WO2025023657A1

WO2025023657A1 - Network node, user equipment and methods performed thereby in communication system

Info

Publication number: WO2025023657A1
Application number: PCT/KR2024/010491
Authority: WO
Inventors: Fuyuan LI; Weiwei Wang; Bin Wang; Hong Wang; Lixiang Xu; Wei Wang
Original assignee: Samsung Electronics Co Ltd
Current assignee: Samsung Electronics Co Ltd
Priority date: 2023-07-21
Filing date: 2024-07-19
Publication date: 2025-01-30
Anticipated expiration: 2026-01-21
Also published as: CN119342529A

Abstract

The present disclosure relates to a 5G communication system or a 6G communication system for supporting a higher data rate than a 4G communication system such as Long Term Evolution (LTE). The present disclosure provides a method performed by a first network node, the method comprising; receiving a first message transmitted by a first node, wherein the first message includes information for indicating to deactivate a quality of service (QoS) flow of user equipment (UE), transmitting the information for indicating to deactivate the QoS flow of the UE to a second node, and deactivating the QoS flow of the UE..

Description

NETWORK NODE, USER EQUIPMENT AND METHODS PERFORMED THEREBY IN COMMUNICATION SYSTEM

The present disclosure relates to the field of communication, and more specifically, to a method performed by a first network node, a method performed by a second network node, a method performed by user equipment, the first network node, the second network node, and the user equipment.

Considering the development of wireless communication from generation to generation, the technologies have been developed mainly for services targeting humans, such as voice calls, multimedia services, and data services. Following the commercialization of 5th-generation (5G) communication systems, it is expected that the number of connected devices will exponentially grow. Increasingly, these will be connected to communication networks. Examples of connected things may include vehicles, robots, drones, home appliances, displays, smart sensors connected to various infrastructures, construction machines, and factory equipment. Mobile devices are expected to evolve in various form-factors, such as augmented reality glasses, virtual reality headsets, and hologram devices. In order to provide various services by connecting hundreds of billions of devices and things in the 6th-generation (6G) era, there have been ongoing efforts to develop improved 6G communication systems. For these reasons, 6G communication systems are referred to as beyond-5G systems.

6G communication systems, which are expected to be commercialized around 2030, will have a peak data rate of tera (1,000 giga)-level bps and a radio latency less than 100μsec, and thus will be 50 times as fast as 5G communication systems and have the 1/10 radio latency thereof.

In order to accomplish such a high data rate and an ultra-low latency, it has been considered to implement 6G communication systems in a terahertz band (e.g., 95GHz to 3THz bands). It is expected that, due to severer path loss and atmospheric absorption in the terahertz bands than those in mmWave bands introduced in 5G, technologies capable of securing the signal transmission distance (that is, coverage) will become more crucial. It is necessary to develop, as major technologies for securing the coverage, radio frequency (RF) elements, antennas, novel waveforms having a better coverage than orthogonal frequency division multiplexing (OFDM), beamforming and massive multiple input multiple output (MIMO), full dimensional MIMO (FD-MIMO), array antennas, and multiantenna transmission technologies such as large-scale antennas. In addition, there has been ongoing discussion on new technologies for improving the coverage of terahertz-band signals, such as metamaterial-based lenses and antennas, orbital angular time pointum (OAM), and reconfigurable intelligent surface (RIS).

Moreover, in order to improve the spectral efficiency and the overall network performances, the following technologies have been developed for 6G communication systems: a full-duplex technology for enabling an uplink transmission and a downlink transmission to simultaneously use the same frequency resource at the same time; a network technology for using satellites, high-altitude platform stations (HAPS), and the like in an integrated manner; an improved network structure for supporting mobile base stations and the like and enabling network operation optimization and automation and the like; a dynamic spectrum sharing technology via collision avoidance based on a prediction of spectrum usage; an use of artificial intelligence (AI) in wireless communication for improvement of overall network operation by using AI from a designing stage for developing 6G and internalizing end-to-end AI support functions; and a next-generation distributed computing technology for overcoming the limit of user equipment (UE) computing ability through reachable super-high-performance communication and computing resources (such as mobile edge computing (MEC), clouds, and the like) over the network. In addition, through designing new protocols to be used in 6G communication systems, developing mechanisms for implementing a hardware-based security environment and safe use of data, and developing technologies for maintaining privacy, attempts to strengthen the connectivity between devices, optimize the network, promote softwarization of network entities, and increase the openness of wireless communications are continuing.

It is expected that research and development of 6G communication systems in hyper-connectivity, including person to machine (P2M) as well as machine to machine (M2M), will allow the next hyper-connected experience. Particularly, it is expected that services such as truly immersive extended reality (XR), high-fidelity mobile hologram, and digital replica could be provided through 6G communication systems. In addition, services such as remote surgery for security and reliability enhancement, industrial automation, and emergency response will be provided through the 6G communication system such that the technologies could be applied in various fields such as industry, medical care, automobiles, and home appliances.

According to an aspect of the present disclosure, there is provided a method performed by a first network node in a communication system, the method comprising: receiving a first message transmitted by a first node, wherein the first message includes information for indicating to deactivate a quality of service (QoS) flow of user equipment (UE), transmitting the information for indicating to deactivate the QoS flow of the UE to a second node, and deactivating the QoS flow of the UE.

According to the method performed by the first network node in the communication system provided by the present disclosure, wherein the method further comprises receiving a second message transmitted by the first node, wherein the second message includes information for indicating to activate the QoS flow of the UE, and activating the QoS flow of the UE, and transmitting the information for indicating to activate the QoS flow of the UE to the second node.

According to the method performed by the first network node in the communication system provided by the present disclosure, wherein the method further comprises: transmitting a third message to a second network node, wherein the third message includes information for indicating whether the deactivation for the QoS flow of the UE is successful.

According to the method performed by the first network node in the communication system provided by the present disclosure, wherein the method further comprises: transmitting a fourth message to the second network node, wherein the fourth message includes information for indicating whether the activation for the QoS flow of the UE is successful.

According to the method performed by the first network node in the communication system provided by the present disclosure, wherein the method further comprises: receiving a fifth message transmitted by a second network node, wherein the fifth message includes configuration information related to deactivating or activating the QoS flow.

According to the method performed by the first network node in the communication system provided by the present disclosure, wherein the configuration information is received when a PDU session or a QoS flow is initially set up or during one of a plurality of training iterations.

According to the method performed by the first network node in the communication system provided by the present disclosure, wherein the configuration information includes at least one of: protocol data unit (PDU) session identification (ID) or QoS flow identification (ID); a time point at which the QoS flow is activated for a first time, a time point at which the QoS flow is deactivated for a first time, a time point at which the QoS flow is activated for a second time, a time point at which the QoS flow is deactivated for a second time, during one training iteration; a periodicity of model training iterations; or a number of model training iterations.

According to the method performed by the first network node in the communication system provided by the present disclosure, wherein the method further comprises: transmitting a sixth message to the second network node, wherein the sixth message includes information for indicating whether the configuration for the deactivation or activation for the QoS flow of the UE is successful.

According to the method performed by the first network node in the communication system provided by the present disclosure, wherein the method further comprises: receiving a seventh message transmitted by the first node, wherein the seventh message includes information for indicating to deactivate the QoS flow of the UE in a case that the user equipment leaves federated learning.

According to the method performed by the first network node in the communication system provided by the present disclosure, wherein the seventh message includes at least one of: protocol data unit (PDU) session identification (ID) or QoS flow identification (ID); a PDU session or QoS flow status, which is set to be deactivated; a user equipment status of the federated learning, which is set to instruct the user equipment to leave federated learning; federated learning application identification, which is used to identify a federated learning training task.

According to the method performed by the first network node in the communication system provided by the present disclosure, wherein the method further comprises: transmitting an eighth message to the second network node, wherein the eighth message includes information for indicating whether the deactivation for the QoS flow of the UE is successful in a case that the user equipment leaves the federated learning.

According to the method performed by the first network node in the communication system provided by the present disclosure, wherein the method further comprises: receiving a ninth message transmitted by the first node, wherein the ninth message includes information for indicating to activate the QoS flow of the UE in a case that the user equipment re-joins the federated learning after leaving the federated learning.

According to the method performed by the first network node in the communication system provided by the present disclosure, wherein the ninth message includes at least one of: protocol data unit (PDU) session identification (ID) or QoS flow identification (ID); a PDU session or QoS flow status, set to be activated; a user equipment status of the federated learning, which is set to instruct the user equipment to join the federated learning; federated learning application identification, which is used to identify a federated learning training task.

According to the method performed by the first network node in the communication system provided by the present disclosure, wherein the method further comprises: transmitting a tenth message to the second network node, wherein the tenth message includes information for indicating whether the activation for the QoS flow is successful in a case that the user equipment re-joins the federated learning after leaving the federated learning.

According to the method performed by the first network node in the communication system provided by the present disclosure, wherein the user equipment leaving the federated learning comprises at least one of: the user equipment leaving the federated learning due to not meeting federated learning training requirements in an iteration after joining federated learning; the user equipment leaving the federated learning due to a failure in model distribution or training result reporting, resulting from an occurrence of network congestion or deterioration in wireless channel quality during a federated learning model distribution stage or a training result reporting stage, or in a case that learning model related data is not received within a predetermined time window during the model distribution stage, or in a case that no training result data is received by the first network node within a predetermined time window during the training result reporting stage; the user equipment leaving the federated learning due to a failure in training during federated learning training stage; or the user equipment leaving the federated learning in a case that the first network node monitors that the service requirements for the QoS flow cannot be fulfilled.

According to the method performed by the first network node in the communication system provided by the present disclosure, wherein the method further comprises: transmitting an eleventh message to the first node and the second node, wherein the eleventh message includes information for indicating to deactivate the QoS flow in a case that the first network node monitors that the service requirements for QoS flow cannot be fulfilled.

According to the method performed by the first network node in the communication system provided by the present disclosure, wherein the first message and the eleventh message include at least one of: protocol data unit (PDU) session identification (ID) or QoS flow identification (ID); or a PDU session or QoS flow status, which is set to be deactivated.

According to the method performed by the first network node in the communication system provided by the present disclosure, wherein the second message include at least one of: protocol data unit (PDU) session identification (ID) or QoS flow identification (ID); or a PDU session or QoS flow status, set to be activated.

According to the method performed by the first network node in the communication system provided by the present disclosure, wherein the third message and the eighth message include at least one of: protocol data unit (PDU) session identification (ID) or QoS flow identification (ID); success response information for indicating a success in the deactivationfor the corresponding PDU session or QoS flow; or rejection response information for indicating a failure in the deactivation for the corresponding PDU session or QoS flow and corresponding cause information.

According to the method performed by the first network node in the communication system provided by the present disclosure, wherein the fourth message and the tenth message include at least one of: protocol data unit (PDU) session identification (ID) or QoS flow identification (ID); success response information for indicating a success in the activation for the corresponding PDU session or QoS flow; or rejection response information for indicating a failure in the activation for the corresponding PDU session or QoS flow and corresponding cause information.

According to the method performed by the first network node in the communication system provided by the present disclosure, wherein the first network node is a base station, the first node includes one of the user equipment or a second network node, and the second node includes the other one of the user equipment or the second network node, and the second network node is a core network device; or the first network node is a centralized unit control plane (CU-CP) of the base station, the first node is a core network device, and the second node is a centralized unit user plane (CU-UP) or a distribution unit (DU) of the base station; or the first network node is a master node (MN), the first node is a core network device, and the second node is an secondary node (SN).

According to the method performed by the first network node in the communication system provided by the present disclosure, wherein the method further comprises: receiving a handover request acknowledge message transmitted by a third network node, and transmitting a radio resource control (RRC) reconfiguration message to the UE, wherein the handover request acknowledge message and the RRC reconfiguration message include at least one of: protocol data unit (PDU) session identification (ID) or QoS flow identification (ID); or a PDU session or QoS flow status, which is set to be deactivated.

