Security for AI / ML model storage and sharing

Abstract

Embodiments include methods of a consumer network function (NFc) of a communication network. The method includes sending a first request for a first access token associated with a machine learning (ML) model to a first NF of a communication network. The first request includes at least one of an analysis identifier (ID) and an interoperability ID associated with the ML model. The method includes receiving a first response including a first access token from the first NF and sending a second request for the ML model to a producer NF (NFp) of the communication network. The second request includes the first access token and at least one of the analysis ID and the interoperability ID. The method includes receiving, from the NFp, a second response including one or more of: an ML model; an identifier of the ML model; and an address of a storage resource associated with a second NF of the communication network from which the ML model can be obtained.

Description

[0001] This application is a divisional application of PCT international application PCT / EP2023 / 076517 filed on September 26, 2023, entitled "Security of AI / ML Model Storage and Sharing", which has entered the Chinese national phase patent application 202380068353.2. Technical Field

[0002] This application generally relates to the field of communication networks, and more specifically to technologies for protecting artificial intelligence / machine learning (AI / ML) models used to generate analyses in communication networks (e.g., 5G core networks). Background Technology

[0003] Currently, fifth-generation (5G) cellular systems are being standardized within the 3rd Generation Partnership Project (3GPP). NR is being developed to achieve maximum flexibility to support a wide variety of distinct use cases. These use cases include enhanced mobile broadband (eMBB), machine-type communication (MTC), ultra-reliable low-latency communication (URLLC), sidelink device-to-device (D2D), and several others.

[0004] At a high level, a 5G system (5GS) consists of an access network (AN) and a core network (CN). The AN provides UE connectivity to the CN, for example, via base stations such as gNBs or ng-eNBs. The CN includes various network functions (NFs) that provide a wide range of different functions, such as session management, connection management, accounting, authentication, etc.

[0005] Figure 1 A high-level view of an exemplary 5G network architecture is shown, consisting of a Next-Generation Radio Access Network (NG-RAN, 199) and a 5G core (5GC, 198). The NG-RAN may include one or more gNodeBs (gNBs) connected to the 5GC via one or more NG interfaces, such as gNBs (100, 150) connected via corresponding interfaces (102, 152). More specifically, gNBs may connect to one or more Access and Mobility Management Functions (AMFs) in the 5GC via corresponding NG-C interfaces and to one or more User Plane Functions (UPFs) in the 5GC via corresponding NG-U interfaces. The 5GC may include various other network functions (NFs), such as Session Management Functions (SMFs).

[0006] Furthermore, gNBs can connect to each other via one or more Xn interfaces (e.g., Xn interface (140) between gNBs (100, 150)). The radio technology used for NG-RAN is generally referred to as "New Radio" (NR). Regarding the NR interface to the UE, each gNB can support Frequency Division Duplex (FDD), Time Division Duplex (TDD), or a combination thereof. Each gNB can provide service for a geographic coverage area comprising one or more cells, and in some cases, various directional beams can also be used to provide coverage within the respective cells. Generally, a DL "beam" is the coverage area of ​​a reference signal (RS) transmitted by the network that can be measured or monitored by the UE.

[0007] NG-RAN 199 is layered into the Radio Network Layer (RNL) and the Transport Network Layer (TNL). The NG-RAN architecture (i.e., NG-RAN logical nodes and the interfaces between them) is defined as part of the RNL. For each NG-RAN interface (NG, Xn, F1), the associated TNL protocols and functions are specified. The TNL provides services for user plane transmission and signaling transmission.

[0008] An NG RAN logical node (e.g., gNB 100) comprises a central unit (CU or gNB-CU, e.g., 110) and one or more distributed units (DUs or gNB-DUs, e.g., 120, 130). A CU is a logical node that hosts higher-level protocols and performs various gNB functions (e.g., controlling the operation of the DU). A DU is a distributed logical node that hosts lower-level protocols and may include various subsets of gNB functions depending on the function partitioning options. Each CU and DU may include various circuitry required to perform its respective functions, including processing circuitry, communication interface circuitry (e.g., transceiver), and power supply circuitry.

[0009] gNB-CU communicates via the corresponding F1 logic interface (e.g., Figure 1 Interfaces 122 and 132 shown are connected to one or more gNB-DUs. However, a gNB-DU can be connected to only a single gNB-CU. The gNB-CU and its connected gNB-DUs are only visible to other gNBs and 5GCs as gNBs. In other words, the F1 interface is not visible outside of the gNB-CU.

[0010] Another change in 5G networks (e.g., 5GC) is the modification and / or replacement of traditional peer-to-peer interfaces and protocols from previous generations of networks using a service-based architecture (SBA). In SBA, network functions (NFs) provide one or more services to one or more service consumers. Typically, the various services are self-contained functions that can be changed and modified in isolation without affecting other services. Furthermore, these services include various "service operations," which are finer-grained divisions of the overall service functionality. Interactions between service consumers and producers can be of the "request / response" or "subscription / notification" type.

[0011] Of particular interest in this disclosure is the 5GC NF (Network Data Analysis Function) (NWDAF). This NF provides network analysis information (e.g., statistics and / or predictive information of past events) to other NFs at the network slice instance level. The NWDAF can collect data from any 5GC NF. Note that a "network slice" is a logical partition of a 5G network that provides specific network capabilities and characteristics, such as supporting specific services. A network slice instance is a collection of NF instances and the network resources (e.g., computing, storage, communication) required to provide the capabilities and characteristics of a network slice.

[0012] Machine learning (ML) is a type of artificial intelligence (AI) that focuses on using data and algorithms to mimic how humans learn, gradually improving accuracy as more data becomes available. ML algorithms build models based on sample (or "training") data, which are then used to make predictions or decisions. ML algorithms can be used in a variety of applications (such as medicine, email filtering, speech recognition, etc.) where developing conventional algorithms to perform the desired tasks is difficult or infeasible. A subset of ML is closely related to computational statistics.

[0013] The 5G system architecture allows any NF to obtain analytical data from the NWDAF using the Data Collection Coordination Function (DCCF) and associated NdCCCF services. The NWDAF can also store and retrieve analytical information from the Analytical Data Repository Function (ADRF). 3GPP TS 23.288 (v17.2.0) specifies the NWDAF as the primary NF for calculating analytical reports and classifies the NWDAF into two sub-functions (or logical functions): the Analysis Logic Function (AnLF), which performs the analytical process; and the Model Training Logic Function (MTLF), which performs the training and retraining of the ML model used by the AnLF. Summary of the Invention

[0014] AI / ML models (or more simply, ML models) are generally considered important intellectual property of their owners (e.g., 5GC vendors), and therefore require constant protection of their confidentiality and integrity. 3GPP is investigating the feasibility of sharing or storing ML models across network equipment that may be supplied by different vendors. Under this arrangement, the ML model should be protected against access and use by consumer NFs supplied by vendors different from the ML model itself. However, no specific solution currently exists for this requirement.

[0015] The purpose of embodiments of this disclosure is to address these and related problems, challenges and / or difficulties, thereby facilitating other advantageous deployments of ML models for network analysis.

[0016] Some embodiments of this disclosure include methods (e.g., processes) for consumer NF (NFc) in a communication network (e.g., 5GC).

[0017] These exemplary methods include sending a first request to a first NF of a communication network for a first access token associated with an ML model. The first request includes one or more of the following associated with the ML model: an analytics ID and an interoperability ID. These exemplary methods also include receiving a first response from the first NF including the first access token. The exemplary method may further include sending a second request to a producer NF (NFp) of the communication network for the ML model. The second request includes the first access token, and at least one of an analytics ID and an interoperability ID. These exemplary methods also include receiving a second response from the NFp including one or more of the following: the ML model; an identifier for the ML model; and an address of a storage resource associated with a second NF of the communication network from which the ML model can be obtained.

[0018] In some embodiments, the first NF is a Network Repository Function (NRF). In other embodiments, the first NF is an Analytics Data Repository Function (ADRF). In some embodiments, the second NF is NFp. In other embodiments, the second NF is ADRF. In some embodiments, one or more of the following are applicable: NFc is NWDAF (AnLF), and NFp is NWDAF (MTLF).

[0019] Other embodiments include exemplary methods (e.g., processes) for NFp in communication networks (e.g., 5GC).

[0020] These exemplary methods include registering information associated with an ML model in the NRF of the communication network. The ML model is generated, owned, and / or maintained by the NFp. The registration information associated with the ML model includes an analytics ID and an interoperability ID. These exemplary methods also include encrypting the ML model and sending a first request to the ADRF of the communication network for storing the encrypted ML model. The first request includes a first address of the encrypted ML model or a storage resource associated with the NFp from which the ML model can be obtained.

[0021] In some embodiments, these exemplary methods may further include receiving a second request for an ML model from an NFc of a communication network. The second request includes a first access token and at least one of an analytics ID and an interoperability ID. These exemplary methods may further include sending a second response to the NFc based on verification of the first access token, the second response including one or more of the following: the ML model; an identifier for the ML model; a first address of a storage resource associated with an NFp; and a second address of a storage resource associated with an ADRF from which the ML model can be obtained.

[0022] In some embodiments, one or more of the following applies: NFc is NWDAF (AnLF) and NFp is NWDAF (MTLF).

[0023] Other embodiments include methods (e.g., procedures) for NRF used in communication networks (e.g., 5GC).

[0024] These exemplary methods may include registering information associated with an ML model generated, owned, and / or maintained by an NFp of the communication network. The registration information associated with the ML model includes an analytics ID and an interoperability ID. These exemplary methods may also include receiving a first request from an NFc of the communication network for a first access token associated with the ML model. The first request includes at least one of the analytics ID and the interoperability ID. These exemplary methods may also include sending a first response to the NFc including the first access token.

[0025] In some embodiments, these exemplary methods may further include receiving a second request for a second access token from a first NF of the communication network. The second request includes at least one of an analysis ID and an interoperability ID, and one of the following:

[0026] - The first address of the storage resource associated with NFp, from which the ML model can be obtained; or

[0027] - A second address of the storage resource associated with the ADRF of the communication network, from which the ML model can be obtained.

[0028] These exemplary methods may also include sending a second response, including a second access token, to the first NF.

[0029] In some embodiments of these examples, the first address of the storage resource associated with NFp is a first generic resource locator (URL), and the second address of the storage resource associated with ADRF is a second URL or fully qualified domain name (FQDN). In some embodiments of these examples, the first NF is NFc. In other embodiments of these examples, the first NF is NFp.

[0030] Other embodiments include methods (e.g., procedures) for ADRF used in communication networks (e.g., 5GC).

[0031] These exemplary methods may include receiving a first request from an NFp in a communication network for storing an encrypted ML model. The first request includes the encrypted ML model or a first address of a storage resource associated with the NFp, from which the encrypted ML model can be obtained. These exemplary methods may also include storing the encrypted ML model in a storage resource associated with an ADRF. These exemplary methods may further include sending a first response to the NFp, the response including a second address of a storage resource associated with the ADRF.

[0032] In some embodiments, the first NF is an NFp. In other embodiments, the first NF is an NFc of the communication network. In some embodiments, the NFc is an NWDAF (AnLF) and / or the NFp is an NWDAF (MTLF). In some embodiments, the first address of the storage resource associated with the NFp is a first URL, and the second address of the storage resource associated with the ADRF is a second URL or FQDN.

[0033] Other embodiments include NFc, NFp, NRF, and ADRF (or network nodes hosting such NFs) configured to perform operations corresponding to any of the exemplary methods described herein. Other embodiments also include a non-transitory computer-readable medium storing computer-executable instructions that, when executed by processing circuitry, configure such a network node or NF to perform operations corresponding to any of the exemplary methods described herein.

[0034] These and other disclosed embodiments can provide various benefits and / or advantages. By providing ML model owners / producers with the ability to protect ML models during various transmission, storage, and retrieval scenarios, the embodiments improve the security of confidential and / or sensitive ML models, thereby facilitating the deployment of such models in multi-vendor communication networks (e.g., 5GC).

[0035] These and other objects, features, and advantages of this disclosure will become apparent when read in conjunction with the accompanying drawings, which are briefly described below. Attached Figure Description

[0036] Figures 1 to 2 Various aspects of an exemplary 5G network architecture are shown.

[0037] Figure 3 A signaling diagram is shown for the network process used to authorize and authenticate the transmission of AI / ML models.

[0038] Figure 4 Signaling diagrams of processes involving NWDAF (AnLF), NRF, NWDAF (MTLF), and ADRF according to some embodiments of this disclosure are shown.

[0039] Figure 5 (which includes) Figures 5A to 5B The diagram illustrates a signaling diagram of another process involving NWDAF (AnLF), NRF, NWDAF (MTLF), and ADRF according to other embodiments of this disclosure.

[0040] Figure 6 Exemplary methods (e.g., processes) for a consumer NF in a communication network are illustrated according to various embodiments of the present disclosure.

[0041] Figure 7 Exemplary methods (e.g., processes) for a producer NF in a communication network according to various embodiments of this disclosure are shown.

[0042] Figure 8 Exemplary methods (e.g., processes) for NRF in a communication network according to various embodiments of this disclosure are shown.

[0043] Figure 9 Exemplary methods (e.g., processes) for ADRF for communication networks according to various embodiments of this disclosure are shown.

[0044] Figure 10 Communication systems according to various embodiments of the present disclosure are shown.

[0045] Figure 11 A UE according to various embodiments of the present disclosure is shown.

[0046] Figure 12 Network nodes according to various embodiments of this disclosure are shown.

[0047] Figure 13 A host computing system according to various embodiments of the present disclosure is shown.

[0048] Figure 14This is a block diagram of a virtualized environment that can virtualize the functions implemented by some embodiments of this disclosure.

[0049] Figure 15 Communication between a host computing system, a network node, and a UE via multiple connections is illustrated according to various embodiments of the present disclosure. Detailed Implementation

[0050] The embodiments briefly summarized above will now be described more fully with reference to the accompanying drawings. These descriptions are provided by way of example to illustrate the subject matter to those skilled in the art and should not be construed as limiting the scope of the subject matter to the embodiments described herein. More specifically, examples illustrating operation of various embodiments according to the advantages described above are provided below.

[0051] Generally, unless a different meaning is explicitly defined and / or implied in the context of use, all terms used herein are to be interpreted according to their ordinary meaning to those skilled in the art. Unless otherwise expressly stated or clearly implied from the context of use, all references to “an / element, device, component, apparatus, step, etc.” should be openly interpreted as referring to at least one instance of an element, device, component, apparatus, step, etc. Unless it must be explicitly stated that an operation is described as occurring after or before another operation and / or implicitly implied that an operation must occur after or before another operation, the operation of any methods and / or processes disclosed herein need not be performed in the exact order disclosed. Where appropriate, any feature of any embodiment disclosed herein may be applied to any other disclosed embodiment. Similarly, where appropriate, any advantage of any embodiment described herein may be applied to any other disclosed embodiment.

