Device and method for efficient ai computation and distribution by interworking access network and edge data center in wireless communication system

Abstract

According to a first aspect of the present invention, an operation method of a terminal for requesting allocation of AI computation comprises the steps of: transmitting an AI computation allocation request message for a predetermined AI computation to an access network which is connected to the terminal and interworks with each of a plurality of distributedly arranged edge data centers; receiving, from the access network that has received the AI computation allocation request message, allocation of uplink resources for transmitting required information for the AI computation, and transmitting the required information to the access network; and acquiring, via the access network, an AI computation result performed using the required information by one or more edge data centers among the plurality of edge data centers.

Description

Device and method for efficiently performing AI computation and distribution by linking an access network and an edge data center in a wireless communication system.

[0001] The present disclosure relates generally to wireless communication systems, and more particularly to devices and methods for efficiently performing AI computation and distribution by linking an access network and an edge data center in a wireless communication system.

[0002] For reference, this application claims priority to Korean Patent Application No. 10-2024-0119271, filed September 3, 2024. The entire contents of that application, which serves as the basis for this priority claim, are incorporated herein by reference.

[0003] A data center is a physical facility that houses IT infrastructure (e.g., computer systems and hardware) to store, process, and manage vast amounts of digital data and information. Simply put, using a data center is similar to connecting your home desktop to the internet, accessing it remotely via a smartphone or laptop, and utilizing its data and computing resources. Therefore, until now, data centers have been no different from using a typical computer remotely.

[0004] However, the landscape of data centers is rapidly changing. AI services and applications are diversifying, growing rapidly, and their capabilities are rapidly increasing. Therefore, securing data centers equipped with the operational resources to support them is becoming increasingly important.

[0005] Based on the discussion described above, the present disclosure provides a device and method for efficiently performing AI computation and distribution by linking an access network and an edge data center in a wireless communication system.

[0006] In addition, the present disclosure provides a device and method for efficiently performing AI computation and distribution by placing an access network and an edge data center close to each other in a wireless communication system.

[0007] The method for operating a terminal for requesting allocation of an AI operation according to the first aspect of the present invention includes the steps of: transmitting an AI operation allocation request message for a predetermined AI operation to an access network that is linked to and connected to each of a plurality of distributed Edge Data Centers; receiving an uplink resource for transmitting necessary information for the AI ​​operation from the access network that has received the AI ​​operation allocation request message; and transmitting the necessary information to the access network; and obtaining, via the access network, a result of an AI operation performed by at least one of the plurality of Edge Data Centers using the necessary information.

[0008] Each of the above multiple edge data centers may have a limited number of connectable access networks.

[0009] For an edge data center among the plurality of edge data centers whose connection has been disconnected, at least one of the necessary information, information about the access network, and information about the terminal may be prevented from being used after disconnection.

[0010] Between each of the above multiple edge data centers and the access network,

[0011] At least one of the operating status, available resources, and number of connected terminals of each of the plurality of edge data centers may be periodically updated.

[0012] The one or more edge data centers may be selected from among the plurality of edge data centers based on the at least one piece of updated information.

[0013] The one or more edge data centers can perform the AI ​​operation by receiving at least one of an AI model required for the AI ​​operation, information required for the AI ​​operation, and a scope of the AI ​​operation from the terminal.

[0014] The one or more edge data centers may be selected from among the remaining edge data centers, excluding the edge data center that first requested the AI ​​operation but rejected it among the plurality of edge data centers.

[0015] When a handover occurs from a first cell included in the access network to a second cell included in an access network different from the access network in the terminal, the AI ​​operation result can be provided from the access network to the other access network.

[0016] A terminal for requesting allocation of an AI operation according to a second aspect of the present invention comprises a transceiver and a control unit operably connected to the transceiver, wherein the control unit transmits, through the transceiver, an AI operation allocation request message for a predetermined AI operation to an access network connected to each of a plurality of distributed edge data centers and the terminal, receives an uplink resource for transmitting necessary information for the AI ​​operation from the access network that has received the AI ​​operation allocation request message, transmits the necessary information to the access network, and obtains, via the access network, an AI operation result performed by at least one edge data center among the plurality of edge data centers using the necessary information.

[0017] A non-transitory computer-readable recording medium storing at least one computer-executable instruction according to a third aspect of the present invention, wherein the at least one instruction, when executed by a processor, causes the processor to perform a method including the steps of: transmitting an AI operation allocation request message for a predetermined AI operation to each of a plurality of distributed Edge Data Centers and an access network connected to the terminal; receiving an uplink resource for transmitting necessary information for the AI ​​operation from the access network that has received the AI ​​operation allocation request message, and transmitting the necessary information to the access network; and obtaining, via the access network, a result of an AI operation performed by at least one of the plurality of Edge Data Centers using the necessary information.

[0018] A computer program stored in a non-transitory computer-readable recording medium according to a fourth aspect of the present invention, wherein the computer program comprises instructions for causing the processor to perform a method including the steps of: transmitting an AI operation allocation request message for a predetermined AI operation to each of a plurality of distributed Edge Data Centers and an access network connected to the terminal; receiving an uplink resource for transmitting necessary information for the AI ​​operation from the access network that has received the AI ​​operation allocation request message and transmitting the necessary information to the access network; and obtaining, via the access network, a result of an AI operation performed by at least one of the plurality of Edge Data Centers using the necessary information.

[0019] Devices and methods according to various embodiments of the present disclosure enable efficient AI computation and distribution by linking an access network and an edge data center in a wireless communication system.

[0020] The effects that can be obtained from the present disclosure are not limited to the effects mentioned above, and other effects that are not mentioned can be clearly understood by a person having ordinary skill in the art to which the present disclosure belongs from the description below.