According to the method performed by the first network node in the communication system provided by the present disclosure, the method comprises: wherein the method further comprises: receiving a handover command transmitted by the second network node, and transmitting a radio resource control (RRC) reconfiguration message to the UE, wherein the handover command and the RRC reconfiguration message include at least one of: protocol data unit (PDU) session identification (ID) or QoS flow identification (ID); or a PDU session or QoS flow status, which is set to be deactivated

According to the method performed by the first network node in the communication system provided by the present disclosure, wherein the method comprises: receiving an multicast and broadcast service (MBS) session resource setup request transmitted by the second network node, wherein the MBS session resource setup request includes at least one of: MBS start time and MBS end time; a sequence of scheduled activation times and non-activation times of MBS session; MBS user identification list, which is used to indicate UE identification to which MBS is to be applied.

According to the method performed by the first network node in the communication system provided by the present disclosure, wherein the method comprises: transmitting an MBS session resource setup response message to the second network node, wherein the MBS session resource setup response message includes at least one of: MBS session identification; a MBS QoS flow setup list, including information on the QoS flow successfully setup.

According to the method performed by the first network node in a the communication system provided by the present disclosure, wherein the method comprises: transmitting a message related to MBS session resources to the second network node, wherein the message related to MBS session resources includes at least one of: MBS session identification; cause information.

According to another aspect of the present disclosure, there is provided a method performed by a second network node in a communication system, the method comprising: transmitting a first message to a first network node and deactivating a quality of service (QoS) flow of user equipment (UE), wherein the first message includes information for indicating to deactivate the QoS flow; and/or transmitting a second message to the first network node and activating the QoS flow of the UE, wherein the second message includes information for indicating to activate the QoS flow.

According to the method performed by the second network node in the communication system provided by the present disclosure, wherein the method further comprises: transmitting the first message to the first network node, after receiving a twelfth message, including information for indicating to deactivate the QoS flow, transmitted by the user equipment.

According to the method performed by the second network node in the communication system provided by the present disclosure, wherein the method further comprises: transmitting the second message to the first network node, after receiving a thirteenth message, including information for indicating to activate the QoS flow, transmitted by the user equipment.

According to the method performed by the second network node in the communication system provided by the present disclosure, wherein the method further comprises: receiving a fourteenth message transmitted by the user equipment, wherein the fourteenth message includes information for indicating to deactivate the QoS flow in a case that the user equipment leaves the federated learning; transmitting, to the first network node, the information for indicating to deactivate the QoS flow in a case that the user equipment leaves the federated learning.

According to the method performed by the second network node in the communication system provided by the present disclosure, wherein the method further comprises: receiving a fifteenth message transmitted by the user equipment, wherein the fifteenth message includes information for indicating to activate the QoS flow in a case that the user equipment re-joins the federated learning after leaving the federated learning; and transmitting, to the first network node, the information for indicating to activate the QoS flow in a case that the user equipment re-joins the federated learning after leaving the federated learning.

According to the method performed by the second network node in the communication system provided by the present disclosure, wherein the method further comprises: receiving a sixteenth message transmitted by the user equipment, wherein the sixteenth message includes information for indicating to deactivate the QoS flow in a case that the user equipment fails in federated learning training; and transmitting, to the first network node, the information for indicating to deactivate the QoS flow in a case that the user equipment fails in federated learning training.

According to the method performed by the second network node in the communication system provided by the present disclosure, wherein the method further comprises: receiving a path switch request message transmitted by a third network node, wherein the path switch request message includes at least one of: protocol data unit (PDU) session identification (ID) or QoS flow identification (ID); or a PDU session or QoS flow status, which is set to be deactivated.

According to the method performed by the second network node in the communication system provided by the present disclosure, wherein the method further comprises: receiving a handover request acknowledge transmitted by the third network node, wherein the handover request acknowledge comprises at least one of: protocol data unit (PDU) session identification (ID) or QoS flow identification (ID); or a PDU session or QoS flow status, which is set to be deactivated.

According to yet another aspect of the present disclosure, there is provided a method performed by user equipment (UE) in a communication system, the method comprising: determining that conditions for deactivating and/or activating a quality of service (QoS) flow of the UE are satisfied, transmitting a first message to a first network node or a second network node, wherein the first message includes information for indicating to deactivate the QoS flow, and/or transmitting a second message to the first network node or the second network node, wherein the second message includes information for indicating to activate the QoS flow.

According to the method performed by the user equipment (UE) in the communication system provided by the present disclosure, wherein the method further comprises: receiving a seventeenth message transmitted by a server, wherein the seventeenth message includes information for instructing that the server decides the user equipment to leave the federated learning.

According to the method performed by the user equipment (UE) in the communication system provided by the present disclosure, wherein the method further comprises: receiving an eighteenth message transmitted by the server, wherein the eighteenth message includes information for instructing that the server decides the user equipment to re-join the federated learning after leaving the federated learning.

According to the method performed by the user equipment (UE) in the communication system provided by the present disclosure, wherein the method further comprises: transmitting a nineteenth message to the server, wherein the nineteenth message includes information for indicating a failure in the federated learning training.

According to another aspect of the present disclosure, there is provided a first network node comprising: a transceiver configured to transmit and receive signals with the outside; and a controller configured to control the transceiver to perform the method performed by the first network node as described above.

According to another aspect of the present disclosure, there is provided a second network node comprising: a transceiver configured to transmit and receive signals with the outside; and a controller configured to control the transceiver to perform the method performed by the second network node as described above.

According to another aspect of the present disclosure, there is provided user equipment (UE) comprising: a transceiver configured to transmit and receive signals with the outside; and a controller configured to control the transceiver to perform the method performed by the user equipment as described above.

According to another aspect of the present disclosure, there is provided a non-transitory computer-readable recording medium having stored thereon a program for performing any one of the above methods when executed by a computer.

For a more complete understanding of the present disclosure and its advantages, reference is now made to the following description taken in conjunction with the accompanying drawings, in which like reference numerals refer to like parts:

Fig. 1 shows an example wireless network according to an embodiment of the present disclosure;

Fig. 2 shows an example base station according to an embodiment of the present disclosure;

Fig. 3 shows an example user equipment according to an embodiment of the present disclosure;

Fig. 4 shows a procedure of example federated learning according to an embodiment of the present disclosure;

Fig. 5 shows an example procedure of GBR QoS Flow-based activation and deactivation mechanism for federated learning training UE-based according to an embodiment of the present disclosure;

Fig. 6 shows another example procedure of GBR QoS Flow-based activation and deactivation mechanism for federated learning training UE-based according to an embodiment of the present disclosure;

Fig. 7 shows an example procedure of a network-triggered QoS flow release or setup procedure according to an embodiment of the present disclosure;

Fig. 8 shows an example procedure of a network-triggered PDU session release or setup procedure according to an embodiment of the present disclosure;

Fig. 9a and Fig. 9b show an example procedure of GBR QoS processing for federated learning training UE in an abnormal scenario according to an embodiment of the present disclosure;

Fig. 10 shows another example procedure of GBR QoS processing for federated learning training UE in an abnormal scenario according to an embodiment of the present disclosure;

Fig. 11 shows another example procedure of GBR QoS processing for federated learning training UE in an abnormal scenario according to an embodiment of the present disclosure;

Fig. 12 shows another example procedure of GBR QoS processing when federated learning training UE handovers based on the Xn interface according to an embodiment of the present disclosure;

Fig. 13 shows another example procedure of GBR QoS processing when federated learning training UE handovers based on the Ng interface according to an embodiment of the present disclosure;

Fig. 14 shows an example procedure triggered by UE in which federated learning training UE joins a multicast MBS for the model distribution according to an embodiment of the present disclosure;

Fig. 15 shows an example procedure triggered by FL server in which federated learning training UE joins a multicast MBS for the model distribution according to an embodiment of the present disclosure.

Before undertaking the DETAILED DESCRIPTION below, it may be advantageous to set forth definitions of certain words and phrases used throughout this patent document. The term "couple" and its derivatives refer to any direct or indirect communication between two or more elements, whether those elements are in physical contact with one another. The terms "transmit", "receive", and "communicate", as well as derivatives thereof, encompass both direct and indirect communication. The terms "include" and "comprise", as well as derivatives thereof, mean inclusion without limitation. The term "or" is inclusive, meaning and/or. The phrase "associated with", as well as derivatives thereof, means to include, be included within, interconnect with, contain, be contained within, connect to or with, couple to or with, be communicable with, cooperate with, interleave, juxtapose, be proximate to, be bound to or with, have, have a property of, have a relationship to or with, or the like. The term "controller" means any device, system or part thereof that controls at least one operation. Such a controller may be implemented in hardware or a combination of hardware and software and/or firmware. The functionality associated with any particular controller may be centralized or distributed, whether locally or remotely. The phrase "at least one of", when used with a list of items, means that different combinations of one or more of the listed items may be used, and only one item in the list may be needed. For example, "at least one of: A, B, and C" includes any of the following combinations: A, B, C, A and B, A and C, B and C, and A and B and C. Likewise, the term "set" means one or more. Accordingly, a set of items can be a single item or a collection of two or more items.

Moreover, various functions described below can be implemented or supported by one or more computer programs, each of which is formed from computer readable program code and embodied in a computer readable medium. The terms "application" and "program" refer to one or more computer programs, software components, sets of instructions, procedures, functions, objects, classes, instances, related data, or a portion thereof adapted for implementation in a suitable computer readable program code. The phrase "computer readable program code" includes any type of computer code, including source code, object code, and executable code. The phrase "computer readable medium" includes any type of medium capable of being accessed by a computer, such as read only memory (ROM), random access memory (RAM), a hard disk drive, a compact disc (CD), a digital video disc (DVD), or any other type of memory. A "non-transitory" computer readable medium excludes wired, wireless, optical, or other communication links that transport transitory electrical or other signals. A non-transitory computer readable medium includes media where data can be permanently stored and media where data can be stored and later overwritten, such as a rewritable optical disc or an erasable memory device.

Definitions for other certain words and phrases are provided throughout this patent document. Those of ordinary skill in the art should understand that in many if not most instances, such definitions apply to prior as well as future uses of such defined words and phrases.

The figures included herein, and the various embodiments used to describe the principles of the present disclosure are by way of illustration only and should not be construed in any way to limit the scope of the disclosure. Further, those skilled in the art will understand that the principles of the present disclosure may be implemented in any suitably arranged wireless communication system.

Figs. 1-3 below describe various embodiments of the present disclosure implemented in wireless communications systems. The descriptions of Figs. 1-3 are not meant to imply physical or architectural limitations to the manner in which different embodiments may be implemented. Different embodiments of the present disclosure may be implemented in any suitably-arranged communications system.

FIG. 1 illustrates an example wireless network according to embodiments of the present disclosure. The embodiment of the wireless network shown in FIG. 1 is for illustration only. Other embodiments of the wireless network 100 could be used without departing from the scope of the present disclosure.

As shown in FIG. 1, the wireless network includes a base station (next generation nodeB, gNB or gNodeB) 101, a gNB 102, and a gNB 103. The gNB 101 communicates with the gNB 102 and the gNB 103. The gNB 101 also communicates with at least one network 130, such as the Internet, a proprietary Internet Protocol (IP) network, or other data network.

The gNB 102 provides wireless broadband access to the network 130 for a first plurality of user equipments (UEs) within a coverage area 120 of the gNB 102. The first plurality of UEs includes a UE 111, which may be located in a small business; a UE 112, which may be located in an enterprise (E); a UE 113, which may be located in a Wi-Fi hotspot (HS); a UE 114, which may be located in a first residence (R1); a UE 115, which may be located in a second residence (R2); and a UE 116, which may be a mobile device (M), such as a cell phone, a wireless laptop, a wireless personal digital assistant (PDA), or the like. The gNB 103 provides wireless broadband access to the network 130 for a second plurality of UEs within a coverage area 125 of the gNB 103. The second plurality of UEs includes the UE 115 and the UE 116, as well as subscriber stations (SS, for example, UEs) 117, 118 and 119. In some embodiments, one or more of the gNBs 101-103 may communicate with each other and with the UEs 111-116 using existing wireless communication techniques, and one or more of the UE 111-119 may communicate directly with each other (e.g., UEs 117-119) using other existing or proposed wireless communication techniques.