[0052] In addition, the following terms are used throughout the description given below:

[0053] - Radio Access Node: As used herein, a “radio access node” (or equivalently a “radio network node,” “radio access network node,” or “RAN node”) can be any node in a radio access network (RAN) used for wirelessly transmitting and / or receiving signals. Some examples of radio access nodes include, but are not limited to, base stations (e.g., gNBs in 3GPP 5G / NR networks or enhanced NodeBs or eNBs in 3GPP LTE networks), base station distributed components (e.g., CUs and DUs), high-power or macro base stations, low-power base stations (e.g., micro, pico, femto, or femto base stations), integrated access backhaul (IAB) nodes, transport points (TPs), transport receiver points (TRPs), remote radio units (RRUs or RRHs), and relay nodes.

[0054] - Core Network Nodes: As used in this document, a “core network node” is any type of node in the core network. Some examples of core network nodes include, for example, Mobility Management Entity (MME), Serving Gateway (SGW), PDN Gateway (P-GW), Policy and Charging Rules Function (PCRF), Access and Mobility Management Function (AMF), Session Management Function (SMF), User Plane Function (UPF), Charging Function (CHF), Policy Control Function (PCF), Authentication Server Function (AUSF), Location Management Function (LMF), etc.

[0055] - Wireless Device: As used herein, a “wireless device” (or “WD” for short) is any type of device capable of, configured, positioned, and / or operable to wirelessly communicate with network nodes and / or other wireless devices. Wireless communication may include transmitting and / or receiving wireless signals using electromagnetic waves, radio waves, infrared waves, and / or other types of signals suitable for transmitting information through the air. Unless otherwise stated, the term “wireless device” is used interchangeably herein with the term “user equipment” (“UE” for short), which have different meanings than the term “network node”.

[0056] - Radio Node: As used herein, "radio node" can be "radio access node" (or equivalent term) or "wireless device".

[0057] - Network Node: As used herein, a “network node” is any node that is part of the radio access network (e.g., a radio access node or equivalent term) or the core network (e.g., the core network node discussed above) of a cellular communication network. Functionally, a network node is a device that is capable of, configured to, arranged to, and / or operable to communicate directly or indirectly with wireless devices and / or with other network nodes or devices in the cellular communication network to enable and / or provide radio access to the wireless devices, and / or perform other functions (e.g., management) in the cellular communication network.

[0058] - Node: As used herein, the term "node" (without prefix) can be any type of node capable of operating in or with a wireless network (including RAN and / or core network), including radio access nodes (or equivalent terms), core network nodes, or wireless devices. However, the term "node" may be limited to a particular type (e.g., radio access node, IAB node) based on its specific characteristics in any given context.

[0059] The above definitions are not exclusive. In other words, the various terms used above may be interpreted and / or described elsewhere in this disclosure using the same or similar terms. However, if any other interpretation and / or description conflicts with the above definitions, the above definitions shall prevail.

[0060] Note that the descriptions presented herein focus on 3GPP cellular communication systems, and therefore frequently use the terminology of 3GPP LTE or similar terms. However, the concepts disclosed herein are not limited to 3GPP systems and can be applied to any communication system from which it can benefit. Furthermore, although the term "cell" is used herein, it should be understood that (particularly for 5G NR) a beam can be used instead of a cell, and therefore the concepts described herein apply equally to both cells and beams.

[0061] Figure 2 An exemplary architecture for 5GC (200) with service-based interfaces and various NFs defined by 3GPP is shown within the control plane (CP). This includes the following items:

[0062] - Application Functions (AF, with a Naf interface) interact with the 5GC to provide information to network operators and subscribe to specific events occurring within the operator's network. AF provides the ability to deliver services at a different layer (i.e., the transport layer) than the layer where services have already been requested (i.e., the signaling layer), controlling flow resources based on content already negotiated with the network. AF transmits dynamic session information to the PCF (via the N5 interface), including a description of the media to be transmitted by the transport layer.

[0063] - The Policy Control Function (PCF, with an NPCF interface) supports a unified policy framework for managing network behavior by providing PCC rules (e.g., regarding the processing of each service data flow under PCC control) to the SMF via the N7 reference point. The PCF provides the SMF with policy control decisions and flow-based charging control (including service data flow detection, gating, and QoS) and flow-based charging (in addition to credit management). The PCF receives session and media-related information from the AF and notifies the AF of service (or user) plane events.

[0064] - User Plane Function (UPF) – Supports processing user plane traffic based on rules received from the SMF, including packet inspection and various execution actions (e.g., event detection and reporting). The UPF communicates with the RAN (e.g., NG-RNA) via reference point N3, with the SMF (discussed below) via reference point N4, and with the external Packet Data Network (PDN) via reference point N6. Reference point N9 is used for communication between two UPFs.

[0065] - Session Management Function (SMF, with Nsmf interface) interacts with the decoupled business (or user) plane, including creating, updating, and deleting Protocol Data Unit (PDU) sessions and managing session contexts using User Plane Functions (UPF), such as for event reporting. For example, SMF performs data flow inspection (based on filter definitions included in PCC rules), online and offline billing interactions, and policy enforcement.

[0066] - The Billing Function (CHF, with an Nchf interface) is responsible for integrating online and offline billing functions. The Billing Function provides quota management (for online billing), reauthorization triggers, rating conditions, etc., and is notified of usage reports from the SMF. Quota management involves granting a specific number of units (e.g., bytes, seconds) to a service. The CHF also interacts with the billing system.

[0067] - The Access and Mobility Management Function (AMF, with a Namf interface) terminates the RAN CP interface and handles all mobility and connectivity management for the UE (similar to the MME in the EPC). The AMF communicates with the UE via the N1 reference point and with the RAN (e.g., NG-RAN) via the N2 reference point.

[0068] - Network Open Function (NEF) with Nnef interface – acts as an entry point to the operator network by securely opening the capabilities and events of the 5GC NF to AFs inside and outside the 5GC, and by providing AFs with a secure way to provide information to the 3GPP network. For example, NEF provides services that allow AFs to provide specific subscription data (e.g., expected UE behavior) for various UEs.

[0069] - Network repository functionality with Nnrf interface (NRF, 220) - provides service registration and discovery, enabling NFs to identify appropriate services available from other NFs.

[0070] - Network Slice Selection Function (NSSF) with Nnssf Interface – A “network slice” is a logical partition of a 5G network that provides specific network capabilities and features, such as supporting specific services. A network slice instance is a collection of NF instances and the network resources (e.g., compute, storage, communication) required to provide the capabilities and features of a network slice. NSSF enables other NFs (e.g., AMFs) to identify network slice instances suitable for the services desired by the UE.

[0071] - Authentication Server Functionality (AUSF) with Nausf Interface – Based on the user’s Home Network (HPLMN), it performs user authentication and computes security key material for various purposes.

[0072] - Network data analysis capabilities with an Nnwdaf interface (NWDAF, 210) are described in more detail above and below.

[0073] - Location Management Function (LMF) with Nlmf interface - Supports various functions related to UE location determination, including UE location determination and obtaining any of the following: DL location measurement or location estimate from the UE; UL location measurement from the NG RAN; and non-UE associated auxiliary data from the NG RAN.

[0074] The Unified Data Management (UDM) function supports the generation of 3GPP authentication credentials, user identity processing, access authorization based on subscription data, and other subscriber-related functions. To provide this functionality, the UDM uses subscription data (including authentication data) stored in the 5GC Unified Data Repository (UDR). In addition to interacting with the UDM, the UDR also supports the storage and retrieval of policy data by the PCF and application data by the NEF. The Data Storage Function (DSF) allows each NF to store its own context.

[0075] The communication link between the UE and the 5G network (AN and CN) can be grouped into two distinct layers. The UE communicates with the CN through the Non-Access Stratum (NAS) and with the AN through the Access Stratum (AS). All NAS communication is conducted via the NAS protocol ( Figure 2 The N1 interface (in the UE) is used between the UE and AMF. The security of communication at these layers is provided by the NAS protocol (for NAS) and the PDCP protocol (for AS).

[0076] 3GPP Release 17 enhances SBA by adding a Data Management Framework, which includes the Data Collection Coordination Function (DCCF) and Messaging Framework Adapter Function (MFAF) as detailed in 3GPP TR23.700-91 (v17.0.0). As mentioned above, the Data Management Framework is backward compatible with the NWDAF functionality of Release 16. For Release 17, the baseline for services provided by the DCCF (e.g., to the NWDAF) is the Release 16 NF service for obtaining data. For example, the baseline for the DCCF service used by the NWDAF consumer to obtain UE mobility data is Namf_EventExposure.

[0077] 3GPP TS 23.288 (v17.2.0) specifies NWDAF as the primary network function for calculating analytics reports. The 5G system architecture allows any NF to use the DCCF function and associated Ndccf services to obtain analytics from NWDAF. NWDAF can also store and retrieve analytics information from the Analytics Data Repository Function (ADRF).

[0078] 3GPP TS 23.288 further classifies NWDAF into two sub-functions (or logical functions): the NWDAF Analysis Logic Function (NWDAF AnLF), which performs the analysis process; and the NWDAF Model Training Logic Function (NWDAF MTLF), which performs the training and retraining of the ML model used by the NWDAF AnLF. In the following text, the terms "AnLF," "NWDAF AnLF," and "NWDAF (AnLF)" will be used interchangeably. Similarly, the terms "MTLF," "NWDAF MTLF," and "NWDAF (MTLF)" will be used interchangeably.

[0079] 3GPP TS 23.288 (v17.2.0) specifies a subscription / notification procedure for consumer NFs to retrieve ML models associated with one or more analytics IDs whenever an NWDAF MTLF has trained a new ML model and becomes available. This is known as ML model provisioning and is implemented by the Nnwdaf_MLModelProvision service.

[0080] 3GPP TR 33.738 (v0.2.0) describes a study on the security aspects of the drivers of network automation in 5G. One of the goals of this study is the security of AI / ML model sharing and storage, which is identified as “Critical Issue #3”. The following text from 3GPP TR 33.378 describes various aspects of this issue. In this paper, from the perspective of AI / ML of interest, “NFc” refers to the consumer NF, and “NFp” refers to the producer NF.

[0081] ***Start of 3GPP Text***

[0082] 5.3.1 Problem Details

[0083] AI / ML models are shared between NWDAF and / or NF (i.e., NWDAF with NWDAF, ADRF with NWDAF, etc.). In different scenarios, the NF producer of the AI / ML model can store the model in ADRF, NWDAF, or other entities.

[0084] The ADRF (Analytical Data Repository Functionality) is being enhanced to store AI / ML models, thereby facilitating the distribution and sharing of these models among NFs. Since AI / ML models and their algorithms are often proprietary (i.e., protected by the designer's intellectual property), it is essential to ensure that only NFs that have actually granted access to the AI / ML models can read and use them. Furthermore, the ADRF itself cannot be considered a completely trusted entity storing sensitive AI / ML data models. These models are indeed exposed and in a quiescent state within the ADRF.

[0085] The current licensing scheme defined by 3GPP for SBA applies only to the service level or resource / operation level. This licensing granularity may be insufficient in AI / ML model sharing scenarios because ADRF (Analysis Data Repository Function) or NWDAF, or any other network function capable of storing AI / ML models, cannot verify whether an NF consumer is authorized to retrieve AI / ML models.

[0086] 5.3.2 Security Threats

[0087] An unauthorized NFc (which in principle has no right to retrieve a specific model stored by NFp) may have permission to access the storage entity and retrieve that model.

[0088] If no protections are in place for accessing and reading AI / ML models in an ADRF stored by NFp, a compromised ADRF could expose algorithms and sensitive data to unauthorized entities that could easily misuse it and / or further distribute it to other entities, creating a greater data security vulnerability.

[0089] 5.3.3 Potential security requirements

[0090] AI / ML models should be protected between entities that generate ML models or store ML models in ADRF (e.g., NWDAF containing MTLF, NFp) and entities that consume the model (NFc).

[0091] The ADRF (Analysis Data Repository Function) or any other network function capable of storing AI / ML models should be able to authorize NFc to retrieve the AI / ML model.

[0092] NF service consumers should be authorized to access AI / ML models in ADRF (or any other NF capable of storing ML models, such as NWDAFMTLF).

[0093] ***End of 3GPP Text***

[0094] 3GPP TR 33.738 (v0.2.0) also describes a solution for authorizing and authenticating AI / ML model transfers, designated "Solution #2". This security solution protects the AI / ML model between the first entity (e.g., NF) that generates the AI / ML model (or stores the AI / ML model in the ADRF) and the second entity (NFc) that consumes the model. In this solution, the ADRF uses an authorization token to verify whether the NFc is permitted to access the ML model.

[0095] Figure 3 The signaling diagram for this solution, used for authorizing and authenticating AI / ML model transfers, is shown. Figure 3 As shown, this signaling occurs between NWDAF (AnLF) / NFc, the authorization server (e.g., NRF), NWDAF (MTLF) / NFp, and ADRF. Although Figure 3 The operations shown are given numerical labels, but this is intended for illustrative purposes and not to require or imply any particular order of operations, unless otherwise stated below.

[0096] In Operation 1, MTLF trains an ML model and sends it to ADRF by invoking the Nadrf_DataManagement_StorageRequest(ML model) service operation. In addition to model metadata, this message may include the ML model ID, analytics ID, vendor ID, MAC or SHA256 signature of the applied binary, the environment required to execute the ML model, a URL / link for retrieving the configuration, and a secret / signature key / certificate for generating authentication credentials. MTLF may send the ML model encrypted with a symmetric key (e.g., an AES key) before storage.

[0097] In Operation 2, the ADRF stores the ML model and response as specified in 3GPP TS 23.288 (v17.6.0), except that the storage is performed by the ADRF. In Operation 3, the NFC (e.g., NWDAF AnLF) contacts the NRF and requests an access token using the existing procedure specified in 3GPP TS 33.501 (v17.7.0). In Operation 4, the NRF sends the access token and MTLF ID using the existing procedure specified in 3GPP TS 23.288.

[0098] In Operation 5, the NWDAF (AnLF) requests the ML Model ID from the NWDAF (MTLF) using the Nnwdaf_MLModelProvision service operation and access token. The NWDAF (MTLF) retrieves the ML Model ID based on the ML analytics ID and / or ADRF ID. The NWDAF (MTLF) also verifies the received access token. In Operation 6, the NWDAF (MTLF) sends the Nnwdaf_MLModelProvision response, which includes the encryption key used to encrypt the AI / ML model in Operation 1. Additionally, the NWDAF (MTLF) may include a one-time credential for accessing the model from the ADRF, including any of the following:

[0099] - Random numbers, which are shared as part of metadata in operation 1;

[0100] - The MAC or hash value of the binary number or random number shared as part of the data in Operation 1;

[0101] - The signing key serves as the private key for the MTLF, where the public portion is shared in Operation 1;

[0102] - Credentials generated from the MTLF's signing key, such as JWT tokens or certificates.

[0103] One-time credentials can be used to limit the number of times an access can be made from an NFC. Even so, "one-time" credentials can also be used as regular authorization tokens for multiple accesses to an ML model, i.e., not just once (as the name suggests).

