Data transmission for artificial intelligence (AI) technology

Abstract

Some techniques are described: triggering, by a first entity, a procedure by which one or more interfaces between the first entity and a second entity are established over a user plane or a control plane, wherein the triggering is based on a first communication of a first message between the first entity and the second entity, and wherein the one or more interfaces are for data communication between the first entity and the second entity; and establishing, by the first entity, the one or more interfaces based on performance of a second communication of a second message between the first entity and the second entity, the second communication being performed in response to the first communication.

Description

TECHNICAL FIELD

[0001] The present disclosure relates generally to digital wireless communications. BACKGROUND

[0002] Mobile telecommunication technology is pushing the world towards an increasingly interconnected and networked society. Next generation systems and wireless communication technologies will need to support a wider range of use case characteristics and provide more complex and sophisticated access requirements and flexibility compared to existing wireless networks.

[0003] Long-Term Evolution (LTE) is a standard for wireless communication for mobile devices and data terminals developed by the 3rd Generation Partnership Project (3GPP). LTE-Advanced (LTE-A) is a wireless communication standard that enhances the LTE standard. The 5th generation wireless system, referred to as 5G, advances the LTE and LTE-A wireless standards and is committed to supporting higher data rates, massive connectivity, ultra-low latency, high reliability, and other emerging service requirements. SUMMARY

[0004] Techniques are disclosed that enable or support data transmission based on Artificial Intelligence (AI) technology. An example wireless communication method includes: triggering, by a first entity, a procedure by which one or more interfaces between the first entity and a second entity are established on a user plane or a control plane, wherein the triggering is based on a first communication of a first message between the first entity and the second entity, and wherein the one or more interfaces are for data communication between the first entity and the second entity; and establishing, by the first entity, the one or more interfaces based on an execution of a second communication of a second message between the first entity and the second entity, the second communication being executed in response to the first communication.

[0005] In some embodiments, the one or more interfaces are established over the user plane, the first message includes a first set of one or more user plane (UP) transport layer information, the second message includes a second set of one or more user plane (UP) transport layer information, and each of the first set of one or more user plane (UP) transport layer information and the second set of one or more user plane (UP) transport layer information indicates an interface and includes a General Packet Radio Service (GPRS) Tunneling Protocol (GTP) tunnel Internet Protocol (IP) address (GTP tunnel IP address) and a General Packet Radio Service (GPRS) Tunneling Protocol (GTP) tunnel endpoint identification (GTP tunnel endpoint identification). In some embodiments, at least one of the first set of one or more user plane (UP) transport layer information or the second set of one or more user plane (UP) transport layer information includes an object identification that identifies an object for which data is transmitted or received over the interface indicated by the at least one user plane (UP) transport layer information. In some embodiments, the object identification includes a user equipment (UE) identification, a cell identification, a slice identification, a base station identification, an artificial intelligence (AI) session identification, a measurement identification, or a Quality of Service (QoS) flow identification.

[0006] In some embodiments, the first set of one or more user plane (UP) transport layer information includes a first set of one or more General Packet Radio Service (GPRS) Tunneling Protocol (GTP) tunnel Internet Protocol (IP) (GTP IP) addresses, the first set of one or more GTP IP addresses being different from a second set of one or more GTP IP addresses included in the second set of one or more UP transport layer information. In some embodiments, the first set of one or more UP transport layer information includes a first set of one or more GTP tunnel endpoint identifiers, the second set of one or more UP transport layer information includes a second set of one or more GTP tunnel endpoint identifiers, and the first set of one or more GTP tunnel endpoint identifiers is different from the second set of one or more GTP tunnel endpoint identifiers or at least some of the first set of one or more GTP tunnel endpoint identifiers is the same as the second set of one or more GTP tunnel endpoint identifiers.

[0007] In some embodiments, the one or more interfaces are established on the control plane, the first message includes a first set of one or more control plane (CP) transport layer addresses, the second message includes a second set of one or more CP transport layer addresses, and each of the first set of one or more CP transport layer addresses and the second set of one or more CP transport layer addresses indicates an interface and includes two Internet Protocol (IP) addresses and one Stream Control Transmission Protocol (SCTP) port number, the IP addresses being associated with establishment of an SCTP connection.

[0008] In some embodiments, at least one of the first set of one or more control plane (CP) transport layer addresses or at least one of the second set of one or more control plane (CP) transport layer addresses includes an object identification that identifies an object for which data is transmitted or received over an interface indicated by the at least one control plane (CP) transport layer address. In some embodiments, the object identification includes a user equipment (UE) identification, a cell identification, a slice identification, a base station identification, an artificial intelligence (AI) session identification, a measurement identification, or a quality of service (QoS) flow identification. In some embodiments, all data communications between the first entity and the second entity are performed over the one or more interfaces established over the user plane, or all data communications between the first entity and the second entity are performed over the one or more interfaces established over the control plane, or at least some data communications between the first entity and the second entity are performed over a first interface established over the user plane and at least some data communications between the first entity and the second entity are performed over a second interface established over the control plane.