Depending on the network type, the term "base station" or "BS" can refer to any component (or collection of components) configured to provide wireless access to a network, such as transmit point (TP), transmit-receive point (TRP), an enhanced (or "evolved") base station (eNodeB or eNB), a 5G base station (gNB), a macrocell, a femtocell, a wireless fidelity (Wi-Fi) access point (AP), or other wirelessly enabled devices. Base stations may provide wireless access in accordance with one or more wireless communication protocols, e.g., 3GPP 5G New Radio (NR), Long Term Evolution (LTE), LTE Advanced (LTE-A), high speed packet access (HSPA), Wi-Fi 802.11a/b/g/n/ac, etc. For the sake of convenience, the various names for a base station-type apparatus and functionality are used interchangeably in this patent document to refer to network infrastructure components that provide wireless access to remote terminals. Also, depending on the network type, the term "user equipment" (UE) can refer to any component such as a mobile station (MS), subscriber station (SS), remote terminal, wireless terminal, receive point, or user device. For the sake of convenience, the various names for a user equipment-type device and functionality are used interchangeably in this patent document to refer to remote wireless equipment that wirelessly accesses a BS, whether the UE is a mobile device (such as a mobile telephone or smartphone) or is normally considered a stationary device (such as a desktop computer or vending machine).

Dotted lines show the approximate extents of the

coverage areas

120 and 125, which are shown as approximately circular for the purposes of illustration and explanation only. It should be clearly understood that the coverage areas associated with gNBs, such as the

coverage areas

120 and 125, may have other shapes, including irregular shapes, depending upon the configuration of the gNBs and variations in the radio environment associated with natural and man-made obstructions.

As described in more detail below, one or more of the UEs 111-119 include circuitry, programing, or a combination thereof. In certain embodiments, and one or more of the gNBs 101-103 includes circuitry, programing, or a combination thereof.

Although FIG. 1 illustrates one example of a wireless network, various changes may be made to FIG. 1. For example, the wireless network could include any number of gNBs and any number of UEs in any suitable arrangement. Also, the gNB 101 could communicate directly with any number of UEs and provide those UEs with wireless broadband access to the network 130. Similarly, each gNB 102-103 could communicate directly with the network 130 and provide UEs with direct wireless broadband access to the network 130. Further, the

gNBs

101, 102, and/or 103 could provide access to other or additional external networks, such as external telephone networks or other types of data networks.

FIG. 2 illustrates an example base station according to embodiments of the present disclosure. The embodiment of the gNB 102 illustrated in FIG. 2 is for illustration only, and the

gNBs

101 and 103 of FIG. 1 could have the same or similar configuration. However, gNBs come in a wide variety of configurations, and FIG. 2 does not limit the scope of the present disclosure to any particular implementation of a gNB.

As shown in FIG 2, the gNB 102 includes multiple antennas 200a-200n, multiple radio frequency (RF) transceivers 201a-201n, transmit (TX) processing circuitry 203, and receive (RX) processing circuitry 204. The gNB 102 also includes a controller/processor 205, a memory 206, and a backhaul or network interface (IF) 207.

The RF transceivers 201a-201n receive, from the antennas 200a-200n, incoming RF signals, such as signals transmitted by UEs in the network 100. The RF transceivers 201a-201n down-convert the incoming RF signals to generate intermediate frequency (IF) or baseband signals. The IF or baseband signals are sent to the RX processing circuitry 204, which generates processed baseband signals by filtering, decoding, and/or digitizing the baseband or IF signals. The RX processing circuitry 204 transmits the processed baseband signals to the controller/processor 205 for further processing.

The TX processing circuitry 203 receives analog or digital data (such as voice data, web data, electronic mail, or interactive video game data) from the controller/processor 205. The TX processing circuitry 203 encodes, multiplexes, and/or digitizes the outgoing baseband data to generate processed baseband or IF signals. The RF transceivers 201a-201n receive the outgoing processed baseband or IF signals from the TX processing circuitry 203 and up-converts the baseband or IF signals to RF signals that are transmitted via the antennas 201a-201n.

The controller/processor 205 may include one or more processors or other processing devices that control the overall operation of the gNB 102. For example, the controller/processor 205 could control the reception of forward channel signals and the transmission of reverse channel signals by the RF transceivers 201a-201n, the RX processing circuitry 204, and the TX processing circuitry 203 in accordance with well-known principles. The controller/processor 205 could support additional functions as well, such as more advanced wireless communication functions.

For instance, the controller/processor 205 could support beam forming or directional routing operations in which outgoing signals from multiple antennas 200a-200n are weighted differently to effectively steer the outgoing signals in a desired direction. Any of a wide variety of other functions could be supported in the gNB 102 by the controller/processor 205.

The controller/processor 205 is also capable of executing programs and other processes resident in the memory 206, such as an operating system (OS). The controller/processor 205 may move data into or out of the memory 206 as required by an executing procedure.

The controller/processor 205 is also coupled to the backhaul or network interface 207. The backhaul or network interface 207 allows the gNB 102 to communicate with other devices or systems over a backhaul connection or over a network. The interface 207 could support communications over any suitable wired or wireless connection(s). For example, when the gNB 102 is implemented as part of a cellular communication system (such as one supporting 5G, LTE, or LTE-A), the interface 207 could allow the gNB 102 to communicate with other gNBs over a wired or wireless backhaul connection. When the gNB 102 is implemented as an access point, the interface 207 could allow the gNB 102 to communicate over a wired or wireless local area network or over a wired or wireless connection to a larger network (such as the Internet). The interface 207 includes any suitable structure supporting communications over a wired or wireless connection, such as an Ethernet or RF transceiver.

The memory 206 is coupled to the controller/processor 205. Part of the memory 206 could include a random access memory (RAM), and another part of the memory 206 could include a Flash memory or other read only memory (ROM).

Although FIG. 2 illustrates one example of gNB 102, various changes may be made to FIG. 2. For example, the gNB 102 could include any number of each component shown in FIG. 2. As a particular example, an access point could include a number of interfaces 207, and the controller/processor 205 could support routing functions to route data between different network addresses. As another particular example, while shown as including a single instance of TX processing circuitry 203 and a single instance of RX processing circuitry 204, the gNB 102 could include multiple instances of each (such as one per RF transceiver). Also, various components in FIG. 2 could be combined, further subdivided, or omitted and additional components could be added according to particular needs.

FIG. 3 illustrates an example user equipment according to embodiments of the present disclosure. The embodiment of the UE 116 illustrated in FIG. 3 is for illustration only, and the UEs 111-115 and 117-119 of FIG. 1 could have the same or similar configuration. However, UEs come in a wide variety of configurations, and FIG. 3 does not limit the scope of the present disclosure to any particular implementation of a UE.

As shown in FIG. 3, the UE 116 includes an antenna 301, a radio frequency (RF) transceiver 302, TX processing circuitry 303, a microphone 304, and receive (RX) processing circuitry 305. The UE 116 also includes a speaker 306, a controller or processor 307, an input/output (I/O) interface (IF) 308, an input device 309, a touchscreen display 310, and a memory 311. The memory 311 includes an OS 312 and one or more applications 313.

The RF transceiver 302 receives, from the antenna 301, an incoming RF signal transmitted by a gNB of the network 100. The RF transceiver 302 down-converts the incoming RF signal to generate an IF or baseband signal. The IF or baseband signal is sent to the RX processing circuitry 305, which generates a processed baseband signal by filtering, decoding, and/or digitizing the baseband or IF signal. The RX processing circuitry 305 transmits the processed baseband signal to the speaker 306 (such as for voice data) or to the processor 307 for further processing (such as for web browsing data).

The TX processing circuitry 303 receives analog or digital voice data from the microphone 304 or other outgoing baseband data (such as web data, e-mail, or interactive video game data) from the processor 307. The TX processing circuitry 303 encodes, multiplexes, and/or digitizes the outgoing baseband data to generate a processed baseband or IF signal. The RF transceiver 302 receives the outgoing processed baseband or IF signal from the TX processing circuitry 303 and up-converts the baseband or IF signal to an RF signal that is transmitted via the antenna 301.

The processor 307 may include one or more processors or other processing devices and execute the OS 312 stored in the memory 311 in order to control the overall operation of the UE 116. For example, the processor 307 could control the reception of forward channel signals and the transmission of reverse channel signals by the RF transceiver 302, the RX processing circuitry 305, and the TX processing circuitry 303 in accordance with well-known principles. In some embodiments, the processor 307 includes at least one microprocessor or microcontroller.

The processor 307 is also capable of executing other processes and programs resident in the memory 311, such as processes for channel state information (CSI) reporting on uplink channel. The processor 307 may move data into or out of the memory 311 as required by an executing procedure. In some embodiments, the processor 307 is configured to execute the applications 313 based on the OS 312 or in response to signals received from gNBs or an operator. The processor 307 is also coupled to the I/O interface 308, which provides the UE 116 with the ability to connect to other devices, such as laptop computers and handheld computers. The I/O interface 308 is the communication path between these accessories and the processor 307.

The processor 307 is also coupled to the touchscreen display 310. The user of the UE 116 may use the touchscreen display 310 to enter data into the UE 116. The touchscreen display 310 may be a liquid crystal display, light emitting diode display, or other display capable of rendering text and/or at least limited graphics, such as from web sites.

The memory 311 is coupled to the processor 307. Part of the memory 311 could include RAM, and another part of the memory 311 could include a Flash memory or other ROM.

Although FIG. 3 illustrates one example of UE 116, various changes may be made to FIG. 3. For example, various components in FIG. 3 could be combined, further subdivided, or omitted and additional components could be added according to particular needs. As a particular example, the processor 307 could be divided into multiple processors, such as one or more central processing units (CPUs) and one or more graphics processing units (GPUs). Also, while FIG. 3 illustrates the UE 116 configured as a mobile telephone or smartphone, UEs could be configured to operate as other types of mobile or stationary devices.

In the communication mechanism, the network side reserves corresponding resources for GBR (Guaranteed Bit Rate) QoS (Quality of Service) flow set up by user equipment (UE). However, sometimes there are cases that the resources reserved for the QoS Flow are temporarily not used in a certain time period(s), resulting in a waste of resources. It is a possible solution to first release and then set up the QoS Flow, but this solution may lead to frequent release and setup of the QoS Flow, resulting in larger signaling overhead.

Various embodiments of the present disclosure provide a method performed by a first network node, by receiving a first message transmitted by a second node, wherein the first message includes information for indicating to deactivate a quality of service (QoS) Flow, and transmitting the first message to a third node, so that when the resources reserved for the QoS Flow are temporarily not used in a certain time period(s), the QoS Flow may be deactivated through the information included in the first message for indicating to deactivate the QoS Flow (e.g., the network side reserves the configuration of the GBR QoS Flow but does not reserve the corresponding resources which can be used for other transmissions, thereby avoiding a waste of resources). In this way, since the QoS Flow doesn't need to be first released and then set up frequently, the large signaling overhead resulting from the frequent release and setup of the QoS Flow is improved/solved.

In addition, with the acceleration of digitalization, a large amount of data will be generated. Through machine learning technology, the treasures hidden in data can be automatically mined. Machine learning models trained on a large amount of data have been applied in various scenarios and are profoundly changing various aspects of our world, such as precision medicine, clinical auxiliary diagnosis, new drug research and development, portrait recognition, voiceprint recognition, one thousand people one thousand faces recommending algorithm, multimodal learning regarding picture, voice, and natural language, and so on. In application, the accuracy, generalization ability, and the like of a model are very important, which rely on the learning of a large amount of data by machines.

Limited by constraints in data privacy security, such as laws and regulations, policy supervision, trade secrets, personal privacy, and the like, multiple data sources cannot directly exchange data, resulting in a "data island" phenomenon, which restricts the further improvement of artificial intelligence model capabilities. The birth of Federated Learning (FL) is to solve this problem.

The combination of federated learning mechanism and mobile communication system is an application trend for the future. Hereinafter, a case that the resources reserved for the QoS Flow are temporarily not used, occurring at a certain stage of the federated learning when the federated learning mechanism is combined with the 5G mobile communication system, will be described as an implementation of the present disclosure. However, it can be understood that the present disclosure is not limited to the federated learning mechanism, and any mechanism in which a case that the resources reserved for the QoS Flow are temporarily not used may occur at a certain period of time, as well as any mechanism that meets similar characteristics, are within the scope of the present disclosure. In addition, the present disclosure is not limited to the 5G communication system, and can also be extended to the 6G communication system or other communication systems that can implement the inventive concept of the present disclosure.