[0104] In operation 7, NWDAF (AnLF) uses the ADRF service procedure to request the ML model, including the one-time credential received in operation 6. In operation 8, ADRF verifies the one-time credential, and if the verification is successful, provides the stored AI / ML model to NWDAF (AnLF).

[0105] As mentioned above, AI / ML models are generally considered important intellectual property of their owners (e.g., 5GC vendors), and therefore their confidentiality and integrity need to be protected at all times. 3GPP is investigating the feasibility of sharing or storing ML models in network equipment that may be provided by different vendors. Under this arrangement, AI / ML models should be protected from access and use by consumer NFs provided by vendors different from the AI / ML models themselves. However, there is currently no specific solution for this requirement. For example, Figure 3 The solution shown cannot provide the required security in a multi-vendor network environment.

[0106] Embodiments of this disclosure address these and other problems, challenges, and / or difficulties by providing secure AI / ML model sharing between NFp (e.g., NWDAF MTLF) and NFc (e.g., NWDAF AnLF) and AI / ML model storage in ADRF. For example, an NFp (e.g., NWDAF MTLF) may be authorized to transfer and store its AI / ML models in an external repository (e.g., ADRF), and / or retrieve its AI / ML models from the repository. As another example, an NFp (e.g., NWDAF MTLF) may be authorized to transfer its AI / ML models to an NFc (e.g., NWDAF AnLF). As yet another example, an NFp (e.g., NWDAF MTLF) may confidentially protect its AI / ML models and / or model location information during the aforementioned transfer scenarios.

[0107] The embodiments of this disclosure can provide various benefits and / or advantages. By providing owners / producers of AI / ML models with the ability to protect AI / ML models during various transmission, storage, and retrieval scenarios, the embodiments improve the security of confidential and / or sensitive AI / ML models, thereby facilitating the deployment of such models in multi-vendor communication networks (e.g., 5GC).

[0108] In the following description of various embodiments, the terms NFp and NWDAF (MTLF) are used interchangeably, as are the terms NFc and NWDAF (AnLF). Similarly, the terms "model," "ML model," and "AI / ML model" are used interchangeably.

[0109] Figure 4 Signaling diagrams illustrating processes involving NWDAF (AnLF) (410), NRF (420), NWDAF (MTLF) (430), and ADRF (440) according to some embodiments of this disclosure are shown. Although Figure 4 The operations shown are given numerical labels, but this is intended for illustrative purposes and not to require or imply any particular order of operations, unless otherwise stated below.

[0110] In Operation 0, the NWDAF (MTLF) trains the ML model and can encrypt it to protect its integrity. The key used for protection can be referenced by a key ID or certificate and is bound to the interoperability ID, ML model ID, analytics ID, vendor ID, etc. It is assumed that NFs authorized for the same interoperability ID, ML model ID, analytics ID, vendor ID, etc., have corresponding keys for encryption / decryption / verification.

[0111] In addition, NWDAF (MTLF) registers its NF profile in the NRF using ML model information, which may include analytics ID, interoperability ID, vendor ID, ML model filter, model URL, model ID, and model authorization information. This model authorization information specifies the scope of authorization used to access the ML model, such as the requester, provider, model owner, and target model information. As a more specific example:

[0112] - The target model is identified by the following: interoperability ID, vendor ID, analytics ID, model owner, model filter, model URL, and model ID; and

[0113] - The scope is identified by the following: allowed requester and / or provider NF type / ID, allowed requester and / or provider vendor ID, allowed interoperability ID, etc.

[0114] In Operation 1, the NWDAF (MTLF) sends the trained ML model to the ADRF for storage by invoking the Nadrf_DataManagement_StorageRequest service operation. The NWDAF (MTLF) includes an SBA token, the encrypted ML model, one or more model identifiers (e.g., ML model ID, analytics ID, vendor ID, etc.), and optional model authorization information to facilitate subsequent access to the model in this message. In Operation 2, the ADRF stores the encrypted ML model and responds with a URL corresponding to the storage location of the ML model file (i.e., within the ADRF). In Operation 3, the NWDAF (MTLF) updates its NF profile in the NRF using the ML model information received from the ADRF (e.g., the URL).

[0115] In some variants, ADRF can register model licensing information in its own NF profile within the NRF (in a context similar to operation 9). In other variants, NWDAF (MTLF) can register model licensing information in the NRF within the ADRF's NF profile (i.e., on behalf of ADRF) (in a context similar to operation 0).

[0116] In Operation 4, NWDAF (AnLF) uses the existing procedures specified in 3GPP TS 23.288 to discover NWDAF (MTLF). In Operation 4a, NWDAF (AnLF) uses the existing procedures specified in 3GPP TS 33.501 to contact NRF to request an access token (“Token 1”). In Operation 4b, NRF provides Token 1 to NWDAF (AnLF) according to these procedures.

[0117] In operation 5, NWDAF (AnLF) retrieves the ML model using the Nnwdaf_MLModelProvision or Nnwdaf_MLModelInfo_Request service operation and the access token (token 1) received in operation 4. If NWDAF (MTLF) stores the model locally, it will perform operation 8 below. If the model is stored in ADRF, operations 6 and 7 are performed after operation 5.

[0118] In operation 6a, the NWDAF (MTLF) requests a token from the NRF to access the ML model by providing the analytics ID, interoperability ID, ML model ID, and model owner information (e.g., MTLF ID). In operation 6b, the NRF verifies whether the NWDAF (MTLF) is authorized to access the requested ML model based on model authorization information previously registered with the NRF (e.g., operations 0, 3). If so, a second access token (“Token 2”) is generated and sent to the NWDAF (MTLF).

[0119] In operation 7a, the NWDAF (MTLF) requests an ML model from the ADRF using the Nadrf_Model_Request service operation, including the analytics ID, interoperability ID, ML model ID, and token 2. In operation 7b, the ADRF verifies whether the NWDAF (MTLF) is authorized to retrieve the ML model based on the received token 2 or the ML model authorization information received in operation 1. If verified, in operation 7c, the ADRF sends the encrypted ML model to the NWDAF (MTLF).

[0120] In Operation 8, the NWDAF (MTLF) sends the ML model to the NWDAF (AnLF) using the Nnwdaf_MLModelProvision response, based on the service requested in Operation 5. In this operation, the ML model may still be encrypted (as in the encrypted ML model received in Operation 7c) or may be decrypted by the NWDAF (MTLF) and sent in plaintext. If the ML model is sent in encrypted form, the NWDAF (MTLF) may include information to help the NWDAF (AnLF) locate the key used for decryption / verification (e.g., the ID, certificate, or certificate URL associated with the key used to protect the ML model).

[0121] In some variations, ML model information is compared with... Figure 4 The methods shown are similar, obtaining the information via a URL, but using different services, messages, and / or protocols. The signaling flow in these embodiments is similar to... Figure 4The signaling flow shown is the same, but other download services, messages, and / or protocols can be used in operations 1b / 1c and 6 through 9. For example, an ML model can be obtained via a URL through a non-specified process assumed to be implemented by the vendor.

[0122] Figure 5 (including) Figures 5A to 5B Figure 5 shows a signaling diagram of another process involving NWDAF (AnLF) (510), NRF (520), NWDAF (MTLF) (530), and ADRF (540) according to other embodiments of this disclosure. Although the operations shown in Figure 5 are given numerical labels, this is intended for illustrative purposes and not to require or imply any particular order of operations unless otherwise stated below.

[0123] Operations 0a to 0b are the same as those described above. Figure 4 The corresponding operation is the same. In operation 1a, NWDAF(MTLF) is not as... Figure 4 Instead of sending an encrypted ML model as in Operation 1, the ADRF sends a URL (“URL1”) containing the ML model and from which it can be obtained. In Operation 1b, the ADRF uses the Nmtlf_Model_Request service operation to send URL1, the ML model ID, and an access token (“Token2”) to obtain / retrieve the ML model. In some variations, the ADRF may also include an analytics ID and / or an interoperability ID. In Operation 1c, the NWDAF (MTLF) provides the ML model in the response based on the verified access token.

[0124] In different variations of Operation 1a, URL1 can be sent in plaintext or encrypted form. If URL1 is encrypted, NWDAF (MTLF) can include information in the message to assist ADRF in locating the key used for decryption / verification (e.g., the ID, certificate, or certificate URL associated with the key used to protect URL1).

[0125] Although not shown in Figure 5, ADRF can obtain token 2 from NRF in a similar manner to how NWDAF (AnLF) obtains token 1 from NRF in operations 6a to 6b below. When issuing token 2, NRF checks whether ADRF can obtain the ML model from the URL based on the model authorization information registered in operation 0.

[0126] Operations 2 to 5a and above Figure 4The corresponding operation is the same. In operation 5b, the NWDAF (MTLF) uses the Nnwdaf_MLModelProvision response service operation to send the address of the ML model to the NWDAF (AnLF). For example, this address could be URL1 corresponding to the encrypted model stored in the NWDAF (MTLF), or URL2 corresponding to the encrypted model stored in the ADRF. In some variations, the NWDAF (MTLF) may include information to help the NWDAF (AnLF) locate the key used to decrypt / verify the ML model (e.g., the ID, certificate, or certificate URL associated with the key used to protect the ML model).

[0127] In different variations of Operation 5b, URL1 / URL2 can be sent in plaintext or encrypted form. If URL1 / URL2 is encrypted, the NWDAF (MTLF) can include information in the message to help the NWDAF (AnLF) locate the key used for decryption / verification (e.g., the ID, certificate, or certificate URL associated with the key used to protect URL1 / URL2).

[0128] In operation 6a, the NWDAF (AnLF) requests a token from the NRF for accessing the ML model via a URL (e.g., a download service). This request includes the analytics ID, interoperability ID, model owner information (e.g., MTLFID), and either URL1 or ULR2 received in operation 5b. In operation 6b, the NRF verifies whether the NWDAF (AnLF) is authorized to access the ML model based on model authorization information registered in the NRF (e.g., operations 0a, 3) and issues token 2.

[0129] If URL1 is received in operation 5b, operations 7a, 8a, and 9a are executed. In operation 7a, provided the analytics ID, interoperability ID, URL1, and token 2, NWDAF (AnLF) invokes the Nmtlf_Model_Download service operation to download the ML model from NWDAF (MTLF). In operation 8a, NWDAF (MTLF) verifies whether it is authorized to retrieve the ML model based on the received token 2 or local ML model authorization. In operation 9a, based on this verification, NWDAF (MTLF) uses a local ML model to download the ML model. Figure 4 The encrypted ML model is sent to NWDAF (MTLF) in a similar manner to operation 8 in the previous section.

[0130] If URL2 is received in operation 5b, operations 7b, 8b, and 9b are executed. In operation 7b, provided the analytics ID, interoperability ID, URL2, and token 2, NWDAF (AnLF) invokes the Nadrf_Model_Download service operation to download the ML model from ADRF. In operation 8b, based on the received token 2 or the model authorization information received in operation 1, NWDAF (AnLF) is authorized to retrieve the ML model. In operation 9b, based on this verification, ADRF... Figure 4 The encrypted ML model is sent to NWDAF (MTLF) in a similar manner to operation 8 in the previous section.

[0131] In some variations, different protocols (e.g., FTP) can be used instead of any of the service-based interfaces (SBIs) used in operations 1b / 1c, 7a / 7c, and 9a / 9c in Figure 5.

[0132] In some variations, instead of discovering NWDAF (MTLF) in Operation 4, NWDAF (AnLF) discovers ADRF via NRF based on interoperability ID, ML model ID, analytics ID, vendor ID, etc. In this case, NWDAF (AnLF) can directly obtain the address of the ML model (e.g., URL or FQDN) from ADRF.

[0133] For example, in operations 4a to 4b, the NWDAF (AnLF) requests and receives an SBA token (token 1) for accessing the ADRF. In this case, operations 5a to 5b are not performed, and in operations 6a to 6b, the NWDAF (AnLF) requests and receives an access token (token 2) for downloading the ML model from the ADRF, for example, via URL 2. Note that in these variations, token 2 can be the same as token 1. Furthermore, URL 2 can be provided in plaintext or encrypted form in operation 6b in a manner similar to that described above.

[0134] exist Figure 4 In some variations of the embodiment shown in Figure 5, if NWDAF (AnLF) wants to receive updates to the ML model via NWDAF (MTLF), NWDAF (AnLF) subscribes to the model update based on the interoperability ID, ML model ID, analytics ID, vendor ID, etc. If the model is updated, NWDAF (MTLF) can encrypt and protect the updated model with a different key than that used for the previous model version. Upon retrieving the updated model from NWDAF (MTLF), the entity can provide information to help NWDAF (AnLF) locate the key used for decryption / verification (e.g., the ID, certificate, or certificate URL associated with the key used to protect the updated model).

[0135] Alternatively, if NWDAF (AnLF) retrieves the updated model from ADRF, NFp notifies NWDAF (AnLF) of the model ID, URL, and / or FQDN, optionally including the new key ID, in a manner similar to operation 5b discussed above. NWDAF (AnLF) obtains the encrypted updated ML model from ADRF and performs decryption and integrity checks using the new key identified by NWDAF (MTLF).

[0136] Although the embodiments have been described above in the specific context of NWDAF and its logical functions MTLF and AnLF, those skilled in the art will understand that the basic principles of the above embodiments are equally applicable to other NFs, logical functions, nodes, etc., that may be called by different names but perform similar operations to MTLF and AnLF.

[0137] You can refer to this. Figures 6 to 9 To further illustrate these embodiments, Figures 6 to 9 Exemplary methods (e.g., processes) for consumer NF, producer NF, NRF, and ADRF are described respectively. In other words, the various features of the operations described below correspond to the various embodiments described above. Figures 6 to 9 The exemplary methods shown can be used in conjunction (e.g., in conjunction with each other and with other processes described herein) to provide the benefits, advantages, and / or solutions to the problems described herein. Although in 6 to Figure 9 These exemplary methods are illustrated by specific boxes in a particular order, but the operations corresponding to the boxes may be performed in a different order than shown, and may be combined and / or divided into boxes and / or operations with functions different from those shown. Optional boxes and / or operations are indicated by dashed lines.

[0138] More specifically, Figure 6 Exemplary methods (e.g., processes) for a consumer NF (NFc) in a communication network (e.g., 5GC) according to various embodiments of this disclosure are shown. As described elsewhere herein, Figure 6 The exemplary method shown can be performed by an NFc (e.g., an NWDAF (AnLF) or a network node hosting an NWDAF (AnLF)).

[0139] The exemplary method may include the operation at box 610, wherein the NFc may send a first request to a first NF of the communication network for a first access token associated with the ML model. The first request includes one or more of the following associated with the ML model: an analytics ID and an interoperability ID. The exemplary method may also include the operation at box 620, wherein the NFc may receive a first response from the first NF including the first access token. The exemplary method may also include the operation at box 630, wherein the NFc may send a second request for the ML model to a producer NF (NFp) of the communication network. The second request includes at least one of the analytics ID and the interoperability ID, as well as the first access token. The exemplary method may also include the operation at box 640, wherein the NFc may receive a second response from the NFp, the second response including one or more of the following: the ML model, an identifier of the ML model, and an address of a storage resource associated with the second NF of the communication network from which the ML model can be obtained.