[0009] In some embodiments, the data communication is performed over the user plane by using a packet having a pre-defined General Packet Radio Service (GPRS) Tunneling Protocol-User (GTP-U) format. In some embodiments, the packet having the GPRS Tunneling Protocol-User (GTP-U) format includes a GPRS Tunneling Protocol-User (GTP-U) header including one or more of: a payload type indicating whether a payload of the packet includes artificial intelligence (AI)-related data, or an artificial intelligence (AI)-related control message, or data that is not related to artificial intelligence (AI); an artificial intelligence (AI) message type indicating a type of the artificial intelligence (AI)-related control message included in the payload; an encoding or decoding method type identifying a technique used to decode the payload; and an object identity indicating an object related to or associated with the payload. In some embodiments, the object includes a user equipment (UE), a cell, a slice, a base station, an artificial intelligence (AI) session, a measurement session, or a quality of service (QoS) flow. In some embodiments, the data communication is performed over the control plane by using a packet having a Stream Control Transmission Protocol (SCTP) format. In some embodiments, the packet having the SCTP format includes a payload including one or more of: a payload type indicating whether the payload of the packet includes artificial intelligence (AI)-related data, or an artificial intelligence (AI)-related control message, or data that is not related to artificial intelligence (AI); an artificial intelligence (AI) message type indicating a type of the artificial intelligence (AI)-related control message included in the payload; an encoding or decoding method type identifying a technique used to decode the payload; and an object identity indicating an object related to or associated with the payload.

[0010] In some embodiments, the object includes a user equipment (UE), a cell, a slice, a base station, an artificial intelligence (AI) session, a measurement session, or a quality of service (QoS) flow. In some embodiments, the first entity includes a first artificial intelligence (AI) operating entity, and the second entity includes a second artificial intelligence (AI) operating entity. In some embodiments, the first entity or the second entity is located in a radio access network (RAN) network node, or in a core network, or in an operation, administration, and maintenance (OAM) server.

[0011] In yet another example aspect, the method described above is embodied in and stored in the form of processor-executable code on a non-transitory computer-readable storage medium. The code embodied on the computer-readable storage medium, when executed by a processor, causes the processor to implement the method described in the present patent.

[0012] In yet another example embodiment, an apparatus is disclosed that is configured as or operable to perform the method described above.

[0013] The above described and other aspects are more fully described in the accompanying drawings, description, and claims. BRIEF DESCRIPTION OF DRAWINGS

[0014] Figure 1 An example block diagram of an artificial intelligence (AI)-based Radio Access Network (RAN) architecture is shown.

[0015] Figure 2 An example flow diagram for establishing an AI interface between two AI entities on a user plane is shown.

[0016] Figure 3 An example flow diagram for establishing an AI interface between two AI entities on a control plane is shown.

[0017] Figure 4 An example flow diagram for transmitting data between two AI entities on a user plane and / or a control plane is shown.

[0018] Figure 5 An example flow diagram for establishing one or more interfaces between two entities is shown.

[0019] Figure 6 An example block diagram of a hardware platform that can be part of a network node or user equipment is shown. DETAILED DESCRIPTION

[0020] By supporting artificial intelligence (AI) techniques based on big data, intelligent wireless networks can be developed. In addition to the general needs of 5G for energy efficiency, load balancing, mobility management, and coverage optimization, general principles and functional frameworks for AI-based 5G wireless networks can be further developed for intelligent wireless networks.

[0021] An AI entity can be described as a logical functional entity that enables intelligent control and optimization of radio access network (RAN) elements and resources via data collection. AI entities can be located in different geographical locations, integrated in different RAN nodes, or as standalone entities, such as an AI server. Currently, there is no solution to establish interfaces for data input and output between AI entities in a wireless network (e.g., a 3GPP network), nor is there a solution on how to transfer signaling control messages for AI function management, AI model management, and AI measurement management, etc. in a wireless network.

[0022] Figure 1 An example block diagram showing a Next Generation Radio Access Network (NG-RAN) architecture and interfaces that can support AI is shown. Figure 1 An AI entity can be located in a 5G Core Network (5GC), in Operations, Administration and Maintenance (OAM), and in one or more RAN nodes (e.g., one or more eNBs or one or more base stations) is shown.

[0023] In current NG-RAN networks, the Xn interface, the F1 interface (between a Central Unit (CU) and a Distributed Unit (DU)), and the NG interface include user plane interfaces based on GPRS Tunneling Protocol (GTP-U), such as Xn-U, F1-U, NG-U interfaces, and control interfaces based on Stream Control Transmission Protocol (SCTP), such as Xn-C, F1-C, NG-C interfaces. Existing open-standardized backhaul interfaces, such as Xn, F1, NG interfaces, cannot support data input and output between AI entities, nor can they support the transmission of signaling control messages for AI function management, AI model management, and AI measurement management, etc.