In FL mode, a cloud server averages models based on iterations, and aggregates together local models partially trained by each terminal device, so as to train a global model. In each training iteration, device uses local training data to carry out training according to models downloaded from a FL server. Then, the devices report mid-term training results to the cloud server through the UL channels of 5G or other wireless communication systems, and then the FL server aggregates training results from respective devices and updates the global model. Next, the updated global model is distributed to respective devices through the DL channels of 5G or other wireless communication systems, and then the devices may perform the training of the next iteration. In this regard, the distribution of the global model or the uploading of the devices' training results are closely related to the services of 5G or associated wireless communication systems.

The combination of federated learning mechanism and mobile communication system will be described in detail with reference to Fig. 4. Please refer to Fig. 4. Fig. 4 shows a basic procedure of FL, which, in one iteration, is basically divided into several stages as follows:

1. The selection stage for the federated learning training UE (401)

a) For a plurality of candidate federated learning training UEs, each UE reports a training resource status to the FL server before each round of training tasks, and then the FL server decides whether the UE participate in or leave this round of training tasks. As an implementation, when there is no change in UE's status, the UE does not need to notify the FL server of the training resource status per iteration. For example, upon a certain reporting of training resource status by the UE, the FL server makes a decision whether the UE join or leave the federation in this iteration. During subsequent iterations, if the UE's status does not change, then reporting may not be performed, and the FL server continues to perform the last selection result. When there is a change in the training resource status of the UE, the updated status is reported to the FL server.

2. The model distribution stage (402)

a) For a UE participating in this round of federated learning training, a 5GS (5G system) needs to set up communication resources in the core network, radio access network and UE side, for example to set up corresponding PDU session, QoS Flow and Data Radio Bearer (DRB) etc. In synchronous federated learning training tasks, in order to guarantee consistent communication performances at each UE, a contribution to basic consistency in model distribution time, and a well control over the completion of the reporting of training results within a certain period, the FL server may require the 5GS to set up GBR (Guaranteed Bit Rate) QoS Flow for the UE participating in federated learning training.

b) In the following, the possible disadvantages of setting up non-GBR QoS Flow for UE participating in federated learning training will be described. If the FL server sets up a non-GBR QoS Flow for a UE participating in federated learning training, since the cells or base stations where the respective UEs are located vary, or even if the respective UEs located in the same cell or base station, the configurations of radio resources for the UEs cannot be guaranteed as the same. In this way, for the distribution of federated models, it is possible that respective UEs complete model downloading at different times, that is, some UEs complete model downloading at an earlier time, whereas some UEs complete model downloading at a later time. In this way, when each UE completes training and reports training results, similar problems will also occur. The training results of respective UEs arrive at the FL server at different times, and there may be a great time span between the first arrival and the last arrival, which can affect the FL server's model aggregation process. In one embodiment, in order to avoid affecting model aggregation process of the FL server because the results of certain UEs are reported too late, the FL server may implement some strategies, such as deciding not to wait for the UEs that have not reported results when exceeding a certain time threshold from starting the model distribution to starting the model aggregation. In this way, however, there will be other problems. Since the training results of the UE in this iteration have not successfully participated in the model aggregation, a waste of resources at the UE side has been made. Therefore, setting up GBR QoS Flow for UE participating in federated learning training is beneficial to facilitate synchronizing the entire process for federation and the rapid convergence of model synchronization. The schemes involved in the present invention are all implemented on the premise that the FL server requires 5GS to set up GBR QoS Flow for federated learning training UE.

3. The training stage for federated learning training UE (402)

a) The UE participating in this iteration of federated learning training conducts corresponding training according to the configuration after completing downloading of the models and configuring of the federated learning training. For the UE in this procedure, if there is no other service, from the perspective of UE energy saving, the UE may notify the network by carrying certain indications (e.g., indication of only federated learning training task or expecting to enter inactive status), expecting the network to decide the UE to enter RRC_INACTIVE status as soon as possible.

4. The UE training result reporting stage (403)

a) For the UE participating in this round of training, after completing training tasks, it is necessary to report training results to the FL server according to the configuration of federated learning training. When the UE completes training in RRC_INACTIVE status, it enters the RRC_CONNECTED status through the RRC resume procedure. Then the UE reports the training results.

5. The Model aggregation stage

a) After the FL server receives the training results of respective UEs participating in federated learning training, or after the FL server simultaneously adopts some strategies and receives the training results of part of the UEs, it decides to aggregate the models.

After completing training task for one iteration, the FL server obtains a new aggregated global model, and the FL server will decide to start the task for the next iteration. This is repeated until the convergence of the model satisfies corresponding accuracy requirements, and thus the task for the federated learning training at this time is completed.

Through the analysis of Fig. 4, with respect to the basic procedure of federated learning training, i.e., the several stages involved in an iteration, the requirements and the usage of GBR QoS only involve the model distribution stage and the UE training result reporting stage. When the UE is in the training stage, it has completed the model downloading, waits, and doesn't report training result until the training result is output. At this stage, the resources reserved for GBR QoS required by federated learning training tasks will not be used, and depending on different models and training tasks, the time length of the UE training is uncertain, which may be either seconds or minutes. Then, if during one iteration, the resources for GBR QoS required by the federated learning training tasks are always, in the UE training procedure, reserved and not used, a great waste of resources will be made. Based on such problem, there is provided herein an optimized processing mechanism for GBR QoS in federated learning training tasks.

Please refer to Fig. 5. Fig. 5 shows an example procedure of GBR QoS Flow-based activation and deactivation mechanism for federated learning training UE-based according to an embodiment of the present disclosure.

As shown in Fig. 5, with respect to the core idea of the optimized processing mechanism for GBR QoS in federated learning training tasks, a UE-based and GBR QoS Flow-based activation and deactivation mechanism or a corresponding PDU Session activation and deactivation mechanism is introduced.

- When the GBR QoS Flow or the corresponding PDU Session is in deactivated status, the 5GC (5G Core Network), RAN (Radio Access Network) and UE side save the configuration of the existing GBR QoS Flow, but the RAN side and 5GC side no longer reserves the corresponding resources, and their previously reserved resources may be allocated to other UEs, so as to avoid the waste of resources.

- When the GBR QoS Flow or the corresponding PDU Session is in activated status, the 5GC, RAN and UE side activate the configuration of the existing GBR QoS Flow, and the RAN side and 5GC side need to reserve corresponding resources to satisfy the GBR QoS requirements.

At present, the specification regarding setting up GBR QoS Flow can be implemented either by means of setting up a new PDU Session, or by means of setting up a new QoS Flow over an existing PDU Session. In either way, the introduced UE-based and GBR QoS Flow-based activation and deactivation mechanism or UE-based activation and deactivation mechanism based on PDU Session corresponding to GBR QoS Flow are the ideas embodied in the present invention, and are within the protection scope of the present invention.

When the FL server decides that a certain UE joins the training task for this iteration, it notifies 5GS to set up the corresponding required GBR QoS Flow and may indicate that the initial status of GBR QoS Flow is configured as "activated". Optionally, it is also possible to configure the initial status of GBR QoS Flow as "deactivated", and then notify the 5GS to activate this GBR QoS Flow when the FL server decides to start the model distribution task. Optionally, following the existing mechanism in specification, when the GBR QoS Flow is set up, it is in activated status by default, without carrying any indication. The RAN needs an indication for activating to restore the GBR QoS Flow to the activated status only after it receives a deactivation indication of GBR QoS Flow. When the GBR QoS Flow is in an activated status, the FL server performs the model distribution. For a certain UE, when the model distribution is completed, the FL server may notify the 5GC, after being aware of the completion, to configure the corresponding GBR QoS Flow as a deactivated status, so that the 5GC, RAN and UE can save the existing GBR QoS Flow configuration, but the RAN side and 5GC side no longer reserve corresponding resources.

With respect to the GBR QoS Flow deactivation indication (502), optionally, it may be triggered by an AF/FL server to notify the 5GC of it, and then the 5GC notifies the RAN of it through an NGAP message, which in turn notifies the UE of it through an RRC message. The indication message may include one or more of the following information:

- PDU Session ID or QoS Flow ID (QFI);

- PDU Session or QoS Flow Status = "deactivated".

With respect to GBR QoS Flow activation indication (501), optionally, it may be triggered by the AF/FL server to notify the 5GC of it, and then the 5GC notifies the RAN of it through an NGAP message, which in turn notifies the UE of it through an RRC message. The indication message may include one or more of the following information:

- PDU Session ID or QoS Flow ID (QFI);

- PDU Session or QoS Flow Status = "activated".

With respect to information of GBR QoS Flow status indication, the NG interface may be used to indicate it to the RAN from the 5GC by means of a new Class1 or Class2 procedure, or the existing NG interface message may be used to indicate it when the initial QoS Flow is set up (e.g., NGAP (NG Application Protocol): INITIAL CONTEXT SETUP REQUEST or PDU SESSION RESOURCE SETUP REQUEST), or it may be indicated through modification signaling of QoS Flow (e.g., NGAP: UE CONTEXT MODIFICATION REQUEST or PDU SESSION RESOURCE MODIFY REQUEST). If the Class1 procedure is used, for the GBR QoS Flow deactivation indication, the corresponding response message may include one or more of the following information:

- PDU Session ID or QoS Flow ID (QFI);

- Success response or Acknowledgement information of "PDU Session deactivated" or "QoS Flow deactivated";

- Rejection information of "PDU Session deactivated" or "QoS Flow deactivated"， which possibly indicates Cause information.

For GBR QoS Flow activation indication, the corresponding response message may include one or more of the following information:

- PDU Session ID or QoS Flow ID (QFI);

- Success response or Acknowledgement information of "PDU Session activated" or "QoS Flow activated";

- Rejection information of "PDU Session activated" or "QoS Flow activated", which possibly indicates Cause information.

Optionally, when the UE completes model downloading, the UE notifies the 5GC of the completion through NAS layer signaling after being aware of this event, and then the 5GC notifies the RAN side to perform the deactivating operation on the corresponding GBR QoS Flow through the above scheme. The enhancement of NAS signaling may optionally be implemented in the following ways, such as adding an IE (information element) to an existing message (such as NAS PDU session modification request), which may include one or more of the following information:

- PDU Session ID or QoS Flow ID (QFI);

- PDU Session or QoS Flow Status = "deactivated".

Optionally, the UE notifies the RAN through an RRC message, and then the RAN notifies the 5GC. The corresponding RRC message or NGAP message may include one or more of the following information:

- PDU Session ID or QoS Flow ID (QFI).

- PDU Session or QoS Flow Status ="deactivated"

The UE may notify the RAN through a new RRC message or by using an existing RRC message, such as UE Assistance Information. The RAN may also notify the 5GC through a new NGAP message or by using an existing NGAP message (e.g., PDU Session Resource Modify Indication or PDU Session Resource Notify procedure).

For the UE training result uploading stage, when the UE completes training, the UE may notify the 5GC through NAS signaling, and then the 5GC notifies the RAN side to perform activating operation on the corresponding GBR QoS Flow through the above scheme, so as to upload the training results to the FL server. When the UE completes training result uploading, the UE may notify the 5GC through NAS layer signaling, and then the 5GC notifies the RAN side to perform deactivating operation on the corresponding GBR QoS Flow through the above scheme.

Optionally, when the UE completes training, and is to upload the training result, the UE may notify the RAN through an RRC message, and then the RAN notifies the 5GC to perform activating operation on the corresponding GBR QoS Flow. The corresponding RRC message or NGAP message may include one or more of the following information:

- PDU Session ID or QoS Flow ID (QFI)

- PDU Session or QoS Flow Status = "activated".

Optionally, when the UE completes training result uploading, the UE may notify the RAN through an RRC message, and then the RAN notifies the 5GC to perform deactivating operation on the corresponding GBR QoS Flow. The corresponding RRC message or NGAP message may include one or more of the following information:

- PDU Session ID or QoS Flow ID (QFI)

- PDU Session or QoS Flow Status = "deactivated".

Optionally, the 5GC may notify the AF/FL server whether the status of the GBR QoS Flow or the corresponding PDU Session of a certain federated learning training UE is in activated or deactivated status.