[0140] In some embodiments, the first NF is a Network Repository Function (NRF). In other embodiments, the first NF is an Analytics Data Repository Function (ADRF). In some embodiments, one or more of the following are applicable:

[0141] -NFc is the analysis logic function NWDAF (AnLF) for network data analysis; and

[0142] - NFp is the model training logic function NWDAF (MTLF) for network data analysis.

[0143] In some embodiments, the second response includes an ML model (e.g., such as...) Figure 4 As shown, the ML model can be encrypted. In this case, the second response also includes information that can be used to locate a key that can be used to decrypt and verify the ML model. Some examples of such information have been discussed above.

[0144] In other embodiments, the second response includes the address of a storage resource associated with the ML model (e.g., as shown in Figure 5), and the exemplary method further includes the following operations marked with corresponding box labels:

[0145] - (650) Send a third request to the first NF for a second access token associated with the ML model, wherein the third request includes at least one of the analysis ID and the interoperability ID and the address of the storage resource associated with the second NF;

[0146] - (660) Receive a third response from the first NF, including the second access token; and

[0147] - (670) Obtain the ML model from the second NF using the second access token and the address of the storage resource associated with the second NF.

[0148] In some of these embodiments, the address of the storage resource is encrypted, and the second response also includes information that can be used to locate a key that can be used to decrypt and verify the address of the storage resource.

[0149] In some embodiments, the address of the storage resource associated with the second NF is a Universal Resource Locator (URL). In other embodiments, the address of the storage resource associated with the second NF is a Fully Qualified Domain Name (FQDN). In some embodiments, the second NF is an NFp. In other embodiments, the second NF is an ADRF.

[0150] also, Figure 7 Exemplary methods (e.g., procedures) for NFp in a communication network (e.g., 5GC) according to various embodiments of this disclosure are shown. As described elsewhere herein, Figure 7 The exemplary method shown can be performed by NFp (e.g., NWDAF (MTLF) or a network node hosting NWDAF (MTLF)).

[0151] The exemplary method includes the operation at box 710, wherein the NFp can register information associated with an ML model in the NRF of the communication network. The ML model is generated, owned, and / or maintained by the NFp, where "and / or" represents any one or more of the three listed properties. The registration information associated with the ML model includes an analytics ID and an interoperability ID. The exemplary method also includes the operation at box 720, wherein the NFp can encrypt the ML model and send a first request to the ADRF of the communication network for storing the encrypted ML model. The first request includes a first address of the encrypted ML model or a storage resource associated with the NFp, from which the ML model can be obtained.

[0152] In some embodiments, the exemplary method may further include the operation at block 750, wherein the NFp may receive a second request for an ML model from the NFc of the communication network. The second request includes at least one of an analytics ID and an interoperability ID, and a first access token. The exemplary method may further include the operation at block 780, wherein, based on verification of the first access token, the NFp may send a second response to the NFc, the second response including one or more of the following: the ML model; an identifier for the ML model; a first address of a storage resource associated with the NFp; and a second address of a storage resource associated with the ADRF, from which the ML model can be obtained.

[0153] In some of these embodiments, the first address of the storage resource associated with NFp is a first generic resource locator (URL), and the second address of the storage resource associated with ADRF is a second URL or fully qualified domain name (FQDN).

[0154] In some embodiments of these embodiments, the first request includes a first address of a storage resource associated with NFp, and the second response includes one of the following:

[0155] - The first address of the storage resource associated with NFp, or

[0156] - The second address of the storage resource associated with ADRF.

[0157] Figure 5 shows an example of these embodiments.

[0158] In some variations of these embodiments, the first address included in the first request is encrypted, and the first request also includes information that can be used to locate a key capable of being used to decrypt and verify the first address. In some variations of these embodiments, the first address or the second address included in the second response is encrypted, and the second response also includes information that can be used to locate a key capable of being used to decrypt and verify the first address or the second address.

[0159] In some other variations, the exemplary method may also include the following operations, labeled with the corresponding box numbers:

[0160] - (730) Receive another request for the ML model from the ADRF, wherein the other request includes a first address of the storage resource associated with the NFp and a second access token;

[0161] - (735) Based on the verification of the second access token, send another response to ADRF including an encrypted ML model; and

[0162] - (740) Then receive the second address of the storage resource associated with the ADRF from the ADRF.

[0163] In some other variations, the registration information associated with the ML model (e.g., from box 710) also includes a first address of the storage resource associated with the NFp, and the exemplary method also includes the operation of box 745, wherein the NFp can update the registration information associated with the ML model in the NRF to include the received second address.

[0164] In some variations of these embodiments, the second response includes a first address of the storage resource associated with the NFp, and the exemplary method further includes the following operations, marked with corresponding box labels:

[0165] - (790) Receive a third request for the ML model from the NFc, wherein the third request includes: a third access token associated with the ML model, a first address, and at least one of an analytics ID and an interoperability ID; and

[0166] - (795) Based on the verification of the third access token, send a third response including the ML model to the NFc.

[0167] In some other variations, the ML model in the third response is encrypted, and the third response also includes information that can be used to locate the key that can be used to decrypt and verify the ML model.

[0168] In other of these embodiments, the exemplary method may also include the following operations, marked with corresponding box numbers:

[0169] - (755) Send a fourth request to the NRF of the communication network for an access token associated with the ML model, wherein the fourth request includes at least one of an analysis ID and an interoperability ID;

[0170] - (760) Receive the requested access token from the NRF;

[0171] - (765) Send a fifth request for the ML model to the ADRF, wherein the fifth request includes at least one of the analysis ID and the interoperability ID and the received access token;

[0172] - (770) Receive the fifth response from ADRF, which includes the ML model.

[0173] The received ML model is then included in the second response sent to the NFc (e.g., in box 780).

[0174] In some embodiments, NFc is NWDAF (AnLF). In some embodiments, NFp is NWDAF (MTLF).

[0175] also, Figure 8 Exemplary methods (e.g., procedures) for NRF in a communication network (e.g., 5GC) according to various embodiments of this disclosure are illustrated. As described elsewhere herein, Figure 8 The exemplary method shown can be performed by the NRF or a network node that hosts the NRF.

[0176] The exemplary method includes the operation of block 810, wherein the NRF can register information associated with an ML model generated, owned, and / or maintained by a producer network function (NFp) of the communication network, wherein "and / or" represents any one or more of the three listed properties. The registration information associated with the ML model includes an analytics ID and an interoperability ID. The exemplary method also includes the operation of block 830, wherein the NRF can receive a first request for a first access token associated with the ML model from an NFc of the communication network. The first request includes at least one of the analytics ID and the interoperability ID. The exemplary method also includes the operation of block 840, wherein the NRF can send a first response to the NFc including the first access token.

[0177] In some embodiments, the exemplary method further includes the operation of block 850, wherein the NRF can receive a second request for a second access token from a first NF of the communication network. The second request includes at least one of an analytics ID and an interoperability ID, and one of the following:

[0178] - The first address of the storage resource associated with NFp, from which the ML model can be obtained; or

[0179] - A second address of the storage resource associated with the ADRF of the communication network, from which the ML model can be obtained.

[0180] The exemplary method may also include the operation of box 860, wherein the NRF may send a second response including a second access token to the first NF.

[0181] In some embodiments of these examples, the first address of the storage resource associated with the NFp is a first URL, and the second address of the storage resource associated with the ADRF is a second URL or FQDN. In some embodiments of these examples, the first NF is an NFc (e.g., as shown in Figure 5). In other embodiments of these examples, the first NF is an NFp (e.g., as shown in Figure 5). Figure 4 (As shown).

[0182] In some embodiments of these embodiments, the registration information associated with the ML model (e.g., in box 810) also includes a first address of a storage resource associated with the NFp, and the exemplary method also includes the operation of box 820, wherein the NRF may, for example, update the registration information based on a request from the NFp to include a second identifier of a storage resource associated with the ADRF.

[0183] In some embodiments, NFc is NWDAF (AnLF). In some embodiments, NFp is NWDAF (MTLF).

[0184] also, Figure 9 Exemplary methods (e.g., procedures) for ADRF in a communication network (e.g., 5GC) according to various embodiments of this disclosure are shown. As described elsewhere herein, Figure 9 The exemplary method shown can be performed by ADRF or a network node that hosts ADRF.

[0185] The exemplary method includes the operation at box 910, wherein the ADRF can receive a first request from an NFp in the communication network for storing an encrypted ML model. The first request includes the encrypted ML model or a first address of a storage resource associated with the NFp, from which the encrypted ML model can be obtained. The exemplary method also includes the operation at box 940, wherein the ADRF can store the encrypted ML model in a storage resource associated with the ADRF. The exemplary method further includes the operation at box 950, wherein the ADRF can send a first response to the NFp, the first response including a second address of a storage resource associated with the ADRF.

[0186] In some embodiments, the first request includes a first address of a storage resource associated with the NFp, and the exemplary method further includes the operation at box 920, wherein the ADRF may send another request for the ML model to the NFp. This other request includes the first address and a second access token. The exemplary method also includes the operation at box 930, wherein the ADRF may receive another response from the NFp including an encrypted ML model. The encrypted model is then stored in the storage resource associated with the ADRF (e.g., in box 940). Figure 5 illustrates examples of these embodiments.

[0187] In other embodiments, the exemplary method further includes the operation at box 960, wherein the ADRF can receive a second request for an ML model from a first NF in the communication. The second request includes at least one of an analytics ID and an interoperability ID, and a third access token. The exemplary method may also include the operation at box 970, wherein, based on verification of the third access token, the ADRF can send a second response including the ML model to the first NF.

[0188] In some embodiments of these examples, the first NF is NFp (e.g., as...). Figure 4 (As shown). In other embodiments of these embodiments, the first NF is the NFc of the communication network (e.g., as shown in Figure 5). In some variations of these embodiments, NFc is NWDAF (AnLF) and / or NFp is NWDAF (MTLF).

[0189] In some of these embodiments, the ML model included in the second response (e.g., in box 970) is encrypted, and the second response also includes information that can be used to locate a key that can be used to decrypt and verify the ML model.

[0190] In some embodiments, the first address of the storage resource associated with NFp is a first URL, and the second address of the storage resource associated with ADRF is a second URL or FQDN.

[0191] Although various embodiments have been described above in terms of methods, techniques and / or processes, those skilled in the art will readily understand that such methods, techniques and / or processes can be embodied by various combinations of hardware and software in various systems, communication devices, computing devices, control devices, apparatuses, non-transitory computer-readable media, computer program products and the like.

[0192] Figure 10 An example of a communication system 1000 according to some embodiments is shown. In this example, the communication system 1000 includes a telecommunications network 1002, which includes an access network 1004 (e.g., a RAN) and a core network 1006, the core network 1006 including one or more core network nodes 1008. The access network 1004 includes one or more access network nodes, such as network nodes 1010a to 1010b (one or more of which may generally be referred to as network node 1010), or any other similar 3GPP access node or non-3GPP access point. Network node 1010 facilitates direct or indirect connections of UEs, for example, connecting UEs 1012a to UEs 1012d (one or more of which may generally be referred to as UE 1012) to the core network 1006 via one or more wireless connections.

[0193] Examples of wireless communication via wireless connection include transmitting and / or receiving wireless signals using electromagnetic waves, radio waves, infrared waves, and / or other types of signals suitable for transmitting information without using wiring, cables, or other conductors. Furthermore, in various embodiments, communication system 1000 may include any number of wired or wireless networks, network nodes, UEs, and / or any other components or systems that can facilitate or participate in the communication of data and / or signals (whether via wired or wireless connections). Communication system 1000 may include any type of communication, telecommunications, data, cellular, radio network, and / or other similar system, and / or interface with any type of communication, telecommunications, data, cellular, radio network, and / or other similar system.

[0194] UE 1012 can be any of a wide variety of communication devices, including wireless devices that are arranged, configured, and / or operable to communicate wirelessly with network node 1010 and other communication devices. Similarly, network node 1010 is arranged, capable, configured, and / or operable to communicate directly or indirectly with UE 1012 and / or with other network nodes or devices in telecommunication network 1002 to achieve and / or provide network access (such as wireless network access) and / or to perform other functions in telecommunication network 1002 (such as management).

[0195] In the depicted example, core network 1006 connects network node 1010 to one or more hosts (such as host 1016). These connections can be direct connections or indirect connections via one or more intermediate networks or devices. In other examples, network nodes may be directly coupled to hosts. Core network 1006 includes one or more core network nodes (e.g., 1008) composed of hardware and software components. The characteristics of these components may be substantially similar to those described with respect to UEs, network nodes, and / or hosts, such that the description is generally applicable to the corresponding components of core network node 1008. Example core network nodes include the functions of one or more of the following: Mobile Switching Center (MSC), Mobility Management Entity (MME), Home Subscriber Server (HSS), Access and Mobility Management Function (AMF), Session Management Function (SMF), Authentication Server Function (AUSF), Subscription Identifier Unhiding Function (SIDF), Unified Data Management (UDM), Security Edge Protection Agent (SEPP), Network Open Function (NEF), and / or User Plane Function (UPF).

[0196] Host 1016 may be owned or under the control of a service provider other than the operator or provider of access network 1004 and / or telecommunications network 1002, and may be operated by or on behalf of the service provider. Host 1016 may host a variety of applications to provide one or more services. Examples of such applications include real-time and pre-recorded audio / video content, data collection services (e.g., retrieving and compiling data about various environmental conditions detected by multiple UEs), analytics functions, social media, functions for controlling or otherwise interacting with remote devices, functions for alarms and monitoring centers, or any other such functions performed by a server.

[0197] As a whole, Figure 10The communication system 1000 enables connections between the UE, network nodes, and hosts. In this sense, the communication system can be configured to operate according to predefined rules or procedures, such as specific standards, including but 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 standard (WiFi); and / or any other suitable wireless communication standards, such as Global Microwave Access Interoperability (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.

[0198] In some examples, telecommunications network 1002 is a cellular network implementing 3GPP standardized features. Therefore, telecommunications network 1002 can support network slicing to provide different logical networks to different devices connected to it. For example, telecommunications network 1002 can 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 IoT services to yet another UE.

[0199] In some examples, UE 1012 is configured to send and / or receive information without direct human interaction. For example, the UE may be designed to send information to access network 1004 according to a predetermined schedule when triggered by internal or external events or in response to a request from access network 1004. Additionally, the UE may be configured to operate in single-RAT mode, multi-RAT mode, or multi-standard mode. For example, the UE may operate using any or a combination of Wi-Fi, NR (New Radio), and LTE, i.e., configured for multi-radio dual connectivity (MR-DC), such as E-UTRAN (Evolved UMTS Terrestrial Radio Access Network) New Radio Dual Connectivity (EN-DC).