[0021] FIG. 1A illustrates an example of an architecture that enables edge computing, according to various embodiments of the present disclosure.

[0022] FIG. 1b illustrates an example of an edge data center interconnection structure according to various embodiments of the present disclosure.

[0023] FIG. 2 illustrates an example of a linkage structure of an access network and a data center according to one embodiment of the present disclosure.

[0024] FIG. 3 illustrates an example of a network structure in an edge data center construction environment according to one embodiment of the present disclosure.

[0025] FIG. 4 illustrates an example of signaling for AI operations in a conventional edge data center interconnection environment, according to various embodiments of the present disclosure.

[0026] FIG. 5 illustrates an example of signaling for AI operations in a case where an access network and an edge data center are directly linked, according to one embodiment of the present disclosure.

[0027] FIG. 6 illustrates a comparison between a conventional scheduling request and a scheduling request for AI computation assistance and delegation, according to one embodiment of the present disclosure.

[0028] FIG. 7 illustrates a signal flow diagram for AI computation assistance and delegation with scheduling request and AICS indicators added to conventional DCI format, according to one embodiment of the present disclosure.

[0029] FIG. 8 illustrates a signal flow diagram for AI computation assistance and delegation without separate indicators using a new AI computation assistance and delegation DCI format A according to one embodiment of the present disclosure.

[0030] FIG. 9 illustrates a signal flow diagram for a case where a terminal hands over from access network X to access network Y, and access network Y is already linked with an edge data center linked with access network X, according to one embodiment of the present disclosure.

[0031] FIG. 10 illustrates a signal flow diagram for a case where a terminal hands over from access network X to access network Y, and access network Y is not linked with an edge data center linked with access network X, according to one embodiment of the present disclosure.

[0032] FIG. 11 illustrates a signal flow diagram for a case where, according to one embodiment of the present disclosure, a terminal hands over from access network X to access network Y, but access network Y is not linked with an edge data center linked with access network X and linkage fails.

[0033] FIG. 12 illustrates a method of operating a terminal according to one embodiment of the present disclosure.

[0034] FIG. 13 illustrates a configuration diagram of a terminal in a wireless communication system according to various embodiments of the present disclosure.

[0035] FIG. 14 illustrates a configuration diagram of an access network in a wireless communication system according to various embodiments of the present disclosure.

[0036] The terms used in this disclosure are used only to describe specific embodiments and may not be intended to limit the scope of other embodiments. The singular expression may include plural expressions unless the context clearly indicates otherwise. Terms used herein, including technical or scientific terms, may have the same meaning as commonly understood by those of ordinary skill in the art described in this disclosure. Terms defined in general dictionaries among the terms used in this disclosure may be interpreted as having the same or similar meaning in the context of the relevant technology, and shall not be interpreted in an idealized or overly formal sense unless explicitly defined in this disclosure. In some cases, even if a term is defined in this disclosure, it cannot be interpreted to exclude embodiments of the present disclosure.

[0037] The various embodiments of the present disclosure described below illustrate a hardware-based approach as an example. However, since the various embodiments of the present disclosure include techniques utilizing both hardware and software, the various embodiments of the present disclosure do not exclude a software-based approach.

[0038] Additionally, in the detailed description and claims of the present disclosure, “at least one of A, B, and C” can mean “only A,” “only B,” “only C,” or “any combination of A, B, and C.” Additionally, “at least one of A, B, or C” or “at least one of A, B, and / or C” can mean “at least one of A, B, and C.”

[0039] The present disclosure relates to a device and method for efficient AI computation and distribution in a wireless communication system. Specifically, the present disclosure describes a technique for efficiently performing AI computation and distribution by linking an access network and an edge data center in a wireless communication system.

[0040] The terms used in the following description, including terms referring to signals, channels, control information, network entities, and device components, are provided for convenience of explanation. Therefore, the present disclosure is not limited to the terms described below, and other terms with equivalent technical meanings may be used.

[0041] Additionally, while this disclosure describes various embodiments using terminology used in certain communication standards (e.g., 3rd Generation Partnership Project (3GPP)), these are merely illustrative examples. The various embodiments of this disclosure can be easily modified and applied to other communication systems.

[0042] A base station may be a network infrastructure that provides wireless access to terminals. A base station may have coverage defined as a certain geographic area based on the distance at which a signal can be transmitted. In addition to a base station, a base station may be referred to as an "access point (AP)", an "eNodeB (eNB)", a "5th generation node", a "wireless point", a "transmission / reception point (TRP)", a "next generation Node B (gNodeB, gNB)", or other terms having equivalent technical meanings.

[0043] Each terminal is a device used by a user and can communicate with a base station via a wireless channel. In some cases, at least one of the terminals can be operated without the user's intervention. That is, at least one of the terminals is a device that performs machine type communication (MTC) and may not be carried by the user. Each of the terminals may be referred to as a 'user equipment (UE)', a 'mobile station', a 'subscriber station', a 'remote terminal', a 'wireless terminal', a 'user device', or other terms having equivalent technical meanings.

[0044] FIG. 1A illustrates an example of an architecture that enables edge computing, according to various embodiments of the present disclosure.

[0045] 3GPP NR (New Radio, 5G) has defined standards to support edge computing functions (TS 23.558, 23.501, 23.502, 23.548, 23.958, 23.222, etc.). However, the current 3GPP standards for edge computing are defined for the wired layer and the application and support layers above it. That is, since the EDN (edge ​​data network) responsible for edge computing is structured to be connected to the back of the core network (CN), only the standards for the wired section have been defined. In other words, the current standard only includes the structure of terminal - access network - core network - EDN. Figure 1a illustrates the structure of terminal - access network - core network - EDN of the current standard.