Through the above method, the problem of a great waste of resources caused by the fact that the resources for GBR QoS are always reserved and not used during the training procedure for the federated learning training UE is solved/improved, and through different messaging methods, the embodiments of the present disclosure can adapt to more flexible scenarios and various different network architectures.

In addition, through the analysis of Fig. 4, with respect to the basic procedure of federated learning training, there are several different stages involved in an iteration. However, it is required to take N different iterations to make the model convergence satisfy the accuracy requirements and to complete model training, wherein the procedures involved in the different iterations are basically the same. In this way, each iteration may be set to be periodic, wherein the time points of the activation and deactivation for GBR QoS may be pre-configured uniformly without having to be configured in each iteration. In this way, the signaling overhead can be reduced.

Please refer to Fig. 6. Fig. 6 shows another example procedure of GBR QoS Flow -based activation and deactivation mechanism for federated learning training UE-based according to an embodiment of the present disclosure. As shown in Fig. 6, by means of pre-configuration, the configuration parameters setting of activation and deactivation may be performed when the PDU Session or GBR QoS Flow is initially set up, or the setting may be changed when adjustment is needed later, or the configuration parameters setting may be performed during one of a plurality of training iterations, or the setting may be changed when adjustment is needed later. As shown in Fig. 6, it is assumed that in one iteration, the model distribution starts at time point T0, the distribution completion is predicted to be at time point T1, the start of the training result reporting is predicted to be at time point T2, and the reporting completion is predicted to be at time point T3. It is assumed that the iteration periodicity is T, and it goes through N iterations. Then the operations of activation and deactivation for GBR QoS Flow may be performed at the following time points in the subsequent iteration procedure.

For the start of model distribution, the activation (601) for GBR QoS Flow may be set at time point T0, and at time points of T0 plus an integer multiple of periodicity T, i.e., at time points T0, T0+T, T0+2T, ..., T0+ (N-1)*T.

For the end of model distribution, the deactivation (602) for GBR QoS Flow may be set at time point T1, and at time points of T1 plus an integer multiple of periodicity T, i.e., at time points T1, T1+T, T1+2T, ..., T1+ (N-1)*T.

For the start of training result reporting, the activation (603) for GBR QoS Flow may be set at time point T2, and at time points of T2 plus an integer multiple of periodicity T, i.e., at T2, T2+T, T2+2T, ..., T2+(N-1)*T.

For the end of training result reporting, the deactivation (604) for GBR QoS Flow may be set at T3, and at time points of T3 plus an integer multiple of periodicity T, i.e., at T3, T3+T, T3+2T, ..., T3+(N-1)*T.

Therefore, the pre-setting of the activation and deactivation for GBR QoS Flow may be configured in an NG interface message or RRC message through the following information:

- PDU Session ID or QoS Flow ID (QFI);

- a first time point T0 which the QoS flow is activated for the first time, a second time point T1 at which the QoS flow is deactivated for the first time, a third time point T2 at which the QoS flow is activated for the second time, a fourth time point T4 at which the QoS flow is deactivated for the second time, during one training iteration (T0 and T1 can be set the start time and end time of the model distribution stage respectively, and T2 and T3 can be set the start time and end time of the training result reporting stage respectively);

- T (length of time, representing a periodicity of iteration);

- N (positive integer, representing a number of iterations),

wherein, the above pre-configured information may be received when the PDU Session or QoS Flow (e.g., GBR QoS Flow) is initially set up or during one of a plurality of training iterations.

With respect to preset information of the activation and deactivation for GBR QoS Flow, the NG interface may be used to indicate it to the RAN from the 5GC by means of a new Class1 or Class2 procedure, or the existing NG interface message may be used to indicate it when the initial QoS Flow is set up (e.g., NGAP: INITIAL CONTEXT SETUP REQUEST or PDU SESSION RESOURCE SETUP REQUEST), or it may be indicated through modification signaling of QoS Flow (e.g., NGAP: UE CONTEXT MODIFICATION REQUEST or PDU SESSION RESOURCE MODIFY REQUEST). If the Class1 procedure is used, for the preset of the activation and deactivation for GBR QoS Flow, the corresponding response message(indicated to the 5GC by the RAN) may include one or more of the following information:

- Success response or Acknowledgement information;

- Rejection information, which possibly indicates Cause information.

By means of pre-configuring the activation and deactivation mechanism for GBR QoS Flow uniformly, it is unnecessary to configure the time points of the activation and deactivation for GBR QoS in each iteration, thereby further reducing the signaling overhead. In addition, it can be understood that, in the above-mentioned method, the pre-configuration may be carried out not just at the beginning of iteration, or instead, the pre-configuring uniformly may be carried out after a plurality of iterations (e.g., after determining appropriate parameters), which is not limited by the present disclosure.

It has been set forth above that when the FL server decides to set up GBR QoS Flow for UE participating in federated learning training, it is beneficial to guarantee that the time progresses of the model distribution and the training result uploading in the federated learning training procedure are generally consistent in each iteration training, thereby achieving a better model aggregation effect and shortening the convergence time of the entire federated model. In order to obtain independent and equally distributed samples from all devices, that is, to give all devices a fair opportunity to contribute to the aggregation model, a large number of UE that satisfy the training requirements and conditions may be needed to participate in federated learning training. The environment in which the UE is located in wireless communication is time-varying and complex, and the UE participating in federated learning training may have the following situations.

1. When the UE has successfully joined the federated learning training tasks, in the next iteration, some situations (such as change in location of UE or deterioration in quality of wireless channel) may make the UE in the selection stage of this iteration be decided by the FL server not to satisfy the training requirements, and then make the UE leaves this iteration. Then, the situation may be restored, or the conditions satisfy the requirements for federated learning training again, so the UE re-joins federated learning training.

2. When the UE is in the model distribution stage or training result reporting stage, situations such as network congestion or deterioration in wireless channel quality may occur. Either within the duration between the start time and the end time of the model distribution stage, if the model data is not received within a certain time window (UE-based implementation), or within the duration between the start time and the end time of the training result reporting stage, if the uplink training result data is not received within a certain time window (RAN-based implementation), it is possible to lead to a failure in model distribution or training result reporting.

3. When the UE is in the training stage, a failure in training may occur due to the UE's own situations, such as increased computing load or other abnormal situations in computing.

In the following, description will be made with regard to the three abnormal situations as mentioned above and the corresponding processing methods.

For the first situation, the final result is that the UE leaves the training task for this iteration, and failed to successfully participate in the federated learning training task for this iteration. However, for the GBR QoS Flow set up by the UE, the network side is still in a status of reserving resources, whereas this iteration will not make use of the resources. Then how to deal with this situation so as to avoid wasting wireless resources and core network resources.

It is a possible way to release the GBR QoS Flow through signaling, and then re-set up the corresponding GBR QoS Flow when the UE re-joins the federation. The corresponding signaling for releasing and setting up GBR QoS Flow will involve NG interface signaling, NAS signaling and RRC signaling. As shown in Fig. 7, Fig. 7 shows an example procedure of a network-triggered QoS Flow release or setup procedure according to an embodiment of the present disclosure.

In Fig. 7, the releasing 701 and setting up 702 of the QoS Flow can be implemented through a PDU session resource modify procedure without releasing the PDU session. As shown in Fig. 8, Fig. 8 shows an example procedure of a network-triggered PDU Session release or setup procedure according to an embodiment of the present disclosure. In Fig. 8, the releasing and setting up of the QoS Flow can be implemented by releasing 801 and then re-setting up 802 the PDU session resources for the entire FL task. No matter which signaling mechanism is used, when a situation that one or more UE(s) with federated learning training tasks frequently leave and re-join the federation in different iterations, a large signaling overhead will be generated. In order to avoid the problems caused by the above situation, the present application proposes an optimized processing mechanism for GBR QoS of federated learning training UE.

Please refer to Figs. 9a and 9b. Figs. 9a and 9b show an example procedure of GBR QoS processing for federated learning training UE in an abnormal scenario according to an embodiment of the present disclosure, i.e., a flowchart of the first optimized processing mechanism for GBR QoS of federated learning training UE. As shown in Figs. 9a and 9b, for the deactivating operation of the FL GBR QoS:

Alt.1 (Alternative 1) 901: When the AF/FL server decides that the UE leaves the federated learning training task for this iteration, it notifies the 5GC, and then the network element in the 5GC notifies the RAN side of the deactivation for the GBR QoS Flow corresponding to the FL training task of this UE or of the deactivation for the corresponding PDU Session. Deactivation means that in a case that the RAN side keeps the configuration of the existing GBR QoS Flow, it will no longer reserve the corresponding resources, so as to avoid wasting resources. Notification can be made in many ways, such as defining a new message of class1 or class2 to notify the RAN, or adding an IE to the existing NGAP message (such as PDU Session Resource Modify Request) to notify the RAN. Such message may include one or more of the following information:

- PDU Session ID or QoS Flow ID (QFI);

- PDU Session or QoS Flow Status = "deactivated".

For a class1 procedure, the corresponding response information may include one or more of the following information:

- PDU Session ID or QoS Flow ID (QFI);

- Rejection response of "PDU Session deactivated" or "QoS Flow deactivated", which possibly indicates Cause information.

In order to avoid the influence on the existing messages or IEs (such as GBR QoS information IE) at the RAN side in the specification, the CN (core network) may notify the RAN by means of defining a new message, such as Downlink FL information Transfer, or may notify the RAN of the event that the UE leaves the federation for a certain federation task, and then the RAN side keeps the configuration of the corresponding GBR QoS Flow of this UE, but releases the reserved resources. The newly defined message may include one or more of the following information:

- PDU Session ID or QoS Flow ID (QFI);

- FL UE Status = "leave the federation";

- FL Application ID (used to identify a FL training task, because one UE may join training tasks for multiple different FL Applications at the same time, and the task for each application sets up a different PDU Session or a different GBR QoS Flow).

In a word, as long as the function of the information notified to the RAN side is to make the RAN side keep the configuration of the existing specified GBR QoS Flow and the RAN side will no longer reserve the corresponding resources, it belongs to the protection scope of the present invention.

Optionally, the RAN side may then notify the UE through an RRC message as needed, and the corresponding PDU session or QoS Flow or DRB enters the deactivated status.

Alt.2 (Alternative 2) 902: When the AF/FL server decides that the UE leaves the federated learning training task for this iteration, it notifies the UE through application layer information, then the UE notifies the 5GC through NAS layer signaling, and then the 5GC notifies the RAN side through the scheme of Alt.1. The enhancement of NAS signaling can optionally be implemented in the following ways, such as adding an IE to the existing message (such as NAS PDU session modification request), which may include one or more of the following information:

- PDU Session ID or QoS Flow ID (QFI);

- PDU Session or QoS Flow Status = "deactivated";

- FL UE Status = "leave the federation";

Alt.3 (Alternative 3) 903: When the AF/FL server decides that the UE leaves the federated learning training task for this iteration, it notifies the UE through application layer information, and then the UE notifies the RAN through an RRC message, and then the RAN notifies the CN. The corresponding RRC message or NGAP message may include one or more of the following information:

- PDU Session ID or QoS Flow ID (QFI);

- PDU Session or QoS Flow Status = "deactivated";

- FL UE Status = "leave the federation";

The UE may notify the RAN through a new RRC message or by using an existing RRC message, such as UE Assistance Information. The RAN may also notify the CN through a new NGAP message or by using an existing NGAP message (such as PDU Session Resource Modify Indication or PDU Session Resource Notify procedure).

When the federated learning training UE is determined by the FL server not to satisfy the training requirements, and the UE is made to change from previously joining to now leaving this iteration, after the GBR QoS Flow or the corresponding PDU Session has been in a deactivated status through the above-mentioned Alternative 1, Alternative 2 or Alternative 3, the situation of the status of UE participating in federated learning training may be restored during a certain subsequent iteration, that is, its conditions satisfy the requirements for federated learning training again, so the UE needs to restore the corresponding PDU Session or GBR QoS Flow to an activated status for re-joining federated learning training. As shown in Fig. 6, for the activating operation of the FL GBR QoS:

Alt.11 (Alternative 11) 904: When the AF/FL server decides the UE re-join federated learning training task for this iteration, it notifies the 5GC, and then the network element in the 5GC notifies the RAN side of the activation for the GBR QoS Flow corresponding to the FL training task of this UE, or the activation for the corresponding PDU Session. Activation means that in a case that the RAN side restore the configuration of the existing GBR QoS Flow, it needs to reserve the corresponding resources to satisfy the corresponding requirements for GBR QoS. Notification can be made in many ways, such as defining a new message of class1 or class2 to notify the RAN, or adding an IE to the existing NGAP message (such as PDU Session Resource Modify Request) to notify the RAN. Such message may include one or more of the following information:

- PDU Session ID or QoS Flow ID (QFI);

- PDU Session or QoS Flow Status = "activated".

