described herein, this time urgency is relaxed which allows for further exploration of the MDAF's analytics results and the possibility to explore further recommendations. The control of the exploration is placed with the MDAF which is responsible to produce additional results using the existing prediction models. In addition, the requests can serve as additional training inputs for combined models to reduce the inference time and produce approximate results directly instead of exploring multiple candidate results.

[0167] One benefit of the present technique can be seen when the consumer node specifies a cost function that the MDAF will use determine the recommendations and / or to evaluate the determined recommendations, and optionally a cost threshold to determine if a candidate recommendation is accepted. Such a cost function provides extra information about the requested analytics to the MDA request and allows to focus on energy or resource demands as well as time efficiency, for example.

[0168] While the present technique has been mostly described with reference to the MDAF, it is to be understood that it is not limited thereto but may be applied to other Functions such as NWDAF, for example. The technique disclosed herein has been described partly with reference to the 3GPP 5G standard, but it is to be understood that it is also applicable to other mobile telecommunication systems, for example according to a subsequent 3GPP standard (e.g., 6G). The "aspects" described herein above may be combined with features of the detailed description of the figures and vice versa. Various modifications to the present technique may be apparent to those skilled in the art. For example, the number and content of indications and request messages may be adapted (e.g., including multiple indications in a single request message, or transmitting these indications with separate request messages). The present disclosure also provides for the exemplary embodiment set forth below.

[0169] EXEMPLARY EMBODIMENT

[0170] A new use case for the consumer (e.g., the consumer node) of MDA (e.g., implemented by the DA node) is disclosed to provide a cost function when requesting for analytics. There is a current assumption for time urgency when requesting analytics that assist in completing a task. However when we remove the need for time urgency it allows for further analytics exploration and the possibility to provide multiple sets of recommended actions that take the cost function into consideration, allowing the consumer the choose which of the actions to take based on their individual needs.

[0171] Operators and vendors alike are constantly driving to reduce their energy consumption. This can be achieved in many ways. MDA currently, can provide to a consumer, predictions and / or recommended actions to take to fulfill a certain task e.g., energy saving in a network. Management Data Analytics currently does not address costs, in particular current energy analytics does not cover any aspects of energy cost or Greenhouse Gas Emissions of recommended actions.

[0172] According to this Exemplary Embodiment, it is proposed to enhance the current MDA solution so that the consumer could introduce some further information when making a request for analytics. This extra information could be for e.g., the availability of various energy sources at different times, or an indication of how urgent it is to complete the task. This could be called an "Energy Cost Function" but could be extended to be a general cost function for other MDA types e.g., Software Upgrade etc. The analytics producer could then provide several recommended actions, that take into consideration, the extra information provided by the consumer.

[0173] Example use case:

[0174] Assume that we have a request for prediction that will affect the decision to schedule a task that we need to assign to be executed on various entities. Tasks and entities may be very diverse - tasks may range from short to very long-time scales (transmissions of software update), an entity may be a single antenna or entire sites. The current assumption of urgency eliminates the opportunity to find other solutions that are more energy or resource-efficient but less time-efficient. This proposal would provide several sets of recommended actions for the consumer to choose between, for example:

[0175] Recommended Action set 1 - most time efficient

[0176] Recommended Action set 2 - most energy efficient

[0177] Recommended Action set 3 - most resource efficient

[0178] It may be that the consumer of analytics is not under any strict time constraints to complete the task at hand so may take the recommendation for the most energy efficient way forward. This could be as simple as executing the task at nighttime when energy costs are lower.

[0179] Potential Requirements

[0180] REQ-COST-FUNC-1 - The producer of analytics should be able to allow an authorized consumer to provide a cost function which provides extra information about the task to be completed.

[0181] NOTE: Instead of, or in addition to a cost function, a producer may provide the usage of time, energy and resources as part of each proposal, or only the cost type of interest to the consumer. In this case, the producer does not need to know anything about the preference of the consumer.