[0024] The example titles of the following sections are used to help understand the disclosed subject matter and do not limit the scope of the claimed subject matter in any way. Thus, one or more features of one example section can be combined with one or more features of another example section. Moreover, 5G terminology is used for clarity of explanation, but the technology disclosed herein is not limited to 5G technology and can be used in wireless systems implementing other protocols.

[0025] 1. Example 1 - Establishing an AI interface on the user plane

[0026] Figure 2An example flow diagram for establishing an AI interface between two AI entities over a user plane is shown. In some embodiments, one purpose of the example AI interface establishment procedure is to establish a GTP-U based interface to exchange AI application level data between AI entities. One or more common AI interface instances can be established to transfer all AI related data between two AI entities, and one or more object specific AI interface instances can be established to transfer AI related data for corresponding objects between two AI entities. The steps shown in Figure 2

[0027] Step 1: AI entity 1 sends a request message (e.g., an AI interface establishment request, an Fl establishment request, an Xn establishment request, or an NG establishment request, etc.) to AI entity 2 to establish one or more common AI interfaces to exchange AI application level data between AI entity 1 and AI entity 2. The request message can carry one or more user plane (UP) transport layer information allocated at AI entity 1, where each UP transport layer information can include a GTP tunnel IP address and / or a GTP tunnel endpoint identifier.

[0028] Step 2: After receiving the request message sent by AI entity 1, AI entity 2 server saves the one or more UP transport layer messages sent by AI entity 1, and sends a response message (e.g., an AI interface establishment response, an Fl establishment response, an Xn establishment response, or an NG establishment response, etc.) to AI entity 1. The response message can carry one or more UP transport layer information allocated at AI entity 2, where each UP transport layer information can include a GTP tunnel IP address and / or a GTP tunnel endpoint identifier.

[0029] ​The GTP tunnel IP address provided by AI entity 2 in each UP transport layer information in the response message is different from the GTP tunnel IP address provided by AI entity 1 in each UP transport layer information in the request message. In some embodiments, the GTP tunnel endpoint identifier provided by AI entity 2 in each UP transport layer information in the response message may be exactly the same, partially the same, or different from the GTP tunnel endpoint identifier provided by AI entity 1 in each UP transport layer information in the request message. For example, if AI entity 1 sends a request message with two UP transport layer information {(IP1#, endpoint 1); (IP#2, endpoint 2)}, such a message can indicate to AI entity 2 that AI entity 1 is requesting the establishment of two tunnels and expects AI entity 2 to return the two tunnel addresses assigned at AI entity 2. Continuing with this example, AI entity 2 can send a response message that includes two UP transport layer information messages, which may include {(IP#3, endpoint 1)(IP#3, endpoint 2)} or {(IP#3, endpoint 3)(IP#3, endpoint 4)} or {(IP#3, endpoint 3)(IP#4, endpoint 4)} to establish two tunnels with AI entity 1.

[0030] After steps 1 and 2, one or more user plane GTP tunnels are established for the public AI interface, and both AI entities can transmit the required AI application-level data between them through the AI ​​interface.

[0031] like Figure 1 As shown, in some embodiments, steps 3 and 4, as further described below, may optionally be performed.

[0032] Step 3: AI Entity 1 sends a specific object request message (e.g., UE AI interface establishment request, SN addition request, or UE context establishment request) to AI Entity 2 to establish one or more AI interfaces, thereby exchanging AI-related data for a specific object between AI Entity 1 and AI Entity 2. This request message may carry one or more UP transport layer information and optionally one or more object identifiers. Each object identifier may be associated with one or more UP transport layer information to indicate which object data the corresponding AI interface is used to transmit. Each UP transport layer information may include a GTP tunnel IP address and / or a GTP tunnel endpoint identifier. In some embodiments, the object identifier may include a UE identifier, a cell identifier, a slice identifier, a base station (e.g., eNB or gNB) identifier, an AI session identifier, a measurement identifier, or a QoS flow identifier.

[0033] Step 4: After receiving the specific object request message sent by AI entity 1, AI entity 2 saves or stores the UP transport layer information allocated at AI entity 1 and sends a response message (e.g., UE AI interface setup response, SN addition response, or UE context setup response, etc.) to AI entity 1. The response message can carry one or more UP transport layer information and optionally one or more object identities, each object identity being associated with one or more UP transport layer information to indicate which object data are transferred by the corresponding AI interface, and wherein each UP transport layer information includes a GTP tunnel IP address and / or a GTP tunnel endpoint identity.

[0034] After step 3 and step 4, one or more GTP tunnel AI interfaces for specific objects are established, through which AI entity can transfer AI application level data associated with this object between AI entities.