- PDU Session ID or QoS Flow ID (QFI);

- Rejection response of "PDU Session activated" or "QoS Flow activated", which possibly indicates Cause information.

In order to avoid the influence on the existing messages or IEs (such as GBR QoS information IE) at the RAN side in the specification, the CN may notify the RAN by means of defining a new message, such as Downlink FL information Transfer, or may notify the RAN of the event that the UE re-joins the federation for a certain federation task, and then the RAN side restores the configuration of the corresponding GBR QoS Flow of this UE, and reserves the corresponding resources. Such message may include one or more of the following information:

- PDU Session ID or QoS Flow ID (QFI);

- FL UE Status = "join the federation";

In a word, as long as the function of the information notified to the RAN side is to make the RAN side reserve the corresponding resources in a case that the configuration of the existing specified GBR QoS Flow is restored, it belongs to the protection scope of the present invention.

Optionally, the RAN side may then notify the UE through an RRC message as needed, and the corresponding PDU session or QoS Flow or DRB will be restored to enter an activated status.

Alt.12 (Alternative 12) 905: When the AF/FL server decides that the UE re-joins the federated learning training task for this iteration, it notifies the UE through application layer information, then the UE notifies the 5GC through NAS layer signaling, and then the 5GC notifies the RAN side through the scheme of Alt.11. The enhancement of NAS signaling can optionally be implemented in the following ways, such as adding an IE to the existing message (such as NAS PDU session modification request), which may include one or more of the following information:

- PDU Session ID or QoS Flow ID (QFI);

- PDU Session or QoS Flow Status = "activated";

- FL UE Status = "join the federation";

Alt.13 (Alternative 13) 906: When the AF/FL server decides that the UE re-joins the federated learning training task for this iteration, it notifies the UE through application layer information, then the UE notifies the RAN through an RRC message, and then the RAN notifies the CN. The corresponding RRC message or NGAP message may include one or more of the following information:

- PDU Session ID or QoS Flow ID (QFI);

- PDU Session or QoS Flow Status = "activated";

- FL UE Status = "join the federation";

The UE may notify the RAN through a new RRC message or by using an existing RRC message, such as UE Assistance Information. The RAN may also notify the CN through a new NGAP message or an existing NGAP message (such as PDU Session Resource Modify Indication or PDU Session Resource Notify procedure).

Through the above method, the problem of a waste of resources and/or a large signaling overhead, due to a case that one or more UE(s) with federated learning training tasks frequently leave and re-join the federation in different iterations, is solved/improved, and through different messaging methods, the embodiments of the present disclosure can adapt to more flexible scenarios and various different network architectures.

For the second situation, when the UE is in the model distribution stage or the training result reporting stage, situations such as network congestion or deterioration in wireless channel quality may occur. Either within the duration between the start time and the end time of the model distribution stage, if the model data is not received within a certain time window (UE-based implementation), or within the duration between the start time and the end time of the training result reporting stage, if the uplink training result data is not received within a certain time window (RAN-based implementation), it is possible to lead to a failure in model distribution or training result reporting. The final result is that the UE failed to participate in the federated learning training task for this iteration successfully. However, for the GBR QoS Flow set up by the UE, the network side is still in the status of reserving resources, but this iteration will not make use of resources. Then how to deal with this situation so as to avoid wasting wireless resources and core network resources.

It is a possible way to release the GBR QoS Flow through signaling, and then re-set up the corresponding GBR QoS Flow when the UE re-joins the federation. The corresponding signaling for releasing and setting up GBR QoS Flow will involve NG interface signaling, NAS signaling and RRC signaling. As shown in Fig. 7, the releasing and setting up of the QoS Flow can be implemented through a PDU session resource modify procedure without releasing the PDU session. As shown in Fig. 8, the releasing and setting up of the QoS Flow can be implemented by releasing and then re-setting up the PDU session resources for the entire FL task. However, no matter which signaling mechanism is used, a large signaling overhead will be generated. In order to avoid the problems caused by the above situation, the present application proposes an optimized processing mechanism for GBR QoS of federated learning training UE.

Please refer to Fig. 10. Fig. 10 shows another example procedure of GBR QoS processing for federated learning training UE in abnormal scenarios according to the embodiment of the present disclosure. As shown in Fig. 10, for the deactivating operation of the FL GBR QoS:

Alt.4 (Alternative 4) 1001: When the RAN side monitors that the service requirements for FL GBR QoS Flow cannot be fulfilled, the existing mechanism may be used to notify the CN. The CN notifies the AF/FL server, and then the AF/FL server decides that the UE leaves the federated learning training task for this iteration, and notifies the 5GC, and then the network element in the 5GC notifies the RAN side of the deactivation for the GBR QoS Flow corresponding to the FL training task of this UE, or the deactivation for the corresponding PDU Session. Deactivation means that in a case that the RAN side keeps the configuration of the existing GBR QoS Flow, it will no longer reserve the corresponding resources, so as to avoid wasting resources. In a way to notify, Alternative 1 in the embodiment shown in Figs. 9a and 9b can be used to implement corresponding actions to notify the RAN and UE.

In one embodiment, Alt.5 (Alternative 5) 1002 can be used. In the above-mentioned abnormal situations, the RAN may notify the CN and UE of deactivation for the GBR QoS Flow corresponding to the FL training task, which may include one or more of the following information:

- PDU Session ID or QoS Flow ID (QFI);

- PDU Session or QoS Flow Status = "deactivated".

The RAN may notify the CN through a new NGAP message or an existing NGAP message (such as PDU Session Resource Modify Indication or PDU Session Resource Notify procedure). Then the CN notifies the AF/FL server, and the FL server decides that the UE leaves the federated learning training for this iteration.

In this regard, for a UE-based implementation within a duration between the start time and the end time of the model distribution stage, if the model data is not received within a certain time window, the UE may notify the FL server of an event of a failure in model distribution through an application layer message, and then the FL server decides that the UE leaves the federated learning training task for this iteration, and then notifies the 5GC. Then, the network element in the 5GC notifies the RAN side of the deactivation for the GBR QoS Flow corresponding to the FL training task of this UE, or the deactivation of the corresponding PDU Session. Deactivation means that in a case that the RAN side keeps the configuration of the existing GBR QoS Flow, it will no longer reserve the corresponding resources, so as to avoid wasting resources. There are many ways to notify, and alternative 1 in the third embodiment can be used to implement corresponding actions to notify the RAN.

When the federated learning training UE is decided by the FL server to change from previously joining to now leaving this iteration, after the GBR QoS Flow or the corresponding PDU Session has been in a deactivated status through the above-mentioned Alternative 4 or Alternative 5, the situation of the status of UE participating in the federation may be restored during a certain subsequent iteration, that is, its conditions satisfy the requirements for the federated learning training again, so the UE needs to restore the corresponding PDU Session or GBR QoS Flow to an activated status for re-joining federated learning training. The activating operation of the FL GBR QoS can be implemented by using the mechanism of Alternative 11, Alternative 12 or Alternative 13 in the third embodiment.

Through the above method, the problem of a waste of resources and/or large signaling overhead, due to network congestion or deterioration in wireless channel quality of the user equipment in the federated learning model distribution stage or in the training result reporting stage, is solved/improved, and through different messaging methods, the embodiments of the present disclosure can adapt to more flexible scenarios and various different network architectures.

In the third situation, when the UE is in the training stage, a failure in training may occur due to the UE's own situations, such as increased computing load or other abnormal situations in computing. The final result is that the UE failed to participate in the federated learning training task for this iteration successfully. However, for the GBR QoS Flow set up by the UE, the network side is still in the status of reserving resources, but this iteration will not make use of resources. Then how to deal with this situation so as to avoid wasting wireless resources and core network resources.

It is a possible way to release GBR QoS Flow through signaling, and then re-set up the corresponding GBR QoS Flow when the UE re-joins the federation. The corresponding signaling for releasing and setting up GBR QoS Flow will involve NG interface signaling, NAS signaling and RRC signaling. As shown in Fig. 4, the releasing and setting up of the QoS Flow can be implemented through a PDU session resource modify procedure without releasing the PDU session. As shown in Fig. 5, the releasing and setting up of the QoS Flow can be implemented by releasing and then re-setting up the PDU session resources for the entire FL task. No matter which signaling mechanism is used, a large signaling overhead will be generated. In order to avoid the problems caused by the above situation, the present application proposes an optimized processing mechanism for GBR QoS of federated learning training UE.

Please refer to Fig. 11. Fig. 11 shows another example procedure of GBR QoS processing for federated learning training UE in abnormal scenarios according to an embodiment of the present disclosure. For the deactivating operation of the FL GBR QoS, the corresponding QoS processing mechanism can use Alternative 1 (1101) or Alternative 2 (1102) or Alternative 3 (1103) in the third embodiment.

The UE notifies the FL server of an event of a failure in training through an application layer message, and then the FL server decides that the UE leaves the federated learning training task for this iteration, and then notifies the 5GC. Then, the network element in the 5GC notifies the RAN side of the deactivation for the GBR QoS Flow corresponding to the FL training task of this UE, or the deactivation for the corresponding PDU Session. Deactivation means that in a case that the RAN side keeps the configuration of the existing GBR QoS Flow, it will no longer reserve the corresponding resources, so as to avoid wasting resources. There are many ways to notify, and alternative 1 in the third embodiment can be used to implement corresponding actions to notify the RAN and UE.

Optionally, when a failure in training occurs to the UE, the UE may notify the 5GC through NAS signaling of the deactivation for the GBR QoS Flow corresponding to the FL training task of this UE, or the deactivation for the corresponding PDU Session, and the 5GC then notifies the RAN side using Alternative 1. And the CN notifies the FL server, and then the FL server decides that the UE leaves the federated learning task for this iteration.

Optionally, when a failure in training occurs to the UE, the UE may notify the RAN through an RRC message, then the RAN notifies the CN, and the CN notifies the FL server of the deactivation for the GBR QoS Flow corresponding to the FL training task of this UE, or the deactivation for the corresponding PDU Session. Then, the FL server decides that the UE leaves the federated learning task for this iteration.

When a failure in training occurs to the federated learning training UE, and the UE is decided by the FL server to change from previously joining to now leaving this iteration, after the GBR QoS Flow or the corresponding PDU Session has been in a deactivated status through the above-mentioned Alternative 1, Alternative 2 or Alternative 3, the situation of the status of UE participating in the federation may be restored during a certain subsequent iteration, that is, its conditions satisfy the requirements for federated learning training again, so the UE needs to restore the corresponding PDU Session or GBR QoS Flow to an activated status for re-joining federated learning training. The activating operation of the FL GBR QoS can be implemented by using the mechanism of Alternative 11, Alternative 12 or Alternative 13 in the third embodiment.

Through the above method, the problem of a waste of resources and/or a large signaling overhead, due to a failure in training of the user equipment in the federated learning training stage, is solved/improved, and through different messaging methods, the embodiments of the present disclosure can adapt to more flexible scenarios and various different network architectures.

In the following, please continue to refer to Fig. 12. Fig. 12 shows another example procedure of GBR QoS processing when a federated learning training UE handovers based on Xn interface according to an embodiment of the present disclosure.

The current mechanism in the specification is that for an Xn-based handover, the handover request message 1201 carries information of GBR QoS Flow and Alternative QoS Parameter Set List needed to be set up, and if the target base station cannot satisfy the required QoS and Alternative QoS due to the current status (such as network congestion), the setting up of this QoS Flow will be rejected.