[0200] In this example, central node 1014 communicates with access network 1004 to facilitate indirect communication between one or more UEs (e.g., UE 1012c and / or 1012d) and network nodes (e.g., network node 1010b). In some examples, central node 1014 may be a controller, router, content source and analyzer, or any other communication device described herein relating to the UE. For example, central node 1014 may be a broadband router that enables the UE to access core network 1006. As another example, central node 1014 may be a controller that sends commands or instructions to one or more actuators in the UE. Commands or instructions may be received from the UE, network node 1010, or via executable code, scripts, procedures, or other instructions in central node 1014. As another example, central node 1014 may be a data collector that acts as a temporary storage device for UE data, and in some embodiments, may perform data analysis or other processing. As another example, central node 1014 may be a content source. For example, for a UE acting as a VR headset, display, speaker, or other media delivery device, the central node 1014 can retrieve VR assets, video, audio, or other media or data related to perception information via a network node, and then provide them directly to the UE after performing local processing and / or adding additional local content. In yet another example, the central node 1014 acts as a proxy server or orchestrator for the UE, particularly if one or more of the UEs are low-energy IoT devices.

[0201] Central node 1014 may have a persistent / persistent or intermittent connection to network node 1010b. Central node 1014 may also allow different communication schemes and / or scheduling between central node 1014 and UEs (e.g., UE 1012c and / or UE 1012d) and between central node 1014 and core network 1006. In other examples, central node 1014 is connected to core network 1006 and / or one or more UEs via a wired connection. Furthermore, central node 1014 may be configured to connect to an M2M service provider via access network 1004, and / or connect to another UE via a direct connection. In some scenarios, a UE may establish a wireless connection with network node 1010 while still being connected via central node 1014 via a wired or wireless connection. In some embodiments, central node 1014 may be a dedicated central node—that is, a central node whose primary function is to route communication from network node 1010b to UE / to network node 1010b. In other embodiments, the central node 1014 may be a non-dedicated central node—that is, a device capable of operating to route communication between the UE and network node 1010b, but additionally capable of operating as a communication start point and / or endpoint for certain data channels.

[0202] Figure 11 A UE 1100 according to some embodiments is illustrated. Examples of UEs include, but are not limited to, smartphones, mobile phones, cellular phones, Voice over IP (VoIP) phones, wireless local loop phones, desktop computers, personal digital assistants (PDAs), wireless cameras, game consoles or devices, music storage devices, playback devices, wearable terminal devices, wireless endpoints, mobile stations, tablet computers, laptop computers, laptop embedded devices (LEEs), laptop-mounted devices (LMEs), smart devices, wireless client devices (CPEs), vehicle-mounted or vehicle-embedded / integrated wireless devices, etc. Other examples include any UE recognized by 3GPP, including narrowband Internet of Things (NB-IoT) UEs, machine-type communication (MTC) UEs, and / or enhanced MTC (eMTC) UEs.

[0203] The UE can support device-to-device (D2D) communication, for example, by implementing 3GPP standards for sidelink communication, Dedicated Short-Range Communication (DSRC), vehicle-to-vehicle (V2V), vehicle-to-infrastructure (V2I), or vehicle-to-everything (V2X). In other examples, the UE may not necessarily be a user in the sense of a human user who owns and / or operates the associated device. Alternatively, the UE may represent a device intended to be sold to or operated by a human user but which may not or initially may not be associated with a particular human user (e.g., a smart sprinkler controller). Alternatively, the UE may represent a device not intended to be sold to or operated by an end user but which may be associated with or operated for the benefit of the user (e.g., a smart power meter).

[0204] UE 1100 includes processing circuitry 1102, operatively coupled via bus 1104 to input / output interface 1106, power supply 1108, memory 1110, communication interface 1112, and other components that may not be explicitly shown. Some UEs may utilize... Figure 11 The components shown may be all or a subset. The level of integration between components can vary depending on the UE. Furthermore, some UEs may contain multiple instances of components, such as multiple processors, memories, transceivers, transmitters, receivers, etc.

[0205] Processing circuitry 1102 is configured to process instructions and data and can be configured to implement any sequential state machine operable to execute instructions stored in memory 1110 as a machine-readable computer program. Processing circuitry 1102 can be implemented as: one or more hardware-implemented state machines (e.g., implemented with discrete logic, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), etc.); programmable logic together with appropriate firmware; one or more stored computer programs, general-purpose processors (e.g., microprocessors or digital signal processors (DSPs)) together with appropriate software; or any combination of the foregoing. For example, processing circuitry 1102 may include multiple central processing units (CPUs).

[0206] In this example, input / output interface 1106 can be configured to provide one or more interfaces to input devices, output devices, or one or more input and / or output devices. Examples of output devices include speakers, sound cards, video cards, displays, monitors, printers, actuators, transmitters, smart cards, other output devices, or any combination thereof. Input devices can allow users to capture information into UE 1100. Examples of input devices include touch-sensitive or presence-sensitive displays, cameras (e.g., digital cameras, digital camcorders, webcams, etc.), microphones, sensors, mice, trackballs, directional keyboards, touchpads, scroll wheels, smart cards, etc. Presence-sensitive displays may include capacitive or resistive touch sensors to sense input from the user. Sensors may be, for example, accelerometers, gyroscopes, tilt sensors, force sensors, magnetometers, optical sensors, proximity sensors, biometric sensors, etc., or any combination thereof. Output devices can use the same type of interface port as input devices. For example, a Universal Serial Bus (USB) port can be used to provide both input and output devices.

[0207] In some embodiments, power supply 1108 is configured as a battery or battery pack. Other types of power sources may be used, such as external power sources (e.g., power outlets), photovoltaic devices, or batteries. Power supply 1108 may also include power supply circuitry for delivering power from power supply 1108 itself and / or external power sources to various parts of UE 1100 via input circuitry or an interface such as a power cable. The delivery of power may, for example, be used to charge power supply 1108. The power supply circuitry may perform any formatting, conversion, or other modifications on the power from power supply 1108 to suit the power for the various components of UE 1100 to which it supplies power.

[0208] Memory 1110 may be or be configured to include memory such as random access memory (RAM), read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), disk, optical disk, hard disk, removable magnetic tape, flash drive, etc. In one example, memory 1110 includes one or more applications 1114, such as an operating system, web browser application, widgets, utility engine, or other applications, and corresponding data 1116. Memory 1110 may store any one or a combination of various operating systems used by UE 1100.

[0209] The memory 1110 can be configured to include multiple physical drive units, such as a redundant array of independent disks (RAID), flash memory, a USB flash drive, an external hard drive, a thumb drive, a pen drive, a key drive, a high-density digital versatile optical disc (HD-DVD) drive, an internal hard drive, a Blu-ray disc drive, a holographic digital data storage (HDDS) disc drive, an external mini dual in-line memory module (DIMM), synchronous dynamic random access memory (SDRAM), external micro DIMM SDRAM, smart card memory (e.g., a tamper-proof module in the form of a universal integrated circuit card (UICC) including one or more subscriber identification modules (SIMs), such as USIM and / or ISIM), other memory, or any combination thereof. The UICC can be, for example, an embedded UICC (eUICC), an integrated UICC (iUICC), or a removable UICC commonly referred to as a "SIM card." The memory 1110 can allow the UE 1100 to access instructions, applications, etc., stored on transient or non-transient storage media to offload or upload data. Articles of art (such as articles of art utilizing communication systems) may be tangibly embodied in or contained in memory 1110, which may be or include a device-readable storage medium.

[0210] Processing circuitry 1102 can be configured to communicate with an access network or other network using communication interface 1112. Communication interface 1112 may include one or more communication subsystems and may include or be communicatively coupled to antenna 1122. Communication interface 1112 may include one or more transceivers for communication (e.g., via one or more remote transceivers capable of wireless communication with another device (e.g., another UE or a network node in the access network). Each transceiver may include transmitter 1118 and / or receiver 1120 suitable for providing network communication (e.g., optical, electrical, frequency allocation, etc.). Furthermore, transmitter 1118 and / or receiver 1120 may be coupled to one or more antennas (e.g., 1122) and may share circuitry, software, or firmware, or alternatively, be implemented separately.

[0211] In the illustrated embodiment, the communication functions of the communication interface 1112 may include cellular communication, Wi-Fi communication, LPWAN communication, data communication, voice communication, multimedia communication, short-range communication such as Bluetooth, near-field communication, location-based communication (e.g., using a Global Positioning System (GPS) to determine location), another type of communication function, or any combination thereof. Communication may be implemented according to one or more communication protocols and / or standards (e.g., IEEE 802.11, Code Division Multiple Access (CDMA), Wideband Code Division Multiple Access (WCDMA), GSM, LTE, New Radio (NR), UMTS, WiMax, Ethernet, Transmission Control Protocol / Internet Protocol (TCP / IP), Synchronous Optical Network (SONET), Asynchronous Transfer Mode (ATM), QUIC, Hypertext Transfer Protocol (HTTP), etc.).

[0212] Regardless of the sensor type, the UE can provide the output of data captured by its sensors via its communication interface 1112 through a wireless connection with a network node. Data captured by the UE's sensors can be transmitted via another UE through the same wireless connection. The output can be periodic (e.g., every 15 minutes if it reports the sensed temperature), random (e.g., to balance the load of reports from several sensors), responsive to a triggering event (e.g., sending an alarm when humidity is detected), responsive to a request (e.g., a user-initiated request), or a continuous stream (e.g., real-time video feed of a patient).

[0213] As another example, the UE includes actuators, motors, or switches associated with a communication interface configured to receive wireless input from a network node via a wireless connection. The state of the actuator, motor, or switch can change in response to the received wireless input. For example, the UE may include a motor that adjusts the control surfaces or rotors of a flying drone based on the received input, or adjusts a robotic arm performing a medical procedure based on the received input.

[0214] When the UE is in the form of an Internet of Things (IoT) device, the UE can be a device used in one or more application areas, including but not limited to urban wearable technology, extended industrial applications, and healthcare. Non-limiting examples of such IoT devices include or embedded in the following devices: connected refrigerators or freezers, televisions, connected lighting devices, electricity meters, robotic vacuum cleaners, voice-controlled smart speakers, home security cameras, motion detectors, thermostats, smoke detectors, door and window sensors, flood / humidity sensors, electronic door locks, connected doorbells, air conditioning systems (such as heat pumps), autonomous vehicles, monitoring systems, weather monitoring devices, vehicle parking monitoring devices, electric vehicle charging stations, smartwatches, fitness trackers, head-mounted displays for augmented reality (AR) or virtual reality (VR), wearable devices for haptic or sensory enhancement, sprinklers, animal or item tracking devices, sensors for monitoring plants or animals, industrial robots, unmanned aerial vehicles (UAVs), and any kind of medical device (such as heart rate monitors or remote-controlled surgical robots). In addition to the above... Figure 11 In addition to the other components described in the UE 1100 shown, a UE in the form of an IoT device also includes circuitry and / or software depending on the intended application of the IoT device.

[0215] As another specific example, in an IoT scenario, a UE can represent a machine or other device that performs monitoring and / or measurement and sends the results of such monitoring and / or measurement to another UE and / or network node. In this case, the UE can be an M2M device, which can be referred to as an MTC device in the 3GPP context. The UE can implement the 3GPP NB-IoT standard. In other scenarios, a UE can represent a vehicle (e.g., a car, bus, truck, ship, and aircraft) or other device capable of monitoring and / or reporting its operational status or other functions associated with its operation.

[0216] In practice, any number of UEs can be used together for a single use case. For example, the first UE can be a drone or integrated into a drone, and provides the drone's speed information (obtained via a speed sensor) to a second UE, which is a remote controller for operating the drone. When the user makes a change from the remote controller, the first UE can adjust the throttle on the drone (e.g., by controlling the actuators) to increase or decrease the drone's speed. The first UE and / or the second UE can also include more than one of the functions described above. For example, the UE can include sensors and actuators, and handle data communication between both the speed sensor and the actuators.

[0217] Figure 12 A network node 1200 according to some embodiments is shown. Examples of network nodes include, but are not limited to, access points (e.g., radio access points) and base stations (e.g., radio base stations, NodeBs, eNBs, and gNBs).

[0218] Base stations can be classified based on the coverage they provide (or, in other words, their transmit power level); therefore, depending on the coverage provided, a base station can be called a femtobase, picobase, microbase, or macrobase. A base station can be a relay node or a relay donor node controlling a relay. Network nodes can also include one or more (or all) portions of a distributed radio base station, such as centralized digital units and / or remote radio units (RRUs), sometimes referred to as remote radio headends (RRHs). These remote radio units may be integrated with an antenna to form an antenna-integrated radio, or they may not be integrated with an antenna to form an antenna-integrated radio. A portion of a distributed radio base station can also be referred to as a node in a distributed antenna system (DAS).

[0219] Other examples of network nodes include multi-transmitter point (multi-TRP) 5G access nodes, multi-standard radio (MSR) devices (e.g., MSR BS), network controllers (e.g., radio network controllers (RNC) or base station controllers (BSC)), base transceiver stations (BTS), transmitter points, transmitter nodes, multi-cell / multicast coordination entities (MCE), operations and maintenance (O&M) nodes, operations support system (OSS) nodes, self-organizing network (SON) nodes, location nodes (e.g., evolved Serving Mobility Location Center (E-SMLC)) and / or minimized drive test (MDT).

[0220] For example, one or more network nodes 1200 may be configured to perform various methods or processes attributable to the NWDAF (or its logical functions) described herein. As a more specific example, one or more network nodes 1200 may be configured to perform operations attributable to the consumer NF (e.g., NWDAF AnLF), producer NF (e.g., NWDAF MTLF), NRF, and ADRF.

[0221] Network node 1200 includes processing circuitry 1202, memory 1204, communication interface 1206, and power supply 1208. Network node 1200 may consist of multiple physically separate components (e.g., NodeB and RNC components, BTS and BSC components, etc.), each with its own corresponding components. In some scenarios where network node 1200 includes multiple separate components (e.g., BTS and BSC components), one or more of these separate components may be shared among multiple network nodes. For example, a single RNC may control multiple NodeBs. In such scenarios, each unique “NodeB and RNC pair” may, in some cases, be considered a single, separate network node. In some embodiments, network node 1200 may be configured to support multiple Radio Access Technologies (RATs). In such embodiments, some components may be replicated (e.g., separate memory 1204 for different RATs), and some components may be reused (e.g., the same antenna 1210 may be shared by different RATs). Network node 1200 may also include multiple sets of various components shown for different wireless technologies (e.g., GSM, WCDMA, LTE, NR, WiFi, Zigbee, Z-wave, LoRaWAN, RFID, or Bluetooth wireless technologies). These wireless technologies may be integrated into the same or different chips or chipsets and other components within network node 1200.

[0222] The processing circuitry 1202 may include one or more of the following: a microprocessor, a controller, a central processing unit, a digital signal processor, an application-specific integrated circuit, a field-programmable gate array, or any other suitable computing device, resource, or combination of hardware, software and / or coding logic, operable to provide network node 1200 functionality, either alone or in combination with other network node 1200 components (e.g., memory 1204).