[0182] REQ-COST-FUNC-2 - The producer of analytics should be able to provide different sets of recommended actions based on different criteria i.e., time efficiency, energy efficiency, resource efficiency.

[0183] Possible Solutions

[0184] It is proposed to enhance the current MDAS solution as specified in TS 28.104 [2]:

[0185] • Add a cost function attribute to the existing MDARequest to allow the consumer to add supplementary information on the task for which the analytics is being requested.

[0186] • Extend the MDAS reporting to give several sets of recommended actions for the consumer to choose from. One possible implementation of this exemplary embodiment is also set out herein above (see e.g. Table 1).

[0187] REFERENCES

[0188] [1] 3GPP TR 22.882, Study on Energy Efficiency as a service criteria, Rel.19

[0189] [2] 3GPP TS 28.104, Management and orchestration; Management Data Analytics

[0190] (MDA), Rel.18

[0191] [3] 3GPP TS 28.809, Management and orchestration; Study on enhancement of

[0192] Management Data Analytics (MDA) Rel.17

[0193] [4] 3GPP TS 28.552, Management and orchestration; 5G performance measurements, Rel.18

[0194] [5] 3GPP TS 23.288, Architecture enhancements for 5G System (5GS) to support network data analytics services, Rel.18

[0195] [6] 3GPP TS 32.422, Telecommunication management; Subscriber and equipment trace; Trace control and configuration management, Rel. 18

[0196] [7] 3GPP TS 32.423, Telecommunication management; Subscriber and equipment trace; Trace data definition and management, Rel. 19

[0197] [8] 3GPP TS 28.406, Telecommunication management; Quality of Experience

[0198] (QoE) measurement collection; Information definition and transport, Rel. 18

[0199] [9] 3GPP TS 28.541, Management and orchestration; 5G Network Resource Model

[0200] (NRM); Stage 2 and stage 3, Rel. 19

[0201]

[0010] 3GPP TS 23.288, Architecture enhancements for 5G System (5GS) to support network data analytics services, Rel. 18

[0202]

[0011] 3GPP TS 23.501, System architecture for the 5G System (5GS), Rel. 19

[0203]

[0012] 3GPP TS 23.502, Procedures for the 5G System (5GS), Rel. 19

Claims

- 36 -CLAIMS1. A method performed by a Data Analytics, DA, node configured to communicate with a consumer node associated with a mobile telecommunication system, the method comprising : determining, based on analytics input data and based on one or more predefined goals, a plurality of recommendations, each recommendation being associated with at least one action to take, wherein the plurality of recommendations differ from one another at least in the respective degrees to which they fulfil the one or more predefined goals; and providing at least one of the determined plurality of recommendations to the consumer node.

2. The method of Claim 1, wherein each recommendation is optimized for a different goal or for a different weighting of the one or more predefined goals.

3. The method of Claim 1 or 2, wherein the one or more predefined goals include(s) at least one of: an energy efficiency; a resource efficiency; a time efficiency; a Quality of Service, QoS.

4. The method of Claim 3, wherein the energy efficiency is defined by at least one energy efficiency parameter selected from: an energy consumption; an amount of greenhouse gas emissions; a percentage of renewable energy; a selection or number of computing resources using renewable energy and / or emitting a predefined maximum amount of greenhouse gases; a use of energy in a time period at which renewable energy is available.

5. The method of Claim 3 or 4, wherein the resource efficiency is defined by at least one resource efficiency parameter selected from: a computing power; a data storage amount; an amount of redistribution of computing resources; a usage of one or more network functions.

6. The method of any one of Claims 3 to 5, wherein the time efficiency is defined by at least one time efficiency parameter selected from: a required time; a delay.

7. The method of any one of Claims 3 to 6, wherein the QoS is defined by at least one QoS parameter selected from: a minimum data transmission rate, a maximum communication latency, a minimum communication security.- 37 -8. The method of any one of Claims 1 to 7, wherein the respective degrees to which the recommendations fulfil the one or more goals are indicative of or correspond to the degrees to which the associated actions fulfil the one or more goals and / or the parameter(s) defining said one or more goals.