[0046] FIG. 1b illustrates an example of an edge data center interconnection structure according to various embodiments of the present disclosure.

[0047] A data center is a physical facility that houses IT infrastructure (e.g., computer systems, hardware) to store, process, and manage vast amounts of digital data and information. Simply put, using a data center is similar to connecting your home desktop to the internet, accessing it remotely via a smartphone or laptop, and utilizing its data and computing resources. Therefore, until now, data centers have been no different from using a typical computer remotely.

[0048] Referring to FIG. 1b, FIG. 1b illustrates a conventional edge data center interconnection structure in which an access network and an edge data center are connected via a core network.

[0049] However, to reduce the burden of AI computation on terminals and improve AI service quality, the construction of edge data centers may become more active. Furthermore, to gain benefits such as delay reduction, an environment where edge data centers are directly connected to the access network (e.g., Figure 2) could be considered, rather than the existing structure where edge data centers are linked behind the core network (e.g., Figure 1b).

[0050] All access networks described in the present disclosure, including FIG. 1a and FIG. 1b, may be concepts including eNB of current LTE (4G) and gNB of NR (5G).

[0051] As mentioned above, the landscape of data centers is rapidly changing. AI services and applications are diversifying, growing rapidly, and their capabilities are rapidly increasing. Therefore, securing data centers equipped with the computational resources to support them is becoming increasingly important. Data centers that support these AI-related computational resources are called edge data centers.

[0052] Recently, AI services and applications have become more diverse and rapidly growing, and their level of sophistication is also rapidly increasing, making it increasingly important to secure data centers with the operational resources to support them.

[0053] Unlike traditional data centers, these edge data centers are specialized for AI computing, absorbing and supporting the demand for AI-related tasks, and the construction of edge data centers is becoming increasingly active.

[0054] In addition, edge data centers, which are positioned closer to terminals to improve the quality of AI services, are being considered. Edge data centers, which support AI computation at the edge, could be an alternative. Because edge data centers are relatively close to user equipment (UE) devices due to their system architecture, they can enhance AI service quality by performing or assisting with tasks such as AI-related computation and inference that terminals must perform when using AI services.

[0055] To implement the above-described effect, the linkage structure of Fig. 2 can be proposed.

[0056] FIG. 2 illustrates an example of a linkage structure of an access network and a data center according to one embodiment of the present disclosure.

[0057] Referring to FIG. 2, if the edge data center is built behind the access network rather than behind the core network (e.g., FIG. 2), computing delay can be reduced, thereby improving computational efficiency.

[0058] Additionally, the edge data center of FIG. 2 can be interconnected with multiple access networks as shown in FIG. 3.

[0059] FIG. 3 illustrates an example of a network structure in an edge data center construction environment according to one embodiment of the present disclosure.

[0060] Referring to Figure 3, it can be seen that an edge data center can be connected to multiple access networks. This connection reduces latency (e.g., computing delay) by not requiring the edge data center to be connected behind the core network (e.g., Figure 2).

[0061] FIG. 4 illustrates an example of signaling for AI operations in a conventional edge data center interconnection environment, according to various embodiments of the present disclosure.

[0062] With the rise of AI services, the importance of data centers is increasing. However, on-device AI, which performs AI computations independently, is also attracting attention as a next-generation AI. In particular, terminal manufacturers are showing a strong interest in on-device AI, preferring AI processing on the terminal itself rather than having AI services hosted in a data center.

[0063] As terminal performance advances, on-device AI services are being actively developed and released. However, there may be cases where it is difficult to handle all AI operations due to limitations in computational resources resulting from the physical size of the terminal.

[0064] In this case, the terminal can distribute the AI ​​computations it needs to perform to the data center or outsource the computations themselves, allowing the terminal to use its computational resources for other services requested by the user. However, if the data center is structurally distant from the terminal, this computational distribution and outsourcing to the data center requires multiple steps, potentially reducing efficiency.

[0065] This conventional edge data center environment is illustrated in Figure 4. Referring to Figure 4, the edge network center can be deployed behind the core network, similar to the structure of Figure 2. This structure requires multiple steps to distribute and delegate AI computations to the data center, which can result in delays and other issues, making it less efficient.

[0066] To solve these problems, a structure like Fig. 5 can be proposed.

[0067] FIG. 5 illustrates an example of signaling for AI operations in a case where an access network and an edge data center are directly linked, according to one embodiment of the present disclosure.

[0068] Referring to Figure 5, the access network and edge network to which a terminal connects can be directly connected without going through the core network. When the access network and edge network are directly connected, the structural distance between edge data centers is shortened, and the required signaling is simplified, enabling efficient distribution and delegation of AI computations.

[0069] The present disclosure proposes a device and method for requesting assistance for AI calculations or entrusting calculations to an edge data center linked to a connected access network when a terminal in the structure of FIG. 5 wishes to perform AI calculations on its own but, under certain conditions, a large load is generated on the AI ​​calculations.

[0070] FIG. 6 illustrates a comparison between a conventional scheduling request and a scheduling request for AI computation assistance and delegation, according to one embodiment of the present disclosure.

[0071] Referring to FIG. 6, box (601) illustrates a case of a conventional scheduling request (SR), and box (603) illustrates a scheduling request for AI computation assistance and delegation of the present disclosure.

[0072] In the case of a conventional scheduling request (601), a terminal can request resource allocation for uplink from an access network (e.g., a base station (eNB, gNB)) along with the presence of data to be transmitted via uplink.

[0073] In contrast, the scheduling request (603) for AI computation assistance and delegation may, in addition to the terminal requesting resource allocation for uplink with the presence of data to be transmitted to the access network (e.g., base station (eNB, gNB)), convey (or request) to the access network that it is willing to request AI computation assistance and delegation to the edge data center.