[0223] In some embodiments, the processing circuitry 1202 includes a system-on-a-chip (SOC). In some embodiments, the processing circuitry 1202 includes a radio frequency (RF) transceiver circuitry 1212 and / or a baseband processing circuitry 1214. In some embodiments, the RF transceiver circuitry 1212 and the baseband processing circuitry 1214 may be on separate chips (or chipsets), boards, or units (e.g., radio units and digital units). In alternative embodiments, some or all of the RF transceiver circuitry 1212 and / or the baseband processing circuitry 1214 may be on the same chip or chipset, board, or unit group.

[0224] Memory 1204 may include any form of volatile or non-volatile computer-readable memory, including but not limited to permanent storage devices, solid-state memory, remotely mounted memory, magnetic media, optical media, random access memory (RAM), read-only memory (ROM), mass storage media (e.g., hard disk), removable storage media (e.g., flash drives, optical discs (CDs), or digital video discs (DVDs)) and / or any other volatile or non-volatile, non-transitory device-readable and / or computer-executable memory device that stores information, data, and / or instructions that can be used by processing circuitry 1202. Memory 1204 may store any suitable instructions, data, or information, including computer programs, software, applications including logic, rules, codes, tables, and / or other instructions that can be executed by processing circuitry 1202 and used by network node 1200. Storage device 1204 may be used to store any calculations performed by processing circuitry 1202 and / or any data received via communication interface 1206. In some embodiments, processing circuitry 1202 and memory 1204 are integrated together.

[0225] Communication interface 1206 is used for wired or wireless communication of signaling and / or data between network nodes, access networks, and / or UEs. As shown, communication interface 1206 includes a port / terminal 1216 for transmitting and receiving data to and from the network, for example, via a wired connection. Communication interface 1206 also includes radio front-end circuitry 1218, which may be coupled to antenna 1210, or in some embodiments, is part of antenna 1210. Radio front-end circuitry 1218 includes a filter 1220 and an amplifier 1222. Radio front-end circuitry 1218 may be connected to antenna 1210 and processing circuitry 1202. Radio front-end circuitry 1218 may be configured to modulate the signal transmitted between antenna 1210 and processing circuitry 1202. Radio front-end circuitry 1218 may receive digital data to be transmitted to other network nodes or UEs via a wireless connection. Radio front-end circuitry 1218 may use a combination of filter 1220 and / or amplifier 1222 to convert the digital data into a radio signal with suitable channel and bandwidth parameters. The radio signal may then be transmitted via antenna 1210. Similarly, when receiving data, antenna 1210 can collect radio signals, which are then converted into digital data by radio front-end circuitry 1218. The digital data can then be passed to processing circuitry 1202. In other embodiments, the communication interface may include different components and / or different combinations of components.

[0226] In some alternative embodiments, network node 1200 does not include a separate radio front-end circuitry 1218; instead, processing circuitry 1202 includes radio front-end circuitry and is connected to antenna 1210. Similarly, in some embodiments, all or some of the RF transceiver circuitry 1212 is part of communication interface 1206. In yet another embodiment, communication interface 1206 includes one or more ports or terminals 1216, radio front-end circuitry 1218, and RF transceiver circuitry 1212 as part of a radio unit (not shown), and communication interface 1206 communicates with baseband processing circuitry 1214, which is part of a digital unit (not shown).

[0227] Antenna 1210 may include one or more antennas or an antenna array configured to transmit and / or receive wireless signals. Antenna 1210 may be coupled to radio front-end circuitry 1218 and may be any type of antenna capable of wirelessly transmitting and receiving data and / or signals. In some embodiments, antenna 1210 is decoupled from network node 1200 and may be connected to network node 1200 via an interface or port.

[0228] Antenna 1210, communication interface 1206, and / or processing circuitry 1202 can be configured to perform any receive operation and / or certain acquire operation described herein by a network node. Any information, data, and / or signals can be received from the UE, another network node, and / or any other network device. Similarly, antenna 1210, communication interface 1206, and / or processing circuitry 1202 can be configured to perform any transmit operation described herein by a network node. Any information, data, and / or signals can be transmitted to the UE, another network node, and / or any other network device.

[0229] Power supply 1208 provides power to the various components of network node 1200 in a form suitable for the various components (e.g., at the voltage and current levels required by each respective component). Power supply 1208 may also include or be coupled to power management circuitry to supply power to the components of network node 1200 for performing the functions described herein. For example, network node 1200 may be connected to an external power source (e.g., mains, power outlet) via input circuitry or an interface (e.g., cable), thereby supplying power to the power circuitry of power supply 1208. As another example, power supply 1208 may include a power source in the form of a battery or battery pack, which is connected to or integrated into the power circuitry. The battery can provide backup power if the external power source fails.

[0230] Embodiments of network node 1200 may include more than Figure 12 Additional components shown are provided to offer certain aspects of the functionality of the network node, including any of the functions described herein and / or any functionality required to support the topics described herein. For example, network node 1200 may include a user interface device to allow information to be input into and output from network node 1200. This allows users to perform diagnostic, maintenance, repair, and other management functions on network node 1200.

[0231] Figure 13 Based on the block diagram of host 1300 described in this document, host 1300 can be... Figure 10 An embodiment of host 1016. As used herein, host 1300 may be or include various combinations of hardware and / or software, including processing resources in a standalone server, blade server, cloud-implemented server, distributed server, virtual machine, container, or server cluster. Host 1300 may provide one or more services to one or more UEs.

[0232] Host 1300 includes processing circuitry 1302 operably coupled via bus 1304 to input / output interface 1306, network interface 1308, power supply 1310, and memory 1312. Other components may be included in other embodiments. The features of these components may be substantially similar to those shown in the previous figures (e.g., Figure 11 and Figure 12 The characteristics described for the device make its description generally applicable to the corresponding components of the host 1300.

[0233] Memory 1312 may include one or more computer programs, including data 1316 and one or more host applications 1314. Data 1316 may include user data, such as data generated by the UE for the host 1300, or data generated by the host 1300 for the UE. Embodiments of host 1300 may utilize only a subset or all of the illustrated components. Host application 1314 may be implemented in a container-based architecture and may provide support for video codecs (e.g., Universal Video Coding (VVC), High Efficiency Video Coding (HEVC), Advanced Video Coding (AVC), MPEG, VP9) and audio codecs (e.g., FLAC, Advanced Audio Coding (AAC), MPEG, G.711), including transcoding for various categories, types, or implementations of UEs (e.g., mobile phones, desktop computers, wearable display systems, head-up display systems). Host application 1314 may also provide user authentication and authorization checks and may periodically report health status, routing, and content availability to a central node (such as a device in the core network or a device at the edge of the core network). Therefore, host 1300 can select and / or indicate different hosts for over-the-top services for the UE. Host application 1314 can support various protocols, such as HTTP Live Streaming (HLS), Real-time Messaging Protocol (RTMP), Real-time Streaming Protocol (RTSP), Dynamic Adaptive Streaming over HTTP (MPEG-DASH), etc.

[0234] Figure 14This is a block diagram illustrating a virtualization environment 1400 in which functionality implemented by some embodiments can be virtualized. In this context, virtualization means creating a virtual version of an apparatus or device that may include a virtualization hardware platform, storage devices, and network resources. As used herein, virtualization can be applied to any device or component thereof described herein, and involves at least a portion of its functionality being implemented as one or more virtual components. Some or all of the functionality described herein can be implemented as virtual components executed by one or more virtual machines (VMs) in one or more virtual environments 1400 hosted by one or more hardware nodes (e.g., hardware computing devices operating as network nodes, UEs, core network nodes, or hosts). Furthermore, in embodiments where virtual nodes do not require radio connectivity (e.g., core network nodes or hosts), the nodes can be fully virtualized.

[0235] Application 1402 (which may alternatively be referred to as a software instance, virtual application, network function, virtual node, virtual network function, etc.) operates in virtualization environment 1400 to implement some of the features, functions, and / or benefits of some embodiments disclosed herein.

[0236] For example, various NFs (or portions thereof) described herein with reference to other figures can be implemented as virtual network functions 1402 in the virtualization environment 1400. As more specific examples, consumer NFs (e.g., NWDAF AnLF), producer NFs (e.g., NWDAF MTLF), NRFs, and / or ADRFs can be implemented as virtual network functions 1402 in the virtualization environment 1400.

[0237] Hardware 1404 includes processing circuitry, memory storing software and / or instructions executable by the hardware processing circuitry, and / or other hardware devices described herein (such as network interfaces, input / output interfaces, etc.). The software can be executed by the processing circuitry to instantiate one or more virtualization layers 1406 (also referred to as hypervisors or virtual machine monitors (VMMs)), provide VMs 1408a and 1408b (one or more of which may generally be referred to as VM 1408), and / or perform any functions, features, and / or benefits described in relation to some embodiments described herein. Virtualization layer 1406 can present a virtual operating platform to VM 1408, which appears as network hardware.

[0238] VM 1408 includes virtual processing, virtual memory, virtual network or interface, and virtual storage, and can be run by a corresponding virtualization layer 1406. Different embodiments of instances of virtual device 1402 can be implemented on one or more VMs 1408, and these implementations can be made in different ways. In some contexts, hardware virtualization is referred to as Network Functions Virtualization (NFV). NFV can be used to unify numerous network device types onto industry-standard high-capacity server hardware, physical switches, and physical storage that can reside in data centers and customer premises equipment (CPE).

[0239] In the context of NFV, VM 1408 can be a software implementation of a physical machine, whose running program behaves as if it were running on a physical, non-virtualized machine. Each VM 1408, along with the portion of hardware 1404 that executes that VM (whether it is dedicated hardware for that VM and / or hardware shared by that VM with other VMs), forms a separate virtual network element. Still within the context of NFV, the virtual network function is responsible for handling the specific network functions running on one or more VMs 1408 above hardware 1404 and corresponding to application 1402.

[0240] Hardware 1404 can be implemented in a standalone network node with general or specific components. Hardware 1404 may implement some functions via virtualization. Alternatively, hardware 1404 may be part of a larger hardware cluster (e.g., in a data center or CPE) where many hardware nodes work together and are managed by management and orchestration 1410, which in particular oversees the lifecycle management of application 1402. In some embodiments, hardware 1404 is coupled to one or more radio units, each radio unit including one or more transmitters and one or more receivers that can be coupled to one or more antennas. The radio units may communicate directly with other hardware nodes via one or more suitable network interfaces and may be used in conjunction with virtual components to provide radio capabilities to virtual nodes (e.g., radio access nodes or base stations). In some embodiments, some signaling may be provided using a control system 1412, which may alternatively be used for communication between hardware nodes and radio units.

[0241] Figure 15 A communication diagram is shown illustrating how host 1502 communicates with UE 1506 via network node 1504 through a partial wireless connection, according to some embodiments. Reference will now be made to... Figure 15 Describe the UE discussed in the preceding paragraphs (e.g., Figure 10 UE1012a and / or Figure 11 UE 1100), network nodes (e.g., Figure 10Network node 1010a and / or Figure 12 Network node 1200) and host (e.g., Figure 10 Host 1016 and / or Figure 13 Example implementations of the host 1300 according to various embodiments.

[0242] Similar to host 1300, embodiments of host 1502 include hardware such as a communication interface, processing circuitry, and memory. Host 1502 also includes software stored in or accessible by host 1502 and executable by the processing circuitry. This software includes a host application operable to provide services to a remote user, such as UE 1506 connected via an over-the-top (OTT) connection 1550 extending between UE 1506 and host 1502. When providing services to a remote user, the host application can provide user data transmitted using OTT connection 1550.

[0243] Network node 1504 includes hardware that enables it to communicate with host 1502 and UE 1506. Connection 1560 can be a direct connection or via a core network (such as...). Figure 10 The connection to the core network (1006) and / or one or more other intermediate networks (e.g., one or more public, private, or hosted networks). For example, an intermediate network could be a backbone network or the Internet.

[0244] UE 1506 includes hardware and software, the software being stored in or accessible by UE 1506 and executable by the UE's processing circuitry. This software includes client applications (e.g., a web browser or operator-specific "application") operable to provide services to human or non-human users via UE 1506, supported by host 1502. In host 1502, the executing host application can communicate with the executing client application via OTT connection 1550, which terminates between UE 1506 and host 1502. When providing services to a user, the UE's client application can receive request data from the host application of the host and, in response to the request data, provide user data. OTT connection 1550 can transmit both request data and user data. The UE's client application can interact with the user to generate user data provided to the host application via OTT connection 1550.

[0245] OTT connection 1550 can be extended via connection 1560 between host 1502 and network node 1504 and via wireless connection 1570 between network node 1504 and UE 1506 to provide connectivity between host 1502 and UE 1506. Connection 1560 and wireless connection 1570, which provide OTT connection 1550, have been abstractly drawn to illustrate communication between host 1502 and UE 1506 via network node 1504, without explicitly involving any intermediate devices and the precise routing of messages via these devices.

[0246] As an example of sending data via OTT connection 1550, in step 1508, host 1502 provides user data, which can be performed by executing a host application. In some embodiments, the user data is associated with a specific human user interacting with UE 1506. In other embodiments, the user data is associated with UE 1506, which shares data with host 1502 without explicit human interaction. In step 1510, host 1502 initiates a transmission to UE 1506 carrying the user data. Host 1502 may initiate the transmission in response to a request sent by UE 1506. This request may be caused by human interaction with UE 1506 or by the operation of a client application executed on UE 1506. According to the teachings of the embodiments described throughout this disclosure, this transmission may be carried out via network node 1504. Therefore, in step 1512, in accordance with the teachings of the embodiments described throughout this disclosure, network node 1504 sends user data carried in a transmission initiated by host 1502 to UE 1506. In step 1514, UE 1506 receives the user data carried in the transmission, which can be performed by a client application running on UE 1506, associated with a host application running by host 1502.

[0247] In some examples, UE 1506 executes a client application that provides user data to host 1502. User data can be provided as a reaction or response to data received from host 1502. Therefore, in step 1516, UE 1506 can provide user data, which can be done by executing the client application. When providing user data, the client application may also consider user input received from a user via the input / output interface of UE 1506. Regardless of the specific manner in which user data is provided, UE 1506 initiates the transmission of user data to host 1502 via network node 1504 in step 1518. In step 1520, in accordance with the teachings of the embodiments described throughout this disclosure, network node 1504 receives user data from UE 1506 and initiates the transmission of the received user data to host 1502. In step 1522, host 1502 receives the user data carried in the transmission initiated by UE 1506.

[0248] One or more of the various embodiments improve the performance of OTT services provided to UE 1506 using OTT connection 1550, in which wireless connection 1570 forms the final part. For example, by providing the owner / producer of the AI / ML model with the ability to protect the AI / ML model during various transmission, storage, and retrieval scenarios, the embodiments improve the security of confidential and / or sensitive AI / ML models, thereby facilitating the deployment of such models in multi-vendor communication networks (e.g., 5GC). In this way, the embodiments facilitate the use of deployed AI / ML models to improve network performance, thereby increasing the value of OTT services delivered through such an improved network.