9. The method of Claim 8, further comprising: selecting at least one of a plurality of analytic models, each being associated with a different degree to which the recommendations determined by said model fulfil the one or more goals and / or the parameter(s) defining said one or more goals, wherein the plurality of recommendations are determined based on the selected at least one analytic model.

10. The method of any one of Claims 1 to 9, wherein the respective degrees to which the recommendations fulfil the one or more goals are indicative of or correspond to the degrees to which the determination of the respective recommendations by the DA node fulfils the one or more goals and / or the parameter(s) defining said one or more goals.

11. The method of Claim 10, further comprising: selecting at least one of a plurality of analytic models, each being associated with a different degree to which the determination by said model fulfils the one or more goals and / or the parameter(s) defining said one or more goals, wherein the plurality of recommendations are determined based on the selected at least one analytic model.

12. The method of any one of Claims 1 to 11, further comprising: obtaining an indication of the one or more predefined goals, wherein the plurality of recommendations differ from one another in the respective degrees to which they fulfil these indicated one or more predefined goals.

13. The method of Claim 12 and one or more of Claims 4 to 7, wherein the indication of the one or more predefined goals includes an indication of the at least one parameter defining the one or more predefined goals.

14. The method of Claim 12 or 13, wherein the indication of the one or more predefined goals includes a cost function.

15. The method of any one of Claims 1 to 14, further comprising: obtaining an indication of a prioritization or selection of potential recommendations, wherein the at least one recommendation is provided to the consumer node based on the indication of the prioritization or selection.

16. The method of any one of Claims 1 to 15, further comprising: obtaining an indication to pause or stop the determination of recommendations; pausing or stopping the determination of the plurality of recommendations in response to obtaining said indication to pause or stop the determination of recommendations, wherein the at least one recommendation provided to the consumer node is part of the recommendation(s) determined before pausing or stopping the determination of the plurality of recommendations.

17. The method of any one of Claims 1 to 16, further comprising: obtaining an indication of at least one trigger condition for providing the at least one recommendation to the consumer node; determining whether the at least one trigger condition is met; and in response to determining that the at least one trigger condition is met, providing the at least one recommendation to the consumer node.

18. The method of Claim 17, wherein the at least one trigger condition relates to a state of the mobile telecommunication system, for example a state of a core network, CN, and / or a state of a radio access network, RAN, of the mobile telecommunication system.

19. The method of any one of Claims 12 to 18, wherein the indication(s) are defined by and / or obtained from a policy associated with the consumer node and known to the DA node.

20. The method of Claim 19, further comprising: based on the indication(s) that were obtained, deriving and / or learning the policy associated with the consumer node for a subsequent provision of recommendations.

21. The method of any one of Claims 12 to 20, further comprising:receiving one or more messages from the consumer node, the one or more messages {i} requesting to provide at least one recommendation, {ii} containing the indication(s), {iii} containing the policy and / or {iv} requesting associating the policy with the consumer node.

22. The method of any one of Claims 1 to 21, further comprising: requesting one or more network nodes associated with the mobile telecommunications system to collect the analytics input data such that at least one data collection condition is fulfilled.

23. The method of Claim 22, wherein the at least one data collection condition comprises a sampling rate, a number of parameters to be monitored and / or a data collection time.

24. The method of Claim 22 or 23, wherein the one or more network nodes comprise a network node configured as a Network Function, NF, of the mobile telecommunication system, for example a NF of the core network, CN, of the mobile telecommunication system.

25. The method of any one of Claims 1 to 24, further comprising: validating one or more of the determined recommendations, wherein the at least one recommendation provided to the consumer node is a validated recommendation.

26. The method of Claim 25, wherein validating a recommendation comprises applying the at least one action associated with said recommendation to a digital twin network.