[0074] Specifically, in the case of a scheduling request (603) for AI computation assistance and delegation, when a terminal requests AI computation assistance and delegation to an edge data center, the terminal may request such a request through an access network.

[0075] In one embodiment, to request AI computation assistance and delegation, a terminal may use SR (Service Request) to notify the access network of its request for AI computation assistance and delegation. When transmitting SR, a separate identifier, AICR (AI computing request), may be newly defined and utilized to distinguish between general uplink and AI computation assistance and delegation requests.

[0076] Specifically, when a terminal transmits AICR together with SR transmission (e.g., transmits AICR = 1), the access network can identify that the terminal is requesting AI computation assistance and delegation.

[0077] Conversely, if the terminal does not send AICR when transmitting SR (e.g., transmits AICR = 0), the access network can identify that the terminal is requesting scheduling for the general uplink.

[0078] As a procedure following FIG. 6, FIG. 7 illustrates a procedure in which, when the access network receives an SR including AICR, the access network schedules an uplink signal for the terminal to transmit AI computation assistance and consignment-related data to the edge data center.

[0079] FIG. 7 illustrates a signal flow diagram for AI computation assistance and delegation with scheduling request and AICS indicators added to conventional DCI format, according to one embodiment of the present disclosure.

[0080] Referring to FIG. 7, a terminal can transmit an SR including AICS to an access network (701).

[0081] When the access network receives an SR including AICR, the access network can transmit information for uplink scheduling for the terminal to transmit AI computation assistance and consignment-related data to the edge data center (703).

[0082] According to one embodiment, information for uplink scheduling may be included in downlink control information (DCI). Specifically, operation (703) may newly define and utilize the delimiter AICS, a field included in the DCI format, to schedule AI computation assistance and delegation of the terminal using the existing DCI format.

[0083] The terminal may transmit an uplink signal for AI computation assistance and delegation to the access network (705) based on the scheduling information of operation (703). According to one embodiment, the scheduling information may be transmitted in a DCI for uplink scheduling. Specifically, if the terminal receives a DCI format for uplink scheduling and confirms that the DCI format contains an AICS (e.g., AICS = 1), the terminal recognizes that the uplink scheduling is for AI computation assistance and delegation, and may transmit the relevant uplink signal.

[0084] The access network can transmit data regarding AI computation assistance and delegation based on the uplink signal of the operation (705) (707).

[0085] FIG. 8 illustrates a signal flow diagram for AI computation assistance and delegation without separate indicators using a new AI computation assistance and delegation DCI format A according to one embodiment of the present disclosure.

[0086] Referring to FIG. 8, FIG. 8 defines a new DCI format (e.g., DCI format A) and illustrates a signaling procedure for an access network to perform uplink scheduling for AI computation assistance and delegation to a terminal.

[0087] The terminal can transmit an SR including AICS in the access network (801).

[0088] The terminal can receive uplink scheduling information for AI computation assistance and delegation from the access network (803).

[0089] In one embodiment, the uplink scheduling information may include DCI format A without a separate identifier (e.g., without AICS). When the terminal receives DCI format A, the terminal may identify the resource region scheduled by the DCI format A as an uplink resource region for AI computation assistance and delegation.

[0090] In conventional 4G and 5G networks, DCI was transmitted using a type of encryption so that only the intended terminal could identify the content. This encryption uses the radio network temporary identifier (RNTI). In this case, the base station can scramble the CRC (cyclic redundancy check) included in the DCI with one of several RNTIs previously assigned to the terminal and transmit it to the terminal. The terminal can then descramble the CRC of the received DCI using the assigned RNTI and interpret the DCI.

[0091] The scrambled bits can only be restored when descrambling with the same RNTI, and since other terminals do not have the RNTI, descrambling is impossible and thus the DCI cannot be interpreted.

[0092] To define the new DCI format A of Figure 8, a new RNTI must also be defined. In one embodiment, the RNTI used for scheduling AI computation assistance and delegation may be defined as an AI Computing-RNTI (AIC-RNTI). The AIC-RNTI may use one of the FFF3 to FFFA (hexadecimal), which are currently not used in 3GPP.

[0093] AIC-RNTI can also be used to generate scrambling sequences for DCI Format A, and for the physical uplink shared channel (PUSCH) and physical downlink shared channel (PDSCH) for AI computation assistance and delegation. This scrambling sequence generation may be similar to that in TS 38.211.

[0094] For example, the initial scrambling sequence for DCI in NR 3GPP TS 38.211 can be determined as shown in Equation 1 below.

[0095]

[0096] In mathematical expression 1 AIC-RNTI can be used.

[0097] Additionally, the initial scrambling sequence for PDSCH in NR 3GPP TS 38.211 can be determined as shown in mathematical expression 2 below.

[0098]

[0099] In mathematical expression 2 AIC-RNTI can be used.

[0100] Additionally, the initial scrambling sequence for PUSCH in NR 3GPP TS 38.211 can be determined as shown in mathematical expression 3 below.

[0101]

[0102] In mathematical expression 3 AIC-RNTI can be used.

[0103] The terminal can transmit uplink information for AI operation assistance and entrustment (805) in response to operation (803).

[0104] According to one embodiment, uplink information for AI computation assistance and delegation may include at least one of the AI ​​computation assistance and delegation order requested by the terminal, an AI model, information and data required for AI computation, or an AI computation range. The AI ​​computation assistance and delegation order requested by the terminal indicates the order of AI computation assistance and delegation sent by the terminal, and may be used to distinguish AI computation assistance and delegation sent from the same terminal on a case-by-case basis.