For the federated learning training UE, when the UE is in the model distribution stage or the training result reporting stage, a handover may occur due to movement. If the target base station rejects to set up GBR QoS Flow due to temporary network congestion during the handover, and then the UE re-joins the federation in a certain iteration after the status is restored, then a situation will occur that the GBR QoS Flow serving for federated learning training is re-set up at the target base station. In order to save signaling overhead, a new mechanism may be taken into consideration to be introduced. Under the above circumstances, the target base station still accepts to set up the GBR QoS Flow of the corresponding federated learning training service, but configures the status of the QoS as a deactivated status, and there is no need for the RAN side to reserve resources. The corresponding handover request acknowledge message 1202 may include one or more of the following information:

- PDU Session ID or QoS Flow ID (QFI);

- PDU Session or QoS Flow Status = "deactivated".

Accordingly, the source base station notifies the UE by transmitting an RRC reconfiguration message 1203, which may include one or more pieces of information:

- PDU Session ID or QoS Flow ID (QFI);

- PDU Session or QoS Flow Status = "deactivated".

Accordingly, the target base station notifies the CN of the deactivation for the GBR QoS Flow corresponding to the FL training task through an NG message, and the path switch request message 1204 may include one or more of the following information:

- PDU Session ID or QoS Flow ID (QFI).

- PDU Session or QoS Flow Status = "deactivated"

Then, the CN notifies 1205 the AF/FL server, and the FL server decides that the UE leaves the federated learning training for this iteration. In a certain subsequent iteration, the status of network congestion may be restored, which is notified to the CN through the existing mechanism, and in turn notified to the AF/FL server. In this way, the UE needs to restore the corresponding PDU Session or GBR QoS Flow to an activated status for re-joining federated learning training. The activating operation of the FL GBR QoS can be implemented by using the mechanism of Alternative 11, Alternative 12 or Alternative 13 in the third embodiment.

Through the above method, the problem of a waste of resources and/or a large signaling overhead, due to a situation that when a handover occurs because the UE in the model distribution stage or the training result reporting stage moves, the target base station rejects to set up GBR QoS Flow because of temporary network congestion, and then the UE re-joins the federation in a certain iteration after the status is restored, is solved/improved, and through different messaging methods, the embodiments of the present disclosure can adapt to more flexible scenarios and various different network architectures.

In the following, please continue to refer to Fig. 13. Fig. 13 shows another example procedure of GBR QoS processing when federated learning training UE handovers based on the Ng interface according to an embodiment of the present disclosure.

The current mechanism in the specification is that for an Ng-based handover, the handover request message 1301 carries information of GBR QoS Flow and Alternative QoS Parameters Set List needed to be set up. If the target base station cannot satisfy the required QoS and Alternative QoS due to the current status (such as network congestion), the setting up of this QoS Flow will be rejected.

For a federated learning training UE, when the UE is in the model distribution stage or the training result reporting stage, a handover may occur due to movement. If the target base station rejects to set up GBR QoS Flow due to temporary network congestion during the handover, and then the UE re-joins the federation in a certain iteration after the status is restored, then the GBR QoS Flow serving for federated learning training will be re-set up at the target base station. In order to save signaling overhead, a new mechanism may be taken into consideration to be introduced. Under the above circumstances, the target base station still accepts to set up the GBR QoS Flow of the corresponding federated learning training service, but configures the status of the QoS as a deactivated status, and there is no need for the RAN side to reserve resources. The corresponding handover request acknowledge message 1302 may include one or more of the following information:

- PDU Session ID or QoS Flow ID (QFI);

- PDU Session or QoS Flow Status = "deactivated".

Accordingly, the CN forwards the status of the setting up of the QoS Flow of the target base station to the source base station, and the corresponding handover command 1303 may include one or more of the following information:

- PDU Session ID or QoS Flow ID (QFI);

- PDU Session or QoS Flow Status = "deactivated".

Accordingly, the source base station notifies the UE by transmitting an RRC reconfiguration message 1304, which may include one or more pieces of information:

- PDU Session ID or QoS Flow ID (QFI);

- PDU Session or QoS Flow Status = "deactivated".

Then, the CN notifies 1305 the AF/FL server, and the FL server decides that the UE leaves the federated learning training for this iteration. In a certain subsequent iteration, the status of network congestion may be restored, which is notified to the CN through the existing mechanism, and in turn notified to the AF/FL server. In this way, the UE needs to restore the corresponding PDU Session or GBR QoS Flow to an activated status for re-joining federated learning training. The activating operation of the FL GBR QoS can be implemented by using the mechanism of Alternative 11, Alternative 12 or Alternative 13 in the third embodiment.

Through the above method, the problem of a waste of resources and/or a large signaling overhead, due to a situation that when a handover occurs because the UE in the model distribution stage or the training result reporting stage moves, the target base station rejects to set up GBR QoS Flow because of temporary network congestion, and the UE re-joins the federation in a certain iteration after the status is restored, is solved/improved, and by setting different messaging methods under another network architecture, the embodiments of the present disclosure can adapt to more flexible scenarios and various different network architectures.

In the above embodiments, with respect to the core idea of the optimized processing mechanism for GBR QoS in federated learning training tasks, a UE-based and GBR QoS Flow-based activation and deactivation mechanism or a corresponding PDU Session activation and deactivation mechanism is introduced. In the case of the split-gNB, after receiving the status indication information about GBR QoS Flow from the 5GC through the NG interface, the CU-CP may further indicate it to the CU-UP through the E1 interface, because the CU-UP is responsible for the resources for the UE's user plane, possibly through a new message or by adding a new IE to an existing message (such as BEARER CONTEXT SETUP REQUEST or BEARER CONTEXT MODIFICATION REQUEST). One or more of the following information may be included in the new message or added to the existing message:

- PDU Session ID or QoS Flow ID (QFI);

- PDU Session or QoS Flow Status = "activated or deactivated".

After receiving the status indication information about GBR QoS Flow from the 5GC through NG interface, the CU-CP may further indicate it to the DU through the F1 interface, because the DU is also responsible for the resources for the UE's user plane, possibly through a new message or by adding a new IE to an existing message (such as UE CONTEXT SETUP REQUEST or UE CONTEXT MODIFICATION REQUEST). One or more of the following information may be included in the new message or added to the existing message:

- PDU Session ID or QoS Flow ID (QFI);

- PDU Session or QoS Flow Status = "activated or deactivated".

For a UE in a dual connectivity, after receiving the status indication information about GBR QoS Flow from the 5GC through the NG interface, the master node (MN) may further indicate it to the SN through the Xn interface, possibly through a new message or by adding a new IE to an existing message (such as S-NODE ADDITION REQUEST, S-NODE MODIFICATION REQUEST or NOTIFICATION CONTROL INDICATION). One or more of the following information may be included in the new message or added to the existing message:

- PDU Session ID or QoS Flow ID (QFI);

- PDU Session or QoS Flow Status = "activated or deactivated".

If the secondary node is also a split-gNB, the CU-CP of the secondary node may also indicate to the CU-UP and DU in the corresponding secondary node through the E1 and F1 interfaces.

With respect to the indication about activation and deactivation for GBR QoS Flow initiated by the RAN, in the scenario of the dual connectivity (DC), the SN may notify the MN of it through the Xn interface, and then the MN notifies the 5GC of it, possibly through a new message or by adding a new IE to an existing message (such as S-NODE MODIFICATION REQUIRED or NOTIFICATION CONTROL INDICATION). One or more of the following information may be included in the new message or added to the existing message:

- PDU Session ID or QoS Flow ID (QFI);

- PDU Session or QoS Flow Status = "activated or deactivated".

Through the above method, the embodiments of the present disclosure can adapt to more flexible scenarios and various different network architectures.

In one embodiment, the UE-based and GBR QoS Flow-based activation and deactivation mechanism or the corresponding PDU Session activation and deactivation mechanism can also be applied to the existing GBR QoS Flow. For the indication of activation and deactivation for GBR QoS Flow, in addition to a way to define a new message, a way to add a new IE to an existing message may be used. One of the possible ways is to add status indication information on the basis of GBR QoS Flow Information IE, which specifically involves IEs in the following specifications.

◆ NGAP 9.3.1.10 GBR QoS Flow Information in TS 38413-h40

◆ XnAP 9.2.3.6 GBR QoS Flow Information in TS 38423-h40

◆ F1AP 9.3.1.46 GBR QoS Flow Information in TS 38473-h41

◆ E1AP 9.3.1.30 GBR QoS Flow Information in TS 37483-h40

In the following, two implementations of carrying out the GBR QoS Flow-based activation and deactivation mechanism by adding a new IE (which indicates the QoS parameters of GBR QoS flows for downlink and uplink) in the current specification are given in the form of table. The first implementation is shown in Table 1 below.

Optionally, in a case of being at the same level as QoS Flow Identifier IE, the indication information (QoS Flow Status) can be added as the status indication information of GBR QoS Flow, as shown in Table 2 below.

Through the above method, the embodiment of the present disclosure can be improved under the definition of the existing specification, and the embodiment of the present disclosure can be compatible with the current specification.

For the status indication information of GBR QoS Flow of UE, during the mobility of the UE, such as a handover occurring based on Xn interface or Ng interface, or a RRC resume procedure occurring on a new base station after a mobility occurs to the UE in the RRC_INACTIVE status, the UE's status of GBR QoS Flow on the source base station will be transferred to the new base station and be maintained. For example, if the status of GBR QoS Flow is a deactivated status on the source base station, the status of GBR QoS Flow will be transferred and maintained as a deactivated status after the UE hands over to the new base station. If the status of GBR QoS Flow is an activated status on the source base station, the status of GBR QoS Flow will be transferred and maintained an activated status after the UE hands over to the new base station, but in some cases, it may also become a deactivated status.

Based on the procedure of federated learning training, it is found that for model distribution, each UE participating in the training receives the same data, in order to further save the resources for the user planes at the RAN side and the 5GC side. Federated learning FL can introduce the mechanism for MBS (Multicast/Broadcast service) for model distribution. Since FL has not only downlink data (for model distribution), but also uplink data (for training result reporting), and for different UEs possibly selected for different iterations, not all UEs in each cell participate in the federation, it is more appropriate to use Multicast service.

Please refer to Fig. 14. Fig. 14 shows an example procedure triggered by UE in which a federated learning training UE joins a multicast MBS for model distribution according to an embodiment of the present disclosure. As shown in Fig. 14, the traditional MBS multicast mechanism is that after the user receives the notification of the multicast service, if the user is interested in this multicast service, the UE transmits NAS signaling 1401 to initiate a request to join in the multicast service. However, for federated learning, the selection of UEs participating in federated learning training in each iteration is decided by the FL server. In order to unify the mechanism for using MBS in FL, and in order to make the MBS serve for FL better, avoiding a situation that the multiple UEs selected by FL could have used MBS mechanism to distribute downlink model data but the UE did not initiate the request to join the MBS, the FL server can further decide 1402 which ones of the UEs participating in the federation are suitable to use the MBS for model distribution, and then notify the CN. In 1403, then the CN initiates to set up of MBS session for the corresponding base station, and then the base station configures the multicast MBS radio resources for the corresponding UE.

Please refer to Fig. 15. Fig. 15 shows an example procedure triggered by FL server in which a federated learning training UE joins a multicast MBS for model distribution according to an embodiment of the present disclosure. As shown in Fig. 15, in the UE selection stage, in addition to reporting the computing resources and wireless channel status, the UE may also report 1501 more detailed location information, such as reporting the cell ID information to which the UE is currently attached, tracking area ID (TAI), base station ID information, etc., so that the FL server may determine whether this UE participates in the federation and decide whether to apply multicast MBS for model distribution, thereby saving the resources for the user planes at the RAN side and the 5GC side. For example, when the FL server finds 1502 that different UEs participating in federated learning training are in the same base station or the same cell, applying multicast MBS for model distribution can save network bearer resources. Using the existing mechanism, the CN may be aware of whether the RAN node supports the multicast MBS function. Then, the CN side may transmit to the RAN a multicast MBS session resource setup request 1503. This signaling is non-UE-associated and based on MBS session, which prepares resources at the RAN side for the UEs using MBS for model distribution under this RAN node. The multicast MBS session resource setup request may include one or more of the following information:

- MBS Session ID;

- Shared NG-U TNL Information at UPF;

- MBS QoS Flows To Be Setup List (including GBR QoS information required by FL);

- Start Time and End Time of MBS, and/or a sequence of scheduled activation times and deactivation times (e.g., the first start time and the first end time, and the periodicity) of the MBS session (indicating the start time and the end time of model distribution in different iteration);

- MBS Session Status (activated, deactivated, …);

- MBS UE ID list (the UE IDs to which the MBS is to be applied under the corresponding base station, which may be represented by RAN UE NGAP ID, and other representations are not excluded, as long as which UEs under the base station the MBS is applied to are indicated).