[0035] 2. Example 2 - Establishing an AI interface on the control plane

[0036] Figure 3 An example flow chart for establishing an AI interface between two AI entities on the control plane is shown. In some embodiments, one purpose of the AI interface establishment procedure is to establish an SCTP based interface to exchange AI application level data between AI entities. One or more common AI interface instances can be established to transfer all AI related data between two AI entities and one or more specific object level AI interface instances can be established to transfer AI related data of corresponding objects between two AI entities. The steps shown in Figure 3 are further described below.

[0037] Step 1: AI entity 1 sends a request message (e.g., AI interface setup request, F1 setup request, Xn setup request, or NG setup request, etc.) to AI entity 2 to establish one or more common AI interfaces to exchange AI application level data between AI entity 1 and AI entity 2. The request message can carry one or more CP transport layer addresses allocated at AI entity 1, wherein each CP transport layer address can include two IP addresses and one SCTP port number for an SCTP connection.

[0038] Step 2: After receiving the request message sent by AI entity 1, AI entity 2 server saves the one or more CP transport layer addresses sent by AI entity 1 and sends a response message (e.g., an AI interface setup response, an Fl setup response, an Xn setup response, or an NG setup response, etc.) to AI entity 1. The response message carries the one or more CP transport layer addresses stored at AI entity 2, where each CP transport layer address can include two IP addresses and one SCTP port number for an SCTP connection.

[0039] For an AI interface based on an SCTP connection, each AI entity can provide two different IP addresses and one port number for an SCTP connection. The two IP addresses provided by AI entity 2 in each CP transport layer address in the response message can be different from the two IP addresses provided by AI entity 1 in each CP transport layer address in the request message, and the port number provided by AI entity 2 in each CP transport layer address in the response message can be the same as the port number provided by AI entity 1 in each CP transport layer address in the request message. For example, if AI entity 1 sends a request message with CP transport layer address information {(IP#1, IP#2, Port#1); (IP#3, IP#4, Port#2); (IP#5, IP#6, Port#2)}, such a message can indicate to AI entity 2 that AI entity 1 is requesting to establish three SCTP connections (or three control plane AI interfaces) and expects AI entity 2 to return three CP transport layer addresses allocated at AI entity 2. Continuing the example, AI entity 2 can send a response message including three CP transport layer addresses, which can include {(IP#7, IP#8, Port#1); (IP#9, IP#10, Port#2); (IP#11, IP#12, Port#2)} to establish three SCTP connections with AI entity 1.

[0040] After step 1 and step 2, one or more control planes for a common AI interface are established through which AI entities can transfer AI application level data required between AI entities.

[0041] As shown in FIG. 1, in some embodiments, steps 3 and 4 as described further below can be optionally performed. Figure 3

[0042] ​Step 3: AI entity 1 sends a request message (e.g., a UE AI interface setup request, a SN addition request, or a UE context setup request, etc.) to AI entity 2 to establish one or more AI interfaces for exchanging AI-related data for specific objects between AI entity 1 and AI entity 2. The request message can carry one or more CP transport layer addresses, and optionally one or more object identities, each object identity being associated with one or more CP transport layer addresses to indicate which object data is transmitted via the corresponding AI interface, and wherein each CP transport layer address includes two IP addresses and one SCTP port number for an SCTP connection. In some embodiments, the object identity can include a UE identity, or a cell identity, or a slice identity, or a base station (e.g., eNB or gNB) identity, or an AI session identity, or a measurement identity, or a QoS flow identity.

[0043] Step 4: After receiving the request message sent by AI entity 1, AI entity 2 saves or stores the CP transport layer addresses allocated at AI entity 1, and sends a response message (e.g., a UE AI interface setup response, a SN addition response, or a UE context setup response, etc.) to AI entity 1. The response message carries one or more CP transport layer addresses, and optionally one or more object identities, each object identity being associated with one or more CP transport layer addresses to indicate which object data is transmitted via the corresponding AI interface, and wherein each CP transport layer address includes two IP addresses and one SCTP port number for an SCTP connection.

[0044] After step 3 and step 4, one or more control plane AI interfaces for specific objects are established, through which AI entities can transmit AI application level data associated with the object between AI entities.

[0045] 3. Example 3 - Exchanging AI data over the AI interface

[0046] Figure 4 An example flow diagram is shown for transmitting data between two AI entities on the user plane and / or the control plane. AI entities (e.g., AI entity 1 and / or AI entity 2) send / receive all AI-related data on the user plane through one or more AI interfaces; or AI entities send / receive all AI-related data on the control plane through one or more AI interfaces; or AI entities send / receive some AI-related data on the user plane through one or more AI interfaces and send / receive other AI-related data on the control plane through one or more AI interfaces.