[0105] The access network can transmit information for AI computation assistance and delegation to the edge data center in response to the operation (805) (807).

[0106] According to one embodiment, information for AI operation assistance and entrustment may include at least one of the IMSI of the terminal requesting AI operation assistance and entrustment, or any terminal ID, information on the telecommunications company to which the access network belongs, or the eNB / gNB ID of the access network sending the information, along with the AI ​​operation assistance and entrustment order requested by the terminal, the AI ​​model, information and data required for the AI ​​operation, and the AI ​​operation range.

[0107] Any terminal ID described above can be used when the access network needs to pseudonymize terminal user information to the edge data center, and the terminal ID can be managed by being mapped to the IMSI or the UE ID assigned by the access network for general communication within the access network.

[0108] The edge data center can perform AI computation assistance and delegation (809) in response to the operation (807).

[0109] In operation (809), each AI computation task can be managed by the edge data center by labeling it based on the terminal IMSI or arbitrary terminal ID that requested AI computation assistance and delegation, the telecommunications company information to which the access network belongs, and the eNB / gNB ID of the access network sending the information.

[0110] In operation (809), after the AI ​​computation task is completed, the edge data center can reply to the corresponding access network with the telecommunications company information to which the access network belongs and the eNB / gNB ID of the access network sending the information.

[0111] The access network can transmit the AI ​​operation result to the terminal based on the terminal IMSI or arbitrary terminal ID that requested AI operation assistance and delegation (811).

[0112] In operation (811), if the access network receives an ACK from the terminal indicating that the AI ​​operation result has been successfully received, the access network can notify the edge data center that the AI ​​operation has been completed.

[0113] In operation (811), the edge data center may, in response to a termination notification received from the edge data center, delete the AI ​​operation-related information and data and reply to the access network with a confirmation message regarding the termination of the AI ​​operation.

[0114] If the access network is connected to multiple edge data centers and requests AI computation assistance and delegation of a terminal to edge data center A, but edge data center A rejects the request for some reason, the access network can request AI computation assistance and delegation to edge data center B by performing operations (801) to (811).

[0115] Figures 9 and 10 illustrate a case where a terminal performs a handover while an edge data center is assisting with AI computation. When a terminal requests AI computation assistance and delegation and performs a handover, the access network requires a signaling procedure to ensure that the AI ​​computation results are accurately returned to the terminal.

[0116] In intra-eNB / gNB handovers, the access network is aware of the AI ​​operation and therefore no separate action is required. However, in inter-eNB / gNB handovers, the AI ​​operation result must be transmitted to the access network to which the terminal has moved. Figures 9 and 10 illustrate inter-eNB / gNB handovers.

[0117] Hereinafter, in the present disclosure, the access network for the cell where the terminal previously resides is referred to as 'X' eNB / gNB, and the access network for the new cell to which the terminal has moved through handover is referred to as 'Y' eNB / gNB.

[0118] FIG. 9 illustrates a signal flow diagram for a case where a terminal hands over from access network X to access network Y, and access network Y is already linked with an edge data center linked with access network X, according to one embodiment of the present disclosure.

[0119] Referring to Figure 9, 'X' can receive a measurement report from a terminal before a handover occurs (901). Based on the measurement report to be received, a handover is commanded to 'Y', a cell with good radio wave strength.

[0120] 'X' can check whether it is connected to the edge data center that requested the AI ​​operation before ordering a handover to 'Y' (903).

[0121] According to one embodiment, operation (903) may include an operation in which, before 'X' commands a handover, 'X' requests a handover of the terminal to 'Y'. In addition, when requesting a handover, 'X' may also send a connection confirmation request message to 'Y' regarding whether the connection with the corresponding edge data center has been established. The connection confirmation request message may include information regarding the ID of the corresponding edge data center.

[0122] 'X' can transmit AI calculation results to the edge data center along with the action (903) and transmit a pause request message (905).

[0123] 'Y' may reply to 'X' with a response to the handover request along with a response to the linkage confirmation request message (907). In the embodiment of Fig. 9, the linkage confirmation request message may include information indicating that it has linked with the corresponding edge data center.

[0124] 'X' can command a handover to a terminal when a handover request is accepted from 'Y' to which the terminal is to be handed over (909).

[0125] In operation (907), if 'Y' responds to the connection confirmation request message that it has connected with the edge data center, 'X' may send an AI operation level change request message to the edge data center (911). In one embodiment, the AI ​​operation level change request message may request that the eNB / gNB ID of the access network be changed to the ID of 'Y'.

[0126] The edge data center can send an AI label change completion message to 'X' (913).

[0127] When 'X' receives an AI label change completion message from the edge data center, a pause cancellation request message and AI operation result for the pause request can be transmitted to the edge data center (915).

[0128] When the AI ​​operation of operation (913) is completed, the edge data center can transmit the AI ​​operation result to 'Y' according to the updated AI task level (917).

[0129] 'Y' can transmit the AI ​​operation result to the terminal that has handed over to its own area (919).

[0130] FIG. 10 illustrates a signal flow diagram for a case where a terminal hands over from access network X to access network Y, and access network Y is not linked with an edge data center linked with access network X, according to one embodiment of the present disclosure.

[0131] Referring to Figure 10, 'X' can receive a measurement report from a terminal before a handover occurs (1001). Based on the measurement report to be received, a handover is commanded to 'Y', a cell with good radio wave strength.

[0132] 'X' can check whether it is connected to the edge data center that requested the AI ​​operation before ordering a handover to 'Y' (1003).

[0133] According to one embodiment, operation (1003) may include an operation in which, before 'X' commands a handover, 'X' requests a handover of the terminal to 'Y'. In addition, when requesting a handover, 'X' may also send a connection confirmation request message to 'Y' regarding whether the connection with the corresponding edge data center has been established. The connection confirmation request message may include information regarding the ID of the corresponding edge data center.