[0249] In the example scenario, host 1502 can collect and analyze plant status information. As another example, host 1502 can process audio and video data that may have been retrieved from the UE for creating mappings. As another example, host 1502 can collect and analyze real-time data to help control vehicle congestion (e.g., control traffic lights). As another example, host 1502 can store monitoring video uploaded by the UE. As another example, host 1502 can store or control access to media content such as video, audio, VR, or AR, which can be broadcast, multicast, or unicast to the UE. As other examples, host 1502 can be used for energy pricing, remote control of non-time-critical power loads to balance generation demand, location services, presentation services (e.g., compiling charts based on data collected from remote devices), or any other function that collects, retrieves, stores, analyzes, and / or transmits data.

[0250] In some examples, a measurement process may be provided for monitoring data rate, latency, and other factors that are the object of improvement in one or more embodiments. Optional network functions may also be present for reconfiguring the OTT connection 1550 between host 1502 and UE 1506 in response to changes in measurement results. The measurement process and / or the network functions for reconfiguring the OTT connection may be implemented in the software and hardware of host 1502 and / or UE 1506. In some embodiments, sensors (not shown) may be deployed in or associated with other devices traversed by the OTT connection 1550; the sensors may participate in the measurement process by providing values ​​of the monitored quantities exemplified above or by providing values ​​of other physical quantities from which the software can calculate or estimate the monitored quantities. Reconfiguring the OTT connection 1550 may include message formats, retransmission settings, preferred routing, etc.; reconfiguration does not require a direct change to the operation of network node 1504. Such processes and functions may be known and practiced in the art. In some embodiments, the measurement may involve proprietary UE signaling that facilitates host 1502's measurement of throughput, propagation time, latency, etc. Measurements can be achieved by having the software use an OTT connection 1550 to send messages (especially empty or “virtual” messages) while monitoring propagation time, errors, etc.

[0251] The foregoing merely illustrates the principles of this disclosure. Various modifications and alterations to the described embodiments will be apparent to those skilled in the art in light of the teachings herein. Therefore, it should be understood that those skilled in the art will be able to design numerous systems, arrangements, and programs that, while not expressly shown or described herein, embody the principles of this disclosure and are therefore within its spirit and scope. As will be understood by those skilled in the art, the various embodiments can be used together and interchangeably.

[0252] As used herein, the terminology may have a conventional meaning in the field of electronic, electrical and / or electronic equipment, and may include, for example, electrical and / or electronic circuits, devices, modules, processors, memories, logic solid-state and / or discrete devices, computer programs or instructions for performing various tasks, processes, calculations, outputs and / or display functions, such as those described herein.

[0253] Any suitable steps, methods, features, functions, or benefits disclosed herein can be performed by one or more functional units or modules of one or more virtual devices. Each virtual device may include multiple such functional units. These functional units may be implemented by processing circuitry, which may include one or more microprocessors or microcontrollers and other digital hardware (which may include digital signal processors (DSPs), application-specific digital logic, etc.). The processing circuitry may be configured to execute program code stored in memory, which may include one or more types of memory, such as read-only memory (ROM), random access memory (RAM), cache memory, flash memory devices, optical storage devices, etc. The program code stored in the memory includes program instructions for executing one or more telecommunications and / or data communication protocols and instructions for executing one or more techniques described herein. In some implementations, the processing circuitry may be used to cause the various functional units to perform corresponding functions according to one or an embodiment of this disclosure.

[0254] As described herein, devices and / or apparatuses may be represented by semiconductor chips, chipsets, or (hardware) modules including such chips or chipsets; however, this does not preclude the possibility that the functionality of a device or apparatus may be implemented as a software module (e.g., including a computer program or computer program product comprising executable software code portions for execution or running on a processor). Furthermore, the functionality of a device or apparatus may be implemented by any combination of hardware and software. A device or apparatus may also be considered as a combination of multiple devices and / or apparatuses, whether they functionally cooperate with each other or are independent of each other. Moreover, devices and apparatuses may be implemented in a distributed manner throughout a system, provided that the functionality of the device or apparatus is preserved. This principle and similar principles are considered to be known to those skilled in the art.

[0255] Unless otherwise defined, all terms used herein, including technical and scientific terms, shall have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure pertains. It will be further understood that the terms used herein should be interpreted in accordance with their meaning in the context of this specification and related art, and not in an ideal or overly formal sense, unless so explicitly defined herein.

[0256] Additionally, certain terms used in this disclosure (including the specification and drawings) may be used synonymously in certain instances (e.g., "data" and "information"). It should be understood that while these terms and / or their terms that may be synonymous with each other may be used synonymously herein, there may be instances where such words are intended not to be used synonymously.

[0257] Example embodiments of the technologies and apparatus described herein include, but are not limited to, the following listed embodiments:

[0258] A1. A method for a consumer network function (NFc) in a communication network, the method comprising:

[0259] Send a first request to the first NF of the communication network for a first access token associated with a machine learning (ML) model, wherein the first request includes one or more of the following associated with the ML model: analysis identifier (ID) and interoperability ID;

[0260] Receive a first response from the first NF, including the first access token;

[0261] A second request for the ML model is sent to the producer NF (NFp) of the communication network, wherein the second request includes at least one of an analysis ID and an interoperability ID, and a first access token; and

[0262] Receive a second response from NFp that includes one of the following:

[0263] ML model, or

[0264] A generic resource locator (URL) for the storage resources associated with the second NF of the communication network, from which the ML model can be obtained.

[0265] A2. The method according to embodiment A1, wherein the first NF is one of the following: Network Repository Function (NRF) or Analytics Data Repository Function (ADRF).

[0266] A3. The method according to any one of embodiments A1 to A2, wherein one or more of the following are applicable:

[0267] NFC stands for Network Data Analysis Function (NWDAF) (AnLF); and

[0268] NFp is the model training logic function NWDAF (MTLF) for network data analysis.

[0269] A4. The method according to any one of embodiments A1 to A3, wherein:

[0270] The second response includes an encrypted ML model; and

[0271] The second response also includes information that can be used to locate the key that can be used to decrypt and verify the ML model.

[0272] A5. The method according to any one of embodiments A1 to A3, wherein:

[0273] The second response includes the URL; and

[0274] The method also includes:

[0275] Send a third request to the first NF for a second access token associated with the ML model, wherein the third request includes one or more of the analysis ID and interoperability ID, as well as a URL;

[0276] Receive a third response from the first NF, including the second access token; and

[0277] Obtain the ML model from the second NF using the second access token and URL.

[0278] A6. The method according to embodiment A5, wherein the second NF from which the ML model is obtained using a URL is one of the following: the producer NF of the communication network or the analytical data repository function (ADRF).

[0279] A7. The method according to any one of embodiments A5 to A6, wherein the URL is encrypted, and the second response further includes information that can be used to locate a key that can be used to decrypt and verify the URL.

[0280] B1. A method for a producer network function (NFp) in a communication network, the method comprising:

[0281] Register information associated with machine learning (ML) models in the Network Repository Function (NRF) of the communication network, where:

[0282] ML models are generated, owned, and / or maintained by NFp, and

[0283] The registered information includes the following: analysis identifier (ID); interoperability ID; and a first generic resource locator (URL) for the storage resource associated with NFp, from which the ML model can be obtained; and

[0284] The ML model is encrypted and a first request for storing the encrypted ML model is sent to the Analytical Data Repository Function (ADRF) of the communication network, wherein the first request includes the encrypted ML model or a first URL.

[0285] B2. The method according to embodiment B1 further includes:

[0286] A second request for an ML model is received from a consumer NF (NFc) in the communication network, wherein the second request includes at least one of an analysis ID and an interoperability ID, and a first access token; and

[0287] Based on the verification of the first access token, a second response is sent to the NFc, which includes one of the following: the ML model; the first URL; or a second URL of the storage resource associated with the ADRF from which the ML model can be obtained.

[0288] B3. The method according to embodiment B2, wherein:

[0289] The first request includes the first URL; and

[0290] The second response may include either the first URL or the second URL.

[0291] B4. The method according to embodiment B3, wherein one or more of the following are applicable:

[0292] The first URL in the first request is encrypted, and the first request also includes information that can be used to locate a key that can be used to decrypt and verify the first URL; and

[0293] The first or second URL included in the second response is encrypted, and the second response also includes information that can be used to locate a key that can be used to decrypt and verify the first or second URL.

[0294] B5. The method according to any one of embodiments B3 to B4 further includes:

[0295] Receive another request for the ML model from ADRF, wherein the other request includes a first URL and a second access token;

[0296] Based on the verification of the second access token, another response including an encrypted ML model is sent to the ADRF; and

[0297] The second URL is then received from the ADRF and the registration information in the NRF is updated to include the second URL.

[0298] B6. The method according to any one of embodiments B3 to B5, wherein:

[0299] The second response includes the first URL; and

[0300] The method also includes:

[0301] Receive a third request for the ML model from the NFC, wherein the third request includes the following items: a third access token associated with the ML model, a first URL, and one or more of an analytics ID and an interoperability ID; and

[0302] Based on the verification of the third access token, a third response including the ML model is sent to the NFc.

[0303] B7. The method according to embodiment B6, wherein the ML model in the third response is encrypted, and the third response further includes information that can be used to locate a key that can be used to decrypt and verify the ML model.

[0304] B8. The method according to embodiment B2 further includes:

[0305] Send a fourth request for an access token associated with the ML model to the Network Repository Function (NRF) of the communication network, wherein the fourth request includes one or more of the analysis ID and interoperability ID;

[0306] Receive the requested access token from the NRF;

[0307] Send a fifth request for the ML model to ADRF, wherein the fifth request includes one or more of the analysis ID and interoperability ID and the received access token;

[0308] Receive a fifth response from ADRF that includes the ML model, and then include the ML model in the second response to NFc.

[0309] B9. The method according to any one of embodiments B1 to B8, wherein one or more of the following are applicable:

[0310] NFC stands for Network Data Analysis Function (NWDAF) (AnLF); and

[0311] NFp is the model training logic function NWDAF (MTLF) for network data analysis.

[0312] C1. A method for a network repository function (NRF) of a communication network, the method comprising:

[0313] Registering information associated with a machine learning (ML) model generated, owned, and / or maintained by a producer network function (NFp) of a communication network, wherein the registered information includes the following items associated with the ML model: analysis identifier (ID); interoperability ID; and a first generic resource locator (URL) of a storage resource associated with the NFp, from which the ML model can be obtained; and

[0314] Receive a first request from the consumer NF (NFc) of the communication network for a first access token associated with the ML model, wherein the first request includes one or more of the analysis ID and interoperability ID;

[0315] Send a first response, including the first access token, to NFc.

[0316] C2. The method according to embodiment C1 further includes:

[0317] Receive a second request for a second access token from the first NF of the communication network, wherein the second request includes the following items:

[0318] Analysis ID and one or more of the interoperability ID; and

[0319] A first URL or a second URL of a storage resource associated with the Analytical Data Repository Function (ADRF) of the communication network, from which the ML model can be obtained; and

[0320] Send a second response, including the second access token, to the first NF.

[0321] C3. The method according to embodiment C2, wherein the first NF is one of the following: NFc or NFp.

[0322] C4. The method according to any one of embodiments C2 to C3 further includes: after registering information associated with the ML model, updating the registered information to include a second URL.

[0323] C5. The method according to any one of embodiments C1 to C4, wherein one or more of the following are applicable:

[0324] NFC stands for Network Data Analysis Function (NWDAF) (AnLF); and

[0325] NFp is the model training logic function NWDAF (MTLF) for network data analysis.

[0326] D1. A method for an Analytical Data Repository Function (ADRF) for a communication network, the method comprising:

[0327] A first request for storing a cryptographic machine learning (ML) model is received from the producer network function (NFp) of the communication network, wherein the first request includes a first generic resource locator (URL) for the cryptographic ML model or a storage resource associated with the NFp, from which the cryptographic ML model can be obtained;

[0328] Store the encrypted ML model in storage resources associated with ADRF; and

[0329] Send a first response to NFp, which includes a second URL of the storage resource associated with ADRF.

[0330] D2. The method according to embodiment D1, wherein:

[0331] The first request includes the first URL; and

[0332] The method also includes:

[0333] Send another request for the ML model to NFp, wherein the second request includes a first URL and a second access token; and

[0334] Receive another response from NFp, including the encrypted ML model.

[0335] D3. The method according to embodiment D1 further includes:

[0336] A second request for the ML model is received from the first NF of the communication, wherein the second request includes one or more of the analysis ID and interoperability ID and a third access token;

[0337] Based on the verification of the third access token, a second response including the ML model is sent to the first NF.

[0338] D4. The method according to embodiment D3, wherein the first NF is one of the following: the NFc of the communication network or the consumer NFc.

[0339] D5. The method according to embodiment D4, wherein one or more of the following are applicable:

[0340] NFC stands for Network Data Analysis Function (NWDAF) (AnLF); and

[0341] NFp is the model training logic function NWDAF (MTLF) for network data analysis.

[0342] D6. The method according to any one of embodiments D3 to D5, wherein the ML model in the second response is encrypted, and the second response further includes information that can be used to locate a key that can be used to decrypt and verify the ML model.

[0343] E1. A consumer network function (NFc) for a communication network, wherein:

[0344] NFc is implemented by operatively coupled communication interface circuitry and processing circuitry, and

[0345] The processing circuit and interface circuit are configured to perform operations corresponding to any of the methods described in embodiments A1 to A7.

[0346] E2. A consumer network function (NFc) of a communication network, the NFc being configured to perform operations corresponding to any of the methods described in embodiments A1 to A7.

[0347] E3. A non-transitory computer-readable medium storing computer-executable instructions that, when executed by processing circuitry associated with a consumer network function (NFc) of a communication network, configure the NFc to perform operations corresponding to any of the methods described in embodiments A1 to A7.

[0348] E4. A computer program product including computer-executable instructions that, when executed by processing circuitry associated with a consumer network function (NFc) of a communication network, configure the NFc to perform operations corresponding to any of the methods described in embodiments A1 to A7.

[0349] F1. A producer network function (NFp) for a communication network, wherein:

[0350] NFp is implemented by operatively coupled communication interface circuitry and processing circuitry; and

[0351] The processing circuit and interface circuit are configured to perform operations corresponding to any of the methods described in embodiments B1 to B9.

[0352] F2. A producer network function (NFp) for a communication network, the NFp being configured to perform operations corresponding to any of the methods described in embodiments B1 to B9.

[0353] F3. A non-transitory computer-readable medium storing computer-executable instructions that, when executed by processing circuitry associated with a producer network function (NFp) of a communication network, configure the NFp to perform operations corresponding to any of the methods described in embodiments B1 to B9.

[0354] F4. A computer program product including computer-executable instructions that, when executed by processing circuitry associated with a producer network function (NFp) of a communication network, configure the NFp to perform operations corresponding to any of the methods described in embodiments B1 to B9.

[0355] G1. A network repository function (NRF) for a communication network, wherein:

[0356] NRF is implemented by operatively coupled communication interface circuitry and processing circuitry, and

[0357] The processing circuit and interface circuit are configured to perform operations corresponding to any of the methods described in embodiments C1 to C5.

[0358] G2. A network repository function (NRF) for a communication network, the NRF being configured to perform operations corresponding to any of the methods described in embodiments C1 to C5.