27. The method of any one of Claims 1 to 26, wherein the at least one action to take comprises or consists of at least one action to take by the consumer node or by another network node configured as a Network Function, NF, of the mobile telecommunication system .

28. The method of any one of Claims 1 to 27, wherein {i} the DA node is configured as Management Data Analytics Function, MDAF, or as Network Data Analytics Function, NWDAF and / or {ii} the consumer node is configured as a Network Function, NF, of the mobile telecommunication system.

29. A Data Analytics, DA, node configured to communicate with a consumer node associated with a mobile telecommunication system, the DA node being configured to: determine, based on analytics input data and based on one or more predefined goals, a plurality of recommendations, each recommendation being associated with at least one action to take, wherein the plurality of recommendations differ from one another at least in the respective degrees to which they fulfil the one or more predefined goals; and provide at least one of the determined plurality of recommendations to the consumer node.

30. The DA node of Claim 29, further configured to perform the method according to any one of Claims 2 to 28.

31. A method performed by a consumer node associated with a mobile telecommunication system and configured to communicate with a Data Analytics, DA, node the method comprising: obtaining, from the DA node, at least one of a plurality of recommendations, the plurality of recommendations being determined by the DA node based on one or more predefined goals, each recommendation being associated with at least one action to take, wherein the plurality of recommendations differ from one another at least in the respective degrees to which they fulfil the one or more predefined goals.

32. The method of Claim 31, wherein each recommendation is optimized for a different goal or for a different weighting of the one or more predefined goals.

33. The method of Claim 31 or 32, wherein the one or more predefined goals include(s) at least one of: an energy efficiency; a resource efficiency; a time efficiency; a Quality of Service, QoS.

34. The method of Claim 33, wherein the energy efficiency is defined by at least one energy efficiency parameter selected from: an energy consumption; an amount of greenhouse gas emissions; a percentage of renewable energy; a selection or number of computing resources using renewable energy and / or emitting a predefined maximum amount of greenhouse gases; a use of energy in a time period at which renewable energy is available.- 41 -35. The method of Claim 33 or 34, wherein the resource efficiency is defined by at least one resource efficiency parameter selected from: a computing power; a data storage amount; an amount of redistribution of computing resources; a usage of one or more network functions.

36. The method of any one of Claims 33 to 35, wherein the time efficiency is defined by at least one time efficiency parameter selected from: a required time; a delay.

37. The method of any one of Claims 33 to 36, wherein the QoS is defined by at least one QoS parameter selected from: a minimum data transmission rate, a maximum communication latency, a minimum communication security.

38. The method of any one of Claims 31 to 37, wherein the respective degrees to which the recommendations fulfil the one or more goals are indicative of or correspond to the degrees to which the associated actions fulfil the one or more goals and / or the parameter(s) defining said one or more goals.

39. The method of Claim 38, wherein the plurality of recommendations are determined by the DA node based on at least one analytic model selected from a plurality of analytic models, each being associated with a different degree to which the recommendations determined by said model fulfil the one or more goals and / or the parameter(s) defining said one or more goals.

40. The method of any one of Claims 31 to 39, wherein the respective degrees to which the recommendations fulfil the one or more goals are indicative of or correspond to the degrees to which the determination of the respective recommendations by the DA node fulfils the one or more goals and / or the parameter(s) defining said one or more goals.

41. The method of Claim 40, wherein the plurality of recommendations are determined by the DA node based on at least one analytic model selected from a plurality of analytic models, each being associated with a different degree to which the determination by said model fulfils the one or more goals and / or the parameter(s) defining said one or more goals.

42. The method of any one of Claims 31 to 41, further comprising:- 42 - providing, to the DA node, an indication of the one or more predefined goals, wherein the plurality of recommendations differ from one another in the respective degrees to which they fulfil these indicated one or more predefined goals.

43. The method of Claim 42 and one or more of Claims 34 to 37, wherein the indication of the one or more predefined goals includes an indication of the at least one parameter defining the one or more predefined goals.