[0047] The AI entity can send / receive packets over the AI interface on the user plane, where the packets can be generated using a predefined GTP-U format, and the packets can include at least one of the following information in the GTP-U header:

[0048] • a payload type indicating whether the payload of the packet includes AI-related data, or AI-related control messages, or non-AI data;

[0049] • an AI message type indicating which type of AI-related control message is included in the payload;

[0050] • an encoding / decoding method type indicating how to decode the payload;

[0051] • an object ID that can be used to indicate an object (e.g., the object can be a UE, a cell, a slice, a gNB, an AI session, a measurement session, or a quality of service (QoS) flow) related or associated with the payload.

[0052] The AI entity sends / receives packets over the AI interface on the control plane, where the packets can be generated using a predefined SCTP format, and at least one of the following information can be included in the SCTP payload:

[0053] • a payload type indicating whether the payload of the packet includes AI-related data, or AI-related control messages, or non-AI data;

[0054] • an AI message type indicating which type of AI-related control message is included in the payload;

[0055] • an encoding / decoding method type indicating how to decode the payload;

[0056] • an object ID that can be used to indicate an object (e.g., the object can be a UE, a cell, a slice, a gNB, an AI session, a measurement session, or a QoS flow) related or associated with the payload.

[0057] Figure 5An example flow diagram is shown for establishing one or more interfaces between two entities. Operation 502 includes triggering, by a first entity, a procedure by which one or more interfaces are established between the first entity and a second entity on a user plane or a control plane, where the triggering is based on a first communication of a first message between the first entity and the second entity, and where the one or more interfaces are for data communication between the first entity and the second entity. Operation 504 includes establishing, by the first entity, the one or more interfaces based on performance of a second communication of a second message between the first entity and the second entity in response to the first communication.

[0058] In some embodiments, the one or more interfaces are established on a user plane, the first message includes a first set of one or more user plane (UP) transport layer information, the second message includes a second set of one or more UP transport layer information, and each of the first set and the second set of UP transport layer information indicates a certain interface and includes a general packet radio service (GPRS) tunneling protocol (GTP) tunnel internet protocol (IP) address and a GTP tunnel endpoint identification. In some embodiments, at least one of the first set of UP transport layer information or the second set of UP transport layer information includes an object identification that identifies an object for which data is transmitted or received over the interface indicated by the at least one UP transport layer information. In some embodiments, the object identification includes a user equipment (UE) identification, a cell identification, a slice identification, a base station identification, an artificial intelligence (AI) session identification, a measurement identification, or a quality of service (QoS) flow identification.

[0059] In some embodiments, the first set of one or more UP transport layer information includes a first set of one or more GTP IP addresses that are different from a second set of one or more GTP IP addresses included in the second set of one or more UP transport layer information. In some embodiments, the first set of one or more UP transport layer information includes a first set of one or more GTP tunnel endpoint identifications, the second set of one or more UP transport layer information includes a second set of one or more GTP tunnel endpoint identifications, and the first set of one or more GTP tunnel endpoint identifications is different from the second set of one or more GTP tunnel endpoint identifications or at least some of the first set of one or more GTP tunnel endpoint identifications are the same as the second set of one or more GTP tunnel endpoint identifications.

[0060] In some embodiments, the one or more interfaces are established on a control plane, the first message includes a first set of one or more control plane (CP) transport layer addresses, the second message includes a second set of one or more CP transport layer addresses, and each of the first set and the second set of CP transport layer addresses indicates a certain interface and includes two Internet Protocol (IP) addresses associated with establishment of a Stream Control Transmission Protocol (SCTP) connection and an SCTP port number.

[0061] In some embodiments, at least one of the first set of CP transport layer addresses or the second set of CP transport layer addresses includes an object identification that identifies an object for which data is transmitted or received through the interface indicated by the at least one CP transport layer address. In some embodiments, the object identification includes a user equipment (UE) identification, a cell identification, a slice identification, a base station identification, an artificial intelligence (AI) session identification, a measurement identification, or a quality of service (QoS) flow identification. In some embodiments, all of the data communications between the first entity and the second entity are performed through the one or more interfaces established on the user plane or all of the data communications between the first entity and the second entity are performed through the one or more interfaces established on the control plane or at least some of the data communications between the first entity and the second entity are performed through the first interface established on the user plane and at least some of the data communications between the first entity and the second entity are performed through the second interface established on the control plane.