[0134] 'X' can transmit AI calculation results to the edge data center along with the action (1003) and transmit a pause request message (1005).

[0135] 'Y' may reply to 'X' with a response to the handover request along with a response to the connection confirmation request message (1007). In the embodiment of Fig. 10, the connection confirmation request message may include information indicating that the connection with the corresponding edge data center is not established.

[0136] 'X' can command a handover to a terminal when a handover request from 'Y' to which the terminal is to be handed over is accepted (1009).

[0137] 'X' can request 'Y' to connect with an edge data center (1011). Action (1011) can be performed when the connection confirmation request message in action (1007) includes information that the edge data center is not connected.

[0138] 'Y' can request connection to the edge data center (1013) in response to action (1011).

[0139] The edge data center can send a link acceptance message to 'Y' (1015).

[0140] 'Y' can send a confirmation message to 'X' confirming that it has connected to the edge data center (1017).

[0141] 'X' may send an AI task level change request message to the edge data center (1019). In one embodiment, the AI ​​task level change request message may request that the eNB / gNB ID of the access network be changed to the ID of 'Y'.

[0142] The edge data center can send an AI label change completion message to 'X' (1021).

[0143] When 'X' receives an AI label change completion message from the edge data center, a pause cancellation request message and AI operation result for the pause request can be transmitted to the edge data center (1023).

[0144] When the AI ​​operation of operation (1021) is completed, the edge data center can transmit the AI ​​operation result to 'Y' according to the updated AI task level (1025).

[0145] 'Y' can transmit the AI ​​operation result to the terminal that has handed over to its own area (1027).

[0146] FIG. 11 illustrates a signal flow diagram for a case where, according to one embodiment of the present disclosure, a terminal hands over from access network X to access network Y, but access network Y is not linked with an edge data center linked with access network X and linkage fails.

[0147] Referring to Figure 11, 'X' can receive a measurement report from a terminal before a handover occurs (1101). Based on the measurement report to be received, a handover is commanded to 'Y', a cell with good radio wave strength.

[0148] 'X' can check whether it is connected to the edge data center that requested the AI ​​operation before ordering a handover to 'Y' (1103).

[0149] According to one embodiment, operation (1103) may include an operation in which, before 'X' commands a handover, 'X' requests a handover of the terminal to 'Y'. In addition, when requesting a handover, 'X' may also transmit a connection confirmation request message to 'Y' regarding whether the connection with the corresponding edge data center has been established. The connection confirmation request message may include information regarding the ID of the corresponding edge data center.

[0150] 'X' can transmit AI calculation results to the edge data center and a pause request message (1105) along with the action (1103).

[0151] 'Y' may reply to 'X' with a response to the handover request along with a response to the connection confirmation request message (1107). In the embodiment of Fig. 11, the connection confirmation request message may include information indicating that the connection with the corresponding edge data center is not established.

[0152] 'X' can command a handover to a terminal when a handover request from 'Y' to which the terminal is to be handed over is accepted (1109).

[0153] 'X' can request 'Y' to connect with an edge data center (1111). Action (1111) can be performed when the connection confirmation request message in action (1107) includes information that the edge data center is not connected.

[0154] 'Y' can request connection to the edge data center (1113) in response to action (1111).

[0155] If the edge data center fails (or rejects) the connection with 'Y', it may send a year failure (or rejection) message to 'Y' (1115). If 'Y' fails (or rejects) the connection with the edge data center, 'X' may not send an AI task level change message to the edge data center.

[0156] When 'X' receives an AI label change completion message from the edge data center, a pause cancellation request message and AI operation result for the pause request can be transmitted to the edge data center (1117).

[0157] The edge data center can transmit the AI ​​computation results to 'X' (1119).

[0158] 'X' can transmit the AI ​​operation result to 'Y' (1121) in response to the action (1119).

[0159] 'Y' can transmit the AI ​​operation result to the terminal that has handed over to its own area in response to the action (1121) (1123).

[0160] FIG. 12 illustrates a method of operating a terminal according to one embodiment of the present disclosure.

[0161] Referring to FIG. 12, a terminal may transmit an SR (Request for AI Computing Assistance and Delegation) to the access network to request AI computation assistance and delegation from an edge data center (1201). According to one embodiment, the SR may include an AICR (AI computing Request) to identify that it is requesting AI computation assistance and delegation.

[0162] The terminal can receive DCI (downlink control information) for uplink scheduling from the access network that received the SR including AICR (1203).

[0163] The terminal may transmit uplink information for AI computation assistance and delegation to the access network based on DCI (1205). According to one embodiment, the uplink information may include the sequence of AI computation assistance and delegation requested by the terminal, the AI ​​model, information and data required for AI computation, and the AI ​​computation scope.

[0164] In one embodiment, DCI may include AI computing scheduling (AICS).

[0165] The terminal can identify, based on AICS, that the DCI is for uplink scheduling for AI computation assistance and delegation.

[0166] In another embodiment, the DCI may include DCI format A. DCI format A may be a newly defined DCI format.

[0167] The terminal can identify that the DCI is related to uplink scheduling for AI computation assistance and delegation without a separate delimiter based on DCI format A.

[0168] According to one embodiment, the DCI may be transmitted to the terminal by scrambling a cyclic redundancy check (CRC) included in the DCI with one of a plurality of radio network temporary identifiers (RNTIs) previously allocated to the terminal.

[0169] According to one embodiment, the terminal may identify the DCI by descrambling the CRC of the DCI received using one of the plurality of RNTIs. That is, the terminal may identify the DCI by descrambling the CRC of the DCI received from the access network using one of the plurality of pre-allocated RNTIs.