After successfully configuring the corresponding UE through an RRC message, the RAN side transmits a multicast MBS session resource setup response message 1504 to the CN, which may include one or more of the following information:

- MBS Session ID;

- Shared NG-U TNL Information at NG-RAN;

- MBS QoS Flows Setup List (including information of successfully setup GBR QoS Flow).

Then the FL server begins to distribute the model data to the UE that joined the multicast.

In order to better adapt to the application of MBS in FL and save resources, the above-mentioned strategy that the FL server decides which ones of the UEs participating in the federation needs to use MBS for model distribution simplifies the signaling procedure, compared with the traditional mechanism for joining and exiting multicast MBS initiated by the UE, thereby better embodying the idea of Network For AI.

For the procedure of setting up multicast MBS session resources initiated by the CN side, if the RAN side rejects it due to some abnormal situations, at this time, the RAN side transmits a multicast MBS session resource setup failure message. This message may include one or more of the following information:

- MBS Session ID;

- Cause.

The CN side may also make necessary modifications to certain parameters (such as a change in a selected UE list, GBR QoS required by FL, an MBS Session Status, MBS start time and MBS end time, etc.) in different iterations or at appropriate occasions, and the CN side may trigger a multicast MBS session resource update procedure. Under normal circumstances, the RAN side transmits a response of successful update, for example, a multicast MBS session resource update response message; or due to some abnormal situations, the RAN side transmits a response of failed update, for example, a multicast MBS session resource update failure message, which may include one or more of the following information:

- MBS Session ID;

- Cause.

Accordingly, if MBS is no longer used for model distribution or federated learning training is completed, the CN obtains corresponding instructions from the AF/FL server, and the CN side may transmit a multicast MBS session resource release request message to the RAN, and the RAN side may transmit a multicast MBS session resource release response to the CN after releasing MBS session resources.

- MBS Session ID;

- Cause.

The above method, which is different from the method triggered by the UE, is a method triggered by the FL server in regard to federated learning training UE joining multicast MBS for model distribution, and it solves/improves the problem that multiple UEs selected by FL could have used MBS mechanism to distribute downlink model data but the UE did not initiate a request to join MBS.

It can be understood that the above various embodiments can be combined with each other to form a new embodiment, which can adapt to more flexible scenarios and various different network architectures through various combinations.

In addition, it can be understood that the message names in the present invention are only examples, and other message names can also be used. For example, the first message may also be referred to as a deactivation indication message, the second message may also be referred to as an activation indication message, the third message may also be referred to as a deactivation response message, the fourth message may also be referred to as an activation response message, the fifth message may also be referred to as a deactivation pre-configuration message, and the sixth message may also be referred to as a deactivation pre-configuration response message, and so on, which is not limited by the present disclosure. Moreover, the words such as "first" and "second" included in the message names of the present invention are only examples of messages and do not represent the execution order.

In addition, it can be understood that the contents contained in respective messages in the above-mentioned embodiments are only examples of the present disclosure, and the contents contained in various similar or identical messages can be replaced with each other, or named with different names, or can be added to or deleted from some items according to the situation, and for the sake of brevity, the meanings of similar content items can be mutually explained according to the context and will not be detailed, and the modified and explained implementations still belong to the scope of protection of the present invention.

At least one embodiment of the present disclosure also provides a non-transitory computer-readable recording medium having stored thereon a program for performing the above methods when executed by a computer.

Various embodiments of the present disclosure can be implemented as computer-readable codes embodied on a computer-readable recording medium from a specific perspective. A computer-readable recording medium is any data storage device that can store data readable by a computer system. Examples of computer-readable recording media may include read-only memory (ROM), random access memory (RAM), compact disk read-only memory (CD-ROM), magnetic tape, floppy disk, optical data storage device, carrier wave (e.g., data transmission via the Internet), and the like. Computer-readable recording media can be distributed through computer systems connected via a network, and thus computer-readable codes can be stored and executed in a distributed manner. Moreover, functional programs, codes and code segments for implementing various embodiments of the present disclosure can be easily interpreted by those skilled in the art to which the embodiments of the present disclosure are applied.

It will be understood that the embodiments of the present disclosure can be implemented in the form of hardware, software, or a combination of hardware and software. The software can be stored as program instructions or computer-readable code executable on a controller on a non-transitory computer-readable medium. Examples of non-transitory computer-readable recording media include magnetic storage media (e.g., ROM, floppy disk, hard disk, etc.) and optical recording media (e.g., CD-ROM, digital video disk (DVD), etc.). Non-transitory computer-readable recording media can also be distributed on network-coupled computer systems, so that computer-readable codes are stored and executed in a distributed manner. The medium can be read by a computer, stored in a memory, and executed by a controller. Various embodiments can be implemented by a computer or a portable terminal including a controller and a memory, and the memory can be an example of a non-transitory computer-readable recording medium suitable for storing program(s) having instructions for implementing embodiments of the present disclosure. The present disclosure can be implemented by a program having codes for specifically implementing the apparatus and method described in the claims. The said program is stored in a machine (or computer) readable storage medium. The said program may be electronically carried on any medium, such as a communication signal transmitted via a wired or wireless connection, and the present disclosure suitably includes its equivalents.

What has been described above is only the specific implementation of the present disclosure, but the protection scope of the present disclosure is not limited to this. Any technicist familiar with this technical field can make various changes or substitutions within the technical scope disclosed in the present disclosure, and these changes or substitutions should be included in the protection scope of the present disclosure. Therefore, the protection scope of the present disclosure should be based on the protection scope of the claims.

Claims

A method performed by a first network node in a communication system, the method comprising:

receiving a first message transmitted by a first node, wherein the first message includes information for indicating to deactivate a quality of service (QoS) flow of user equipment (UE),

transmitting the information for indicating to deactivate the QoS flow of the UE to a second node, and deactivating the QoS flow of the UE.
The method according to claim 1, wherein the method further comprises:

receiving a second message transmitted by the first node, wherein the second message includes information for indicating to activate the QoS flow of the UE, and

activating the QoS flow of the UE, and transmitting the information for indicating to activate the QoS flow of the UE to the second node,

wherein the second message include at least one of:

protocol data unit (PDU) session identification (ID) or QoS flow identification (ID); or

a PDU session or QoS flow status, which is set to be activated.
The method according to claim 1, wherein the method further comprises:

receiving a fifth message transmitted by a second network node,

wherein the fifth message includes configuration information related to deactivating or activating the QoS flow,

wherein the configuration information is received when a PDU session or a QoS flow is initially set up or during one of a plurality of training iterations.
The method according to claim 3, wherein the configuration information includes at least one of:

protocol data unit (PDU) session identification (ID) or QoS flow identification (ID);

a time point at which the QoS flow is activated for a first time, a time point at which the QoS flow is deactivated for a first time, a time point at which the QoS flow is activated for a second time, a time point at which the QoS flow is deactivated for a second time, during one training iteration;

a periodicity of model training iterations; or

a number of model training iterations.
The method according to claim 1, wherein the method further comprises:

receiving a seventh message transmitted by the first node,

wherein the seventh message includes information for indicating to deactivate the QoS flow of the UE in a case that the user equipment leaves federated learning,

wherein the seventh message includes at least one of:

protocol data unit (PDU) session identification (ID) or QoS flow identification (ID);

a PDU session or QoS flow status, which is set to be deactivated;

a user equipment status of the federated learning, which is set to instruct the user equipment to leave the federated learning;

federated learning application identification, which is used to identify a federated learning training task,

wherein the user equipment leaving the federated learning comprises at least one of:

the user equipment leaving the federated learning due to not meeting federated learning training requirements in one iteration after joining the federated learning;

the user equipment leaving the federated learning due to a failure in model distribution or training result reporting, resulting from an occurrence of network congestion or deterioration in wireless channel quality during a federated learning model distribution stage or a training result reporting stage, or in a case that learning model related data is not received within a predetermined time window during the model distribution stage, or in a case that no training result data is received by the first network node within a predetermined time window during the training result reporting stage;

the user equipment leaving the federated learning due to a failure in training during the federated learning training stage; or

the user equipment leaving the federated learning in a case that the first network node monitors that service requirements for the QoS flow cannot be fulfilled.
The method according to claim 1, wherein the method further comprises:

receiving a ninth message transmitted by the first node,

wherein the ninth message includes information for indicating to activate the QoS flow of the UE in a case that the user equipment re-joins the federated learning after leaving the federated learning,

wherein the ninth message includes at least one of:

protocol data unit (PDU) session identification (ID) or QoS flow identification (ID);

a PDU session or QoS flow status, which is set to be activated;

a user equipment status of the federated learning, which is set to instruct the user equipment to join the federated learning;

federated learning application identification, which is used to identify a federated learning training task.
The method according to claim 1, wherein the method further comprises:

transmitting an eleventh message to the first node and the second node,

wherein the eleventh message includes information for indicating to deactivate the QoS flow in a case that the first network node monitors that the service requirements for the QoS flow cannot be fulfilled.
The method according to claim 1 or 7, wherein the first message and the eleventh message include at least one of:

protocol data unit (PDU) session identification (ID) or QoS flow identification (ID); or

a PDU session or QoS flow status, which is set to be deactivated.
The method according to any one of claims 1 to 7, wherein

the first network node is a base station, the first node includes one of the user equipment or the second network node, and the second node includes the other one of the user equipment or the second network node, and the second network node is a core network device; or

the first network node is a centralized unit control plane (CU-CP) of the base station, the first node is a core network device, and the second node is a centralized unit user plane (CU-UP) or a distribution unit (DU) of the base station; or

the first network node is a master node (MN), the first node is a core network device, and the second node is a secondary node (SN).
The method according to claim 1, wherein the method further comprises:

receiving a handover request acknowledge message transmitted by a third network node, and

transmitting a radio resource control (RRC) reconfiguration message to the UE,

wherein the handover request acknowledge message and the RRC reconfiguration message include at least one of:

protocol data unit (PDU) session identification (ID) or QoS flow identification (ID); or

a PDU session or QoS flow status, which is set to be deactivated.
The method according to any one of claims 1 to 10, wherein the method further comprises:

receiving a handover command transmitted by the second network node, and

transmitting the radio resource control (RRC) reconfiguration message to the UE,

wherein the handover command and the RRC reconfiguration message include at least one of:

protocol data unit (PDU) session identification (ID) or QoS flow identification (ID); or

a PDU session or QoS flow status, which is set to be deactivated.
A method performed by a second network node, the method comprising:

transmitting a first message to a first network node and deactivating a quality of service (QoS) flow of user equipment (UE), wherein the first message includes information for indicating to deactivate the QoS flow; and/or

transmitting a second message to the first network node and activating the QoS flow of the UE, wherein the second message includes information for indicating to activate the QoS flow.
The method according to claim 12, further comprising:

receiving a path switch request message and/or a handover request acknowledge transmitted by a third network node,

wherein the path switch request message includes at least one of:

protocol data unit (PDU) session identification (ID) or QoS flow identification (ID); or

a PDU session or QoS flow status, which is set to be deactivated,

wherein the handover request acknowledge includes at least one of:

protocol data unit (PDU) session identification (ID) or QoS flow identification (ID); or

a PDU session or QoS flow status, which is set to be deactivated.
A method performed by user equipment (UE) in a communication system, the method comprising:

determining that conditions for deactivating and/or activating a quality of service (QoS) flow of the UE are satisfied,

transmitting a first message to a first network node or a second network node, wherein the first message includes information for indicating to deactivate the QoS flow, and/or

transmitting a second message to the first network node or the second network node, wherein the second message includes information for indicating to activate the QoS flow.
A node device, comprising:

a transceiver; and

a controller coupled with the transceiver and configured to perform the method of any one of claims 1 to 14.