[0062] In some embodiments, data communication is performed on the user plane using packets with a predefined General Packet Radio Service (GPRS) Tunneling Protocol-User (GTP-U) format. In some embodiments, packets with the GTP-U format include a GTP-U header that includes one or more of the following: payload type, indicating whether the packet's payload includes artificial intelligence (AI) related data, AI-related control messages, or AI-unrelated data; AI message type, indicating which type of AI-related control messages are included in the payload; encoding or decoding method type, identifying the technology used to decode the payload; and object identifier, indicating an object associated with or related to the payload. In some embodiments, the object includes a user equipment (UE), cell, slice, base station, AI session, measurement session, or Quality of Service (QoS) stream. In some embodiments, data communication is performed on the control plane using packets with a Flow Control Transport Protocol (SCTP) format. In some embodiments, a packet having the SCTP format includes a payload that includes one or more of the following: a payload type indicating whether the payload of the packet includes artificial intelligence (AI) related data, AI related control messages, or data unrelated to AI; an AI message type indicating which type of AI related control messages are included in the payload; a type of encoding or decoding method identifying the technology used to decode the payload; and an object identifier indicating an object related to or associated with the payload.

[0063] In some embodiments, the object includes a user equipment (UE), a cell, a slice, a base station, an AI session, a measurement session, or a quality of service (QoS) stream. In some embodiments, the first entity includes an entity performing a first artificial intelligence (AI) operation, and the second entity includes an entity performing a second AI operation. In some embodiments, the first or second entity is located in a radio access network (RAN) node, in the core network, or in an operations, administration, and maintenance (OAM) server.

[0064] Figure 6 An exemplary block diagram of a hardware platform 600 is shown, which may be part of a network node (e.g., an eNB, a base station, or a core network). The hardware platform 600 includes at least one processor 610 and a memory 605 storing instructions thereon. These instructions configure the hardware platform 600 to execute when run by the processor 610. Figures 1 to 5 The operations described in the various embodiments described in this patent document are as follows: Transmitter 615 transmits or sends information or data to another node. For example, a network node transmitter may transmit a message to another network node. Receiver 620 receives information or data transmitted or sent by another node. For example, a base station network node may receive messages from the core network.

[0065] The following sections describe example techniques to enable and / or perform AI data transfer:

[0066] Clause 1 - A method for AI data transfer, comprising: establishing one or more AI interface instances between two AI entities, wherein the AI interface is on a user plane or a control plane; and an AI entity sending AI related data to another AI entity through the AI interface and receiving AI related data from another AI entity through the AI interface.

[0067] Clauses 2 and 3 can establish one or more AI interfaces on a user plane.

[0068] Clause 2 - The method according to clause 1, comprising: the AI entity sending a request message to another AI entity to establish one or more AI interfaces between the two AI entities on a user plane. The request message carries one or more UP transport layer information at the sending end AI entity, and optionally carries one or more object identities, each object identity being associated with one or more UP transport layer information to indicate which object data the corresponding AI interface is used to transmit; and the AI entity receiving a response message from another AI entity, the response message carrying one or more UP transport layer information at the other AI entity, and optionally carrying one or more object identities, each object identity being associated with one or more UP transport layer information.

[0069] Clause 3 - The method according to clause 2, comprising: the UP transport layer information at the AI entity comprises at least one of the following information: GTP tunnel IP address, GTP tunnel endpoint identification.

[0070] Clauses 4-5 can establish one or more AI interfaces on a control plane.

[0071] Clause 4 - The method according to clause 1, comprising: the AI entity sending a request message to another AI entity to establish one or more AI interfaces between the two AI entities on a control plane. The request message carries one or more CP transport layer addresses at the sending end AI entity, and optionally carries one or more object identities, each object identity being associated with one or more CP transport layer addresses to indicate which object data the corresponding AI interface is used to transmit; and the AI entity receiving a response message from another AI entity, the response message carrying one or more CP transport layer addresses at the other AI entity, and optionally carrying one or more object identities, each object identity being associated with one or more CP transport layer addresses.

[0072] Clause 5 - The method according to clause 4, comprising: the CP transport layer address at the AI entity comprises two IP addresses and one SCTP port number, the IP addresses are used for the SCTP connection.

[0073] Clause 6 describes object identification, clause 7 describes the transmission through AI interfaces: only on UP; only on CP; both on UP and CP.

[0074] Clause 6 - The method according to clauses 2, 4, comprising: the object identification is a UE identification, or a cell identification, or a slice identification, or a gNB identification, or an AI session identification, or a measurement identification, or a Qos flow identification.

[0075] Clause 7 - The method according to clause 1, comprising: the AI entity sends / receives all AI related data through one or more AI interfaces on the user plane; or the AI entity sends / receives all AI related data through one or more AI interfaces on the control plane; or the AI entity sends / receives some of the AI related data through one or more AI interfaces on the user plane and sends / receives other of the AI related data through one or more AI interfaces on the control plane.

[0076] Clause 8 describes an example user plane packet format, clause 9 describes an example control plane packet format.