[0359] G3. A non-transitory computer-readable medium storing computer-executable instructions that, when executed by processing circuitry associated with a network repository function (NRF) of a communication network, configure the NRF to perform operations corresponding to any of the methods described in embodiments C1 to C5.

[0360] G4. A computer program product including computer-executable instructions that, when executed by processing circuitry associated with a network repository function (NRF) of a communication network, configure the NRF to perform operations corresponding to any of the methods described in embodiments C1 to C5.

[0361] H1. An analytical data repository function (ADRF) for a communication network, wherein:

[0362] ADRF is implemented by operatively coupled communication interface circuitry and processing circuitry, and

[0363] The processing circuit and interface circuit are configured to perform operations corresponding to any of the methods described in embodiments D1 to D6.

[0364] H2. An analytical data repository function (ADRF) for a communication network, the ADRF being configured to perform operations corresponding to any of the methods described in embodiments D1 to D6.

[0365] H3. A non-transitory computer-readable medium storing computer-executable instructions that, when executed by processing circuitry associated with an analytical data repository function (ADRF) of a communication network, configure the ADRF to perform operations corresponding to any of the methods described in embodiments D1 to D6.

[0366] H4. A computer program product including computer-executable instructions that, when executed by processing circuitry associated with an analytical data repository function (ADRF) of a communication network, configure the ADRF to perform operations corresponding to any of the methods described in embodiments D1 to D6.

Claims

1. A method for a consumer network function (NFc) of a communication network, the method comprising: Send (610) a first request for a first access token associated with a machine learning ML model to a first network function NF of the communication network, wherein the first request includes at least one of the following associated with the ML model: analysis identifier ID and interoperability ID; Receive (620) a first response including the first access token from the first NF; Send (630) a second request for the ML model to the producer NF "NFp" of the communication network, wherein the second request includes at least one of the analysis ID and the interoperability ID and the first access token; and Receive a second response from the NFp (640) including one or more of the following: The ML model, The identifier of the ML model, and The address of the storage resource associated with the second NF of the communication network, from which the ML model can be obtained.

2. The method according to claim 1, wherein, The first NF is one of the following: Network Repository Function (NRF); or Analytics Data Repository Function (ADRF).

3. The method according to any one of claims 1 to 2, wherein, One or more of the following apply: The NFC is the analysis logic function NWDAF (AnLF) of the network data analysis function; and The NFp is the model training logic function NWDAF (MTLF) of the network data analysis function.

4. The method according to any one of claims 1 to 3, wherein: The second response includes the ML model, which is encrypted; and The second response also includes information that can be used to locate a key, which can be used to decrypt and verify the ML model.

5. The method according to any one of claims 1 to 3, wherein: The second response includes the address of the storage resource associated with the second NF; as well as The method further includes: Send (650) a third request to the first NF for a second access token associated with the ML model, wherein the third request includes the following: the address of the storage resource associated with the second NF, and at least one of the analytics ID and the interoperability ID; Receive (660) a third response including the second access token from the first NF; and The ML model (670) is obtained from the second NF using the second access token and the address of the storage resource associated with the second NF.

6. The method according to claim 5, wherein, The address of the storage resource is encrypted, and the second response also includes information that can be used to locate a key, which can be used to decrypt and verify the address of the storage resource.

7. The method according to any one of claims 1 to 6, wherein, The address of the storage resource associated with the second NF is a Universal Resource Locator (URL) or a Fully Qualified Domain Name (FQDN).

8. The method according to any one of claims 1 to 7, wherein, The second NF is one of the following: the NFp of the communication network or the analytical data repository function ADRF.

9. A method for a producer network function (NFp) in a communication network, the method comprising: Register (710) information associated with machine learning ML models in the Network Repository Function (NRF) of the communication network, wherein: The ML model is generated, owned, and / or maintained by the NFp, and The registration information associated with the ML includes an analytics identifier ID and an interoperability ID; and The ML model is encrypted (720) and a first request for storing the encrypted ML model is sent to the Analysis Data Repository Function (ADRF) of the communication network, wherein the first request includes the encrypted ML model or a first address of a storage resource associated with the NFp from which the ML model can be obtained.

10. The method of claim 9, further comprising: A second request for the ML model is received (750) from the Consumer Network Function (NFc) of the communication network, wherein the second request includes at least one of the analysis ID and the interoperability ID, and a first access token; and Based on the verification of the first access token, a second response (780) is sent to the NFc, the second response including one or more of the following: The ML model, The identifier of the ML model, The first address of the storage resource associated with the NFp, or The second address of the storage resource associated with the ADRF, from which the ML model can be obtained.

11. The method of claim 10, wherein: The first address of the storage resource associated with the NFp is a first generic resource locator (URL); and The second address of the storage resource associated with the ADRF is a second URL or a fully qualified domain name (FQDN).

12. The method according to any one of claims 10 to 11, wherein: The first request includes a first address of the storage resource associated with the NFp; and The second response includes a first address of the storage resource associated with the NFp or a second address of the storage resource associated with the ADRF.

13. The method according to claim 12, wherein, One or more of the following apply: The first address included in the first request is encrypted, and the first request also includes information that can be used to locate a key, which can be used to decrypt and verify the first address; and The first address or the second address included in the second response is encrypted, and the second response also includes information that can be used to locate a key that can be used to decrypt and verify the first address or the second address.

14. The method according to any one of claims 12 to 13, further comprising: Receive (730) another request for the ML model from the ADRF, wherein the other request includes a second access token and a first address of the storage resource associated with the NFp; Based on the verification of the second access token, another response including an encrypted ML model is sent to the ADRF (735); and Then, the second address of the storage resource associated with the ADRF is received (740) from the ADRF.

15. The method according to claim 14, wherein, The registration information associated with the ML model also includes a first address of the storage resource associated with the NFp, and the method further includes updating (745) the registration information associated with the ML model in the NRF to include the received second address.

16. The method according to any one of claims 12 to 15, wherein: The second response includes a first address of the storage resource associated with the NFp; as well as The method further includes: A third request for the ML model is received (790) from the NFc, wherein the third request includes at least one of the analytics ID and the interoperability ID, the first address, and a third access token associated with the ML model; and Based on the verification of the third access token, a third response including the ML model is sent to the NFc (795).

17. The method according to claim 16, wherein, The ML model included in the third response is encrypted, and the third response also includes information that can be used to locate a key that can be used to decrypt and verify the ML model.

18. The method according to any one of claims 10 to 11, further comprising: Send (755) a fourth request for an access token associated with the ML model to the Network Repository Function (NRF) of the communication network, wherein the fourth request includes at least one of the analytics ID and the interoperability ID; Receive the requested access token from the NRF (760); Send (765) a fifth request for the ML model to the ADRF, wherein the fifth request includes at least one of the analytics ID and the interoperability ID and the received access token; The ADRF receives (770) a fifth response including the ML model, which is then included in the second response given to the NFc.

19. The method according to any one of claims 9 to 18, wherein, One or more of the following apply: The NFC is the analysis logic function NWDAF (AnLF) of the network data analysis function; and The NFp is the model training logic function NWDAF (MTLF) of the network data analysis function.

20. A method for a Network Repository Function (NRF) of a communication network, the method comprising: Registration (810) of information associated with machine learning (ML) models generated, owned, and / or maintained by the producer network function (NFp) of the communication network, wherein the registration information associated with the ML model includes an analytics identifier ID and an interoperability ID; and Receive (830) a first request for a first access token associated with the ML model from the Consumer Network Function (NFc) of the communication network, wherein the first request includes at least one of the analytics ID and the interoperability ID; Send (840) a first response including the first access token to the NFc.

21. The method of claim 20, further comprising: A second request for a second access token is received (850) from the first network function NF of the communication network, wherein the second request includes the following: At least one of the analysis ID and the interoperability ID; and One of the following: a first address of a storage resource associated with the NFp, from which the ML model can be obtained; or a second address of a storage resource associated with the Analysis Data Repository Function (ADRF) of the communication network, from which the ML model can be obtained; and Send (860) a second response including the second access token to the first NF.

22. The method according to claim 21, wherein: The first address of the storage resource associated with the NFp is a first generic resource locator (URL); and The second address of the storage resource associated with the ADRF is a second URL or a fully qualified domain name (FQDN).

23. The method according to any one of claims 21 to 22, wherein, The first NF is either NFc or NFp.

24. The method according to any one of claims 21 to 23, wherein, The registration information associated with the ML model also includes a first address of the storage resource associated with the NFp, and the method further includes updating (820) the registration information to include a second address of the storage resource associated with the ADRF.

25. The method according to any one of claims 21 to 24, wherein, One or more of the following apply: The NFC is the analysis logic function NWDAF (AnLF) of the network data analysis function; and The NFp is the model training logic function NWDAF (MTLF) of the network data analysis function.

26. A method for an Analysis Data Repository Function (ADRF) for a communication network, the method comprising: A first request for storing an encrypted machine learning (ML) model is received (910) from the producer network function (NFp) of the communication network, wherein the first request includes the encrypted ML model or a first address of a storage resource associated with the NFp from which the encrypted ML model can be obtained; The encrypted ML model is stored (940) in a storage resource associated with the ADRF; and Send a (950) first response to the NFp, the first response including a second address of the storage resource associated with the ADRF.

27. The method of claim 26, wherein: The first request includes a first address of the storage resource associated with the NFp; as well as The method further includes: Send (920) another request for the ML model to the NFp, wherein the other request includes the first address and the second access token; and Receive (930) another response from the NFp including the encrypted ML model, which is then stored in the storage resource associated with the ADRF.

28. The method of claim 26, further comprising: A second request for the ML model is received (960) from the first network function NF of the communication network, wherein the second request includes at least one of the analysis ID and the interoperability ID and a third access token; Based on the verification of the third access token, a second response including the ML model is sent to the first NF (970).

29. The method according to claim 28, wherein, The first NF is either the NFp of the communication network or the consumer NF"NFc".

30. The method according to claim 29, wherein, One or more of the following apply: The NFC is the analysis logic function NWDAF (AnLF) of the network data analysis function; and The NFp is the model training logic function NWDAF (MTLF) of the network data analysis function.

31. The method according to any one of claims 28 to 30, wherein, The ML model included in the second response is encrypted, and the second response also includes information that can be used to locate a key, which can be used to decrypt and verify the ML model.

32. The method according to any one of claims 26 to 31, wherein: The first address of the storage resource associated with the NFp is a first generic resource locator (URL); and The second address of the storage resource associated with the ADRF is a second URL or a fully qualified domain name (FQDN).

33. A consumer network function NFc (410, 510, 1008, 1200, 1402) configured to operate in a communication network (198, 200, 1006), wherein: The NFC is implemented by processing circuitry (1202, 1404) and communication interface circuitry (1206, 1404) that are operatively coupled together, and The processing circuit and the communication interface circuit are configured as follows: Send a first request for a first access token associated with a machine learning ML model to a first network function NF (420, 440, 520, 540, 1008, 1200, 1402) of the communication network, wherein the first request includes one or more of the following associated with the ML model: analysis identifier ID; and interoperability ID; Receive a first response including the first access token from the first NF; A second request for the ML model is sent to the producer NF"NFp" (430, 530, 1008, 1200, 1402) of the communication network, wherein the second request includes at least one of the analysis ID and the interoperability ID, and the first access token; and Receive a second response from the NFp, the second response comprising one or more of the following: The ML model, The identifier of the ML model, and The address of the storage resource associated with the second NF (430, 440, 530, 540, 1008, 1200, 1402) of the communication network, from which the ML model can be obtained.

34. The NFC according to claim 33, wherein, The processing circuit and the communication interface circuit are further configured to perform operations corresponding to the method according to any one of claims 2 to 8.

35. A producer network function NFp (430, 530, 1008, 1200, 1402) configured to operate in a communication network (198, 200, 1006), wherein: The NFp is implemented by processing circuitry (1202, 1404) and communication interface circuitry (1206, 1404) that are operatively coupled together, and The processing circuit and the communication interface circuit are configured as follows: Information associated with machine learning (ML) models is registered in the Network Repository Functions (NRFs) (420, 520, 1008, 1200, 1402) of the communication network, wherein: The ML model is generated, owned, and / or maintained by the NFp, and The registration information associated with the ML includes an analytics identifier ID and an interoperability ID; and The ML model is encrypted and a first request for storing the encrypted ML model is sent to the Analysis Data Repository Function (ADRF) (440, 540, 1008, 1200, 1402) of the communication network. The first request includes the encrypted ML model or a first address of a storage resource associated with the NFp from which the ML model can be obtained.

36. A network repository function NRF (420, 520, 1008, 1200, 1402) configured to operate in a communication network (198, 200, 1006), wherein: The NRF is implemented by processing circuitry (1202, 1404) and communication interface circuitry (1206, 1404) that are operatively coupled together, and The processing circuit and the communication interface circuit are configured as follows: Registration information associated with machine learning (ML) models generated, owned, and / or maintained by the producer network function (NFp) (430, 530, 1008, 1200, 1402) of the communication network, wherein the registration information associated with the ML model includes an analytics identifier ID and an interoperability ID; and A first request for a first access token associated with the ML model is received from the consumer network function NFc (410, 510, 1008, 1200, 1402) of the communication network, wherein the first request includes at least one of the analysis ID and the interoperability ID; Send a first response, including the first access token, to the NFc.

37. An analytical data repository function ADRF (440, 540, 1008, 1200, 1402) is configured to operate in a communication network (198, 200, 1006), wherein: The ADRF is implemented by processing circuitry (1202, 1404) and communication interface circuitry (1206, 1404) that are operatively coupled together, and The processing circuit and the communication interface circuit are configured as follows: A first request for storing an encrypted machine learning (ML) model is received from the producer network function (NFp) (430, 530, 1008, 1200, 1402) of the communication network, wherein the first request includes the encrypted ML model or a first address of a storage resource associated with the NFp, from which the encrypted ML model can be obtained; The encrypted ML model is stored in the storage resource associated with the ADRF; and Send a first response to the NFp, the first response including a second address of the storage resource associated with the ADRF.

38. The ADRF according to claim 37, wherein, The processing circuit and the communication interface circuit are further configured to perform operations corresponding to the method according to any one of claims 27 to 32.

39. A non-transitory computer-readable medium (1204, 1404) storing computer-executable instructions, which, when executed by processing circuitry (1202, 1404) associated with an analytical data repository function ADRF (440, 540, 1008, 1200, 1402), configure the ADRF to perform operations corresponding to the method according to any one of claims 26 to 32, wherein the ADRF (440, 540, 1008, 1200, 1402) is configured to operate in a communication network (198, 200, 1006).

40. A computer program product (1204a, 1404a) including computer-executable instructions, which, when executed by processing circuitry (1202, 1404) associated with an analytical data repository function ADRF (440, 540, 1008, 1200, 1402), configure the ADRF to perform operations corresponding to the method according to any one of claims 26 to 32, wherein the ADRF (440, 540, 1008, 1200, 1402) is configured to operate in a communication network (198, 200, 1006).

Images