[0170] According to one embodiment, the plurality of RNTIs include an AIC-RNTI (AI Computing - RNTI) for AI computation assistance and delegation, and the AIC-RNTI may include one of FFF3 to FFFA.

[0171] The terminal can receive AI computation results generated based on uplink information from the edge data center via an access network (1207).

[0172] According to one embodiment, the access network may transmit to the edge data center, along with uplink information for AI computation assistance and delegation, the international mobile subscriber identity (IMSI) of the terminal requesting AI computation assistance and delegation, or any terminal ID, information on the telecommunications company to which the access network belongs, and the base station ID of the access network sending the telecommunications company information.

[0173] In one embodiment, any terminal ID may be used when the access network needs to anonymize the terminal's user information to the edge data center, and any terminal ID may be managed by being mapped to the IMSI in the access network, or the terminal ID assigned by the access network for general communication.

[0174] FIG. 13 illustrates a configuration diagram of a terminal in a wireless communication system according to various embodiments of the present disclosure. The configuration illustrated in FIG. 13 can be understood as a configuration of a terminal. Terms such as "... unit," "... unit," etc., used hereinafter, refer to a unit that processes at least one function or operation, which can be implemented using hardware, software, or a combination of hardware and software.

[0175] Referring to FIG. 13, the terminal may include a communication unit (1310), a storage unit (1320), and a control unit (1330).

[0176] The communication unit (1310) may perform functions for transmitting and receiving signals via a wireless channel. For example, the communication unit (1310) may perform a conversion function between a baseband signal and a bit stream according to the physical layer specifications of the system. For example, when transmitting data, the communication unit (1310) may generate complex symbols by encoding and modulating a transmission bit stream. When receiving data, the communication unit (1310) may restore a reception bit stream by demodulating and decoding the baseband signal. In addition, the communication unit (1310) may upconvert a baseband signal to an RF band signal and transmit it through an antenna, and downconvert an RF band signal received through the antenna to a baseband signal. For example, the communication unit (1310) may include a transmission filter, a reception filter, an amplifier, a mixer, an oscillator, a DAC, an ADC, etc.

[0177] Additionally, the communication unit (1310) may include multiple transmit / receive paths. Furthermore, the communication unit (1310) may include at least one antenna array composed of multiple antenna elements. In terms of hardware, the communication unit (1310) may be composed of digital circuits and analog circuits (e.g., a radio frequency integrated circuit (RFIC)). Here, the digital circuits and analog circuits may be implemented in a single package. Additionally, the communication unit (1310) may include multiple RF chains. Furthermore, the communication unit (1310) may perform beamforming.

[0178] The communication unit (1310) transmits and receives signals as described above. Accordingly, all or part of the communication unit (1310) may be referred to as a "transmitter," a "receiver," or a "transceiver." Furthermore, in the following description, transmission and reception performed via a wireless channel may be used to mean that the communication unit (1310) performs the processing described above.

[0179] The storage unit (1320) can store data such as basic programs, application programs, and setting information for the operation of the terminal. The storage unit (1320) can be composed of volatile memory, non-volatile memory, or a combination of volatile and non-volatile memory. In addition, the storage unit (1320) can provide stored data upon request from the control unit (1330).

[0180] The control unit (1330) can control the overall operations of the terminal. For example, the control unit (1330) can transmit and receive signals through the communication unit (1310). In addition, the control unit (1330) can record and read data in the storage unit (1320). The control unit (1330) can perform the functions of the protocol stack required by the communication standard. To this end, the control unit (1330) may include at least one processor or microprocessor, or may be a part of a processor. In addition, a part of the communication unit (1310) and the control unit (1330) may be referred to as a CP (communication processor).

[0181] According to various embodiments, the control unit (1330) can control the terminal to perform operations according to the various embodiments described above.

[0182] FIG. 14 illustrates a configuration diagram of an access network in a wireless communication system according to various embodiments of the present disclosure. The configuration illustrated in FIG. 14 can be understood as a configuration of an access network. Terms such as "part," "unit," etc., used hereinafter, refer to a unit that processes at least one function or operation, which can be implemented using hardware, software, or a combination of hardware and software.

[0183] Referring to FIG. 14, the access network may include a wireless communication unit (1410), a backhaul communication unit (1420), a storage unit (1430), and a control unit (1440).

[0184] The wireless communication unit (1410) can transmit and receive wireless signals via a wireless channel. For example, the wireless communication unit (1410) can perform a conversion function between baseband signals and bit streams according to the physical layer specifications of the system. Furthermore, when transmitting data, the wireless communication unit (1410) can generate complex symbols by encoding and modulating the transmission bit stream. When receiving data, the wireless communication unit (1410) can restore the reception bit stream by demodulating and decoding the baseband signal.

[0185] The wireless communication unit (1410) can upconvert a baseband signal into an RF (radio frequency) band signal and transmit it through an antenna, and downconvert an RF band signal received through the antenna into a baseband signal. To this end, the wireless communication unit (1410) can include a transmission filter, a reception filter, an amplifier, a mixer, an oscillator, a digital to analog convertor (DAC), an analog to digital convertor (ADC), and the like.

[0186] The wireless communication unit (1410) may include a plurality of transmission and reception paths, and the wireless communication unit (1410) may include at least one antenna array composed of a plurality of antenna elements.

[0187] In terms of hardware, the wireless communication unit (1410) may include a digital unit and an analog unit, and the analog unit may include a plurality of sub-units depending on operating power, operating frequency, etc. The digital unit may be implemented with at least one processor (e.g., a digital signal processor (DSP)).