[0077] Clause 8. The method according to clause 1, comprising: the AI entity sends / receives a packet through an AI interface on the user plane, wherein the packet is in GTP-U format and includes at least one of the following information in the GTP-U header:

[0078] • a payload type, which is used to indicate that the payload is AI data, or AI control message, or non-AI data;

[0079] • an AI message type, which is used to indicate which type of AI control message is contained in the payload;

[0080] • an encoding / decoding method type, which is used to indicate how to decode the payload;

[0081] • an object ID, which is used to indicate the object (e.g., UE, cell, slice, gNB, AI session, measurement session, QoS flow) associated with the payload.

[0082] Clause 9 - The method according to clause 2, comprising: the AI entity sends / receives a packet through an AI interface on the control plane, wherein the packet is in SCTP format and includes at least one of the following information in the SCTP payload:

[0083] • Payload Type, which is used to indicate whether the payload is AI data, or AI control packet, or non-AI data;

[0084] • AI Packet Type, which is used to indicate which kind of AI control packet is contained in the payload;

[0085] • Encoding / Decoding Method Type, which is used to indicate how to decode the payload;

[0086] • Object ID, which is used to indicate the object (e.g., UE, cell, slice, gNB, AI session, measurement session, QoS flow) related to the payload.

[0087] In this document, the term “exemplary” is used to mean “an example of” and, unless otherwise noted, does not imply ideal or preferred embodiments.

[0088] Some embodiments described herein are described in the general context of methods or processes, which can be implemented in one embodiment by a computer program product comprising computer-executable instructions, such as program code, contained on a computer-readable medium, which computer program product is executed by a computer in a networked environment. The computer-readable medium can include removable and non-removable storage devices including, but not limited to, Read Only Memory (ROM), Random Access Memory (RAM), compact discs (CDs), digital versatile discs (DVDs), etc. Thus, computer-readable medium can include non-transitory storage media. Generally, program modules can include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. Computer-executable instructions, associated data structures, and program modules represent examples of program code for executing steps of the methods disclosed herein. The particular sequence of such computer-executable instructions or associated data structures represents examples of corresponding acts for implementing the functions described in such steps or processes.

[0089] Some disclosed embodiments can be implemented as devices or modules using hardware circuitry, software, or a combination thereof. For example, hardware circuitry implementations can include discrete analog and / or digital components, such as integrated as part of a printed circuit board. Alternatively or additionally, disclosed components or modules can be implemented as an application specific integrated circuit (ASIC) and / or field programmable gate array (FPGA) device. Some implementations additionally or alternatively include a digital signal processor (DSP), which is a specialized microprocessor optimized for the operation needs of digital signal processing associated with the disclosed functionality of the present application. Similarly, various components or subcomponents within each module can be implemented in software, hardware, or firmware. Connections between the modules and / or components within the modules can be provided using any of the connection methods and media known in the art, including but not limited to communication over the Internet, wired or wireless networks using appropriate protocols.

[0090] Although numerous details have been described herein, these should not be construed as limiting the scope of the claimed application or the scope of potential protection. Rather, the described details are intended to describe particular embodiments and particular features of those embodiments. Certain features described herein in the context of separate embodiments can also be implemented in combination in a single embodiment. Conversely, various features described in the context of a single embodiment can also be implemented separately or in any suitable subcombination. Moreover, although features can be described above as acting in certain combinations and even initially claimed as such, one or more features from a claimed combination can in some cases be removed from the combination and the claimed combination can then be directed to a subcombination or variation of a subcombination. Similarly, while operations are depicted in certain sequences in the drawings, this should not be understood as requiring such sequences or order of operations, or that all illustrated operations be performed, to achieve desirable results.

[0091] Only some implementations and examples are described and other implementations, improvements, and variations can be implemented based on what is described and illustrated in this disclosure.

Claims

1. A method of wireless communication, comprising: triggering, by a first entity, a procedure by which one or more interfaces between the first entity and a second entity are established over a user plane or a control plane wherein the triggering is based on a first communication of a first message between the first entity and the second entity, and wherein the one or more interfaces are for data communication between the first entity and the second entity; and establishing, by the first entity, the one or more interfaces based on an execution of a second communication of a second message between the first entity and the second entity, the second communication being executed in response to the first communication; wherein the first entity and the second entity are each an artificial intelligence (AI) -operated entity, the one or more interfaces are one or more AI interfaces, and wherein the data communication over the one or more interfaces is over a data packet containing at least one AI-specific field in a header or a payload; wherein the one or more interfaces are established over the user plane, wherein the first message includes a first set of one or more user plane (UP) transport layer information, wherein the second message includes a second set of one or more user plane (UP) transport layer information, and wherein each of the first set of one or more user plane (UP) transport layer information and the second set of one or more user plane (UP) transport layer information indicates an interface and includes a General Packet Radio Service (GPRS) Tunneling Protocol (GTP) tunnel Internet Protocol (IP) address and a GTP tunnel endpoint identification; wherein the first message further includes one or more object identifications, each of the object identifications being associated with the one or more UP transport layer information to indicate which object data is transmitted over a corresponding AI interface of the one or more AI interfaces.