44. The method of Claim 42 or 43, wherein the indication of the one or more predefined goals includes a cost function.

45. The method of any one of Claims 31 to 44, further comprising: providing, to the DA node, an indication of a prioritization or selection of potential recommendations, wherein the at least one recommendation obtained from the DA node is based on the indication of the prioritization or selection.

46. The method of any one of Claims 31 to 45, further comprising: providing, to the DA node, an indication to pause or stop the determination of recommendations, said indication causing the DA node to pause or stop the determination of the plurality of recommendations, wherein the at least one recommendation obtained from the DA node is part of the recommendation(s) determined by the DA node before the determination of the plurality of recommendations is paused or stopped.

47. The method of any one of Claims 31 to 46, further comprising: providing, to the DA node, an indication of at least one trigger condition for providing the at least one recommendation to the consumer node, said indication causing the DA node to determine whether the at least one trigger condition is met and to provide the at least one recommendation to the consumer node in response to determining that the at least one trigger condition is met.

48. The method of Claim 47, wherein the at least one trigger condition relates to a state of the mobile telecommunication system, for example a state of a core network, CN, and / or a state of a radio access network, RAN, of the mobile telecommunication system.- 43 -49. The method of any one of Claims 42 to 48, wherein the indication(s) are defined by and / or provided via a policy associated with the consumer node and known to the DA node.

50. The method of Claim 49, further comprising: requesting the DA node to derive and / or learn the policy associated with the consumer node based on the provided indication(s) for a subsequent provision of recommendations.

51. The method of any one of Claims 42 to 50, further comprising: transmitting one or more messages to the DA node, the one or more messages {i} requesting to provide at least one recommendation, {ii} containing the indication(s), {iii} containing the policy and / or {iv} requesting associating the policy with the consumer node.

52. The method of any one of Claims 31 to 51, further comprising: requesting the DA node to request one or more network nodes associated with the mobile telecommunication system to collect the analytics input data such that at least one data collection condition is fulfilled.

53. The method of Claim 52, wherein the at least one data collection condition comprises a sampling rate, a number of parameters to be monitored and / or a data collection time.

54. The method of Claim 52 or 53, wherein the one or more network nodes comprise a network node configured as a Network Function, NF, of the mobile telecommunication system, for example a NF of the core network, CN, of the mobile telecommunication system.

55. The method of any one of Claims 31 to 54, further comprising: requesting the DA node to validate one or more of the determined recommendations, wherein the at least one recommendation obtained from the DA node is a validated recommendation.

56. The method of Claim 55, wherein validating a recommendation comprises applying the at least one action associated with said recommendation to a digital twin network.- 44 -57. The method of any one of Claims 31 to 56, wherein the at least one action to take comprises or consists of at least one action to take by the consumer node or by another network node configured as a Network Function, NF, of the mobile telecommunication system.

58. The method of any one of Claims 31 to 57, wherein {i} the DA node is configured as Management Data Analytics Function, MDAF, or as Network Data Analytics Function, NWDAF and / or {ii} the consumer node is configured as a Network Function, NF, of the mobile telecommunication system.

59. A consumer node associated with a mobile telecommunication system and configured to communicate with a Data Analytics, DA, node the consumer node being configured to: obtain, from the DA node, at least one of a plurality of recommendations, the plurality of recommendations being determined by the DA node based on one or more predefined goals, each recommendation being associated with at least one action to take, wherein the plurality of recommendations differ from one another at least in the respective degrees to which they fulfil the one or more predefined goals.

60. The consumer node of Claim 59, further configured to perform the method of any one of Claims 32 to 58.

61. A computer program, comprising instructions that, when executed by processing circuitry, cause the processing circuitry to carry out the method according to any of Claims 1 to 28 or 31 to 58.

62. A carrier containing the computer program of Claim 61, wherein the carrier is one of an electronic signal, optical signal, radio signal, or a computer-readable medium.

Images