[0188] The wireless communication unit (1410) can transmit and receive wireless signals as described above. Accordingly, all or part of the wireless communication unit (210) may be referred to as a "transmitter," a "receiver," or a "transceiver or transceiver." Furthermore, in the following description, transmission and reception performed via a wireless channel may include processing performed by the wireless communication unit (1410) as described above.

[0189] The backhaul communication unit (1420) may provide an interface for communicating with other nodes within the network. That is, the backhaul communication unit (1420) may convert a bit string transmitted from an access network to other nodes, such as other access nodes, other access networks, upper nodes, and core networks, into a physical signal, and may convert a physical signal received from other nodes into a bit string.

[0190] The storage unit (1430) can store data such as basic programs, applications, and configuration information for the operation of the access network. The storage unit (1430) can be configured with volatile memory, non-volatile memory, or a combination of volatile and non-volatile memory. In addition, the storage unit (1430) can provide stored data upon request from the control unit (1440).

[0191] The control unit (1440) can control the overall operations of the access network. For example, the control unit (1440) can transmit and receive signals via the wireless communication unit (1410) or the backhaul communication unit (1420). Furthermore, the control unit (1440) can record and read data from the storage unit (1430). Furthermore, the control unit (1440) can perform the functions of the protocol stack required by the communication standard.

[0192] For this purpose, the control unit (1440) may include at least one processor.

[0193] According to various embodiments of the present disclosure, the control unit (1440) can control the access network to perform operations according to the various embodiments described above.

[0194] The methods according to the embodiments described in the claims or specification of the present disclosure may be implemented in the form of hardware, software, or a combination of hardware and software.

[0195] When implemented in software, a computer-readable storage medium storing one or more programs (software modules) may be provided. The one or more programs stored in the computer-readable storage medium are configured for execution by one or more processors within an electronic device. The one or more programs include instructions that cause the electronic device to execute methods according to the embodiments described in the claims or specification of the present disclosure.

[0196] These programs (software modules, software) may be stored in random access memory, non-volatile memory including flash memory, read only memory (ROM), electrically erasable programmable read only memory (EEPROM), magnetic disc storage devices, compact disc-ROM (CD-ROM), digital versatile discs (DVDs) or other forms of optical storage devices, magnetic cassettes, or may be stored in memories formed by a combination of some or all of these. In addition, each configuration memory may include multiple copies.

[0197] Additionally, the program may be stored on an attachable storage device that is accessible via a communication network, such as the Internet, an intranet, a local area network (LAN), a wide area network (WAN), a storage area network (SAN), or a combination thereof. Such a storage device may be connected to a device performing an embodiment of the present disclosure via an external port. Additionally, a separate storage device on the communication network may be connected to a device performing an embodiment of the present disclosure.

[0198] In the specific embodiments of the present disclosure described above, components included in the disclosure are expressed in the singular or plural form, depending on the specific embodiment presented. However, the singular or plural expressions are selected to suit the presented situation for convenience of explanation, and the present disclosure is not limited to singular or plural components. Components expressed in the plural form may be composed of singular elements, or components expressed in the singular form may be composed of plural elements.

[0199] While the detailed description of this disclosure has described specific embodiments, it should be understood that various modifications are possible without departing from the scope of this disclosure. Therefore, the scope of this disclosure should not be limited to the described embodiments, but should be defined not only by the scope of the claims described below, but also by equivalents thereof.

Claims

1. In the method of operating a terminal for requesting allocation of AI operations, A process of transmitting an AI operation allocation request message for a given AI operation to each of multiple distributed edge data centers and an access network connected to the terminal, A process of allocating uplink resources for transmitting necessary information for the AI ​​operation from the access network that has received the AI ​​operation allocation request message, and transmitting the necessary information to the access network; A process of obtaining the AI ​​computation result performed by one or more edge data centers among the plurality of edge data centers using the necessary information via the access network, How the terminal operates.

2. In claim 1, Each of the above multiple edge data centers, The number of connectable access networks is limited to a preset number. How the terminal operates.

3. In claim 1, For an edge data center among the above multiple edge data centers whose connection has been disconnected, at least one of the necessary information, information about the access network, and information about the terminal is taken measures to prevent it from being used after disconnection. How the terminal operates.

4. In claim 1, Between each of the above multiple edge data centers and the access network, At least one of the operation status, available resources, and number of connected terminals of each of the above multiple edge data centers is periodically updated. How the terminal operates.

5. In claim 4, One or more of the above edge data centers, A selection from among the plurality of edge data centers based on at least one of the updated information How the terminal operates.

6. In claim 5, One or more of the above edge data centers, The AI ​​operation is performed by receiving at least one of the AI ​​model required for the AI ​​operation, the information required for the AI ​​operation, and the scope of the AI ​​operation from the terminal. How the terminal operates.

7. In claim 1, One or more of the above edge data centers, Among the above multiple edge data centers, the edge data center that requested the AI ​​operation first but was rejected is selected from the remaining edge data centers. How the terminal operates.

8. In claim 1, When a handover occurs from a first cell included in the access network to a second cell included in an access network different from the access network in the terminal, the AI ​​operation result is provided from the access network to the other access network. How the terminal operates.

9. For a terminal requesting allocation of AI operations, Transmitter and receiver, A control unit operably connected to the above transmitter and receiver, The above control unit, through the above transceiver unit, Each of the multiple distributed edge data centers is linked to each other and an access network connected to the terminal transmits an AI operation allocation request message for a given AI operation, By receiving the AI ​​operation allocation request message from the access network, uplink resources for transmitting the necessary information for the AI ​​operation are allocated, and the necessary information is transmitted to the access network. One or more edge data centers among the plurality of edge data centers obtain the AI ​​computation results performed using the necessary information via the access network. Terminal.

Images