2. The method of claim 1, wherein, at least one of the first set of one or more user plane (UP) transport layer information or at least one of the second set of one or more user plane (UP) transport layer information includes an object identification that identifies an object for which data is transmitted or received over an interface indicated by the at least one user plane (UP) transport layer information.

3. The method of claim 2, wherein, the object identification includes a user equipment (UE) identification, a cell identification, a slice identification, a base station identification, an artificial intelligence (AI) session identification, a measurement identification, or a quality of service (QoS) flow identification.

4. The method of claim 1, wherein, The first set of one or more user plane (UP) transport layer information includes a first set of one or more GTP IP addresses, which are different from a second set of one or more GTP IP addresses included in the second set of one or more user plane (UP) transport layer information.

5. The method of claim 1, wherein, The first set of one or more user plane (UP) transport layer information includes a first set of one or more GTP tunnel endpoint identifiers, wherein the second set of one or more user plane (UP) transport layer information includes a second set of one or more GTP tunnel endpoint identifiers, and wherein the first set of one or more GTP tunnel endpoint identifiers are different from the second set of one or more GTP tunnel endpoint identifiers, or at least some of the first set of one or more GTP tunnel endpoint identifiers are the same as the second set of one or more GTP tunnel endpoint identifiers.

6. The method of claim 1, wherein The one or more interfaces are established on the control plane, wherein the first message includes a first set of one or more control plane (CP) transport layer addresses, wherein the second message includes a second set of one or more control plane (CP) transport layer addresses, and wherein each of the first set of one or more control plane (CP) transport layer addresses and the second set of one or more control plane (CP) transport layer addresses indicates an interface and includes two Internet Protocol (IP) addresses and one Stream Control Transmission Protocol (SCTP) port number, the Internet Protocol (IP) addresses being associated with establishment of a Stream Control Transmission Protocol (SCTP) connection.

7. The method of claim 6, wherein, At least one of the first set of one or more control plane (CP) transport layer addresses or at least one of the second set of one or more control plane (CP) transport layer addresses includes an object identification of an object, wherein data for the object is transmitted or received through the interface indicated by the at least one control plane (CP) transport layer address.

8. The method of claim 7, wherein, The object identification includes a user equipment (UE) identification, a cell identification, a slice identification, a base station identification, an artificial intelligence (AI) session identification, a measurement identification, or a quality of service (QoS) flow identification.

9. The method of claim 1, wherein, all data communications between the first entity and the second entity are performed through the one or more interfaces established on the user plane, or wherein all data communications between the first entity and the second entity are performed through the one or more interfaces established on the control plane, or wherein at least some data communications between the first entity and the second entity are performed through a first interface established on the user plane, and at least some data communications between the first entity and the second entity are performed through a second interface established on the control plane.

10. The method of claim 1, wherein, Data communication is performed over the user plane by using packets having a predefined GTP-U format.

11. The method of claim 10, wherein, The packet having the GTP-U format comprises a GTP-U header comprising one or more of: a payload type indicating whether the payload of the packet comprises artificial intelligence (AI) related data, or artificial intelligence (AI) related control messages, or data not related to artificial intelligence (AI), an artificial intelligence (AI) message type indicating a type of artificial intelligence (AI) related control message comprised in the payload, an encoding or decoding method type identifying a technique to decode the payload, and an object identity indicating an object related to or associated with the payload.

12. The method of claim 11, wherein, The object comprises a user equipment (UE), a cell, a slice, a base station, an artificial intelligence (AI) session, a measurement session, or a quality of service (QoS) flow.

13. The method of claim 1, wherein, The data communication is performed over the control plane by using packets having a stream control transmission protocol (SCTP) format.

14. The method of claim 13, wherein, The packet having the stream control transmission protocol (SCTP) format comprises a payload comprising one or more of: a payload type indicating whether the payload of the packet comprises artificial intelligence (AI) related data, or artificial intelligence (AI) related control messages, or data not related to artificial intelligence (AI), an artificial intelligence (AI) message type indicating a type of artificial intelligence (AI) related control message comprised in the payload, an encoding or decoding method type identifying a technique to decode the payload, and an object identity indicating an object related to or associated with the payload.

15. The method of claim 14, wherein the object comprises a user equipment (UE), a cell, a slice, a base station, an artificial intelligence (AI) session, a measurement session, or a quality of service (QoS) flow.

16. The method of any one of claims 1 to 15, wherein, The first entity or the second entity is located in a radio access network (RAN) network node, or in a core network, or in an operation, administration and maintenance (OAM) server.

17. An apparatus for wireless communication, comprising a processor configured to implement a method according to one or more of claims 1 to 16.

18. A non-transitory computer readable program storage medium having stored thereon, a code which, when executed by a processor, causes the processor to implement a method according to one or more of claims 1 to 16.

Images