A network resource allocation method for 5G-TSN

Abstract

The present invention relates to a network resource allocation method for 5G-TSN, and belongs to the field of communications technology. This method allocates 5G-TSN network resources. Based on the resource allocation requirements of data streams on the TSN side, the resource allocation requirements of data streams on the 5G side are determined by combining the end-to-end transmission requirements and TSN transmission requirements of data streams in different scenarios, as well as the transmission requirements of data streams on the 5G side. This method can effectively allocate resources in 5G-TSN, thereby ensuring low-latency and high-reliability transmission of data streams in 5G-TSN.

Description

Technical Field

[0001] The present invention belongs to the field of communication technology and relates to a network resource allocation method for 5G-TSN. Background Art

[0002] 5G, with its large bandwidth, low latency, and high reliability, can meet the flexible mobility and differentiated business processing capabilities required by industrial equipment. It will promote wireless applications in various augmented reality / virtual reality terminals, robots, automated guided vehicles (AGVs), and on-site production line equipment, helping to achieve widespread adoption of flexible factory production. However, industrial applications may require networks with deterministic latency, jitter, and higher reliability, which poses unprecedented challenges for 5G. Time-Sensitive Networking (TSN), which provides deterministic forwarding for services with strong real-time requirements and is compatible with Ethernet protocols, has gained popularity in the industrial control network sector. However, with the influx of intelligent devices such as automated guided vehicles (AGVs) in smart factories, the demand for wireless and mobile connectivity for industrial terminals is becoming increasingly urgent. Wired TSN cannot meet the digital and intelligent needs of smart factories. The convergence of 5G and TSN is a key evolutionary trend in smart factory networks.

[0003] When 5G and TSN converge, networks face challenges such as resource allocation, data transmission, and deterministic scheduling. Network resource allocation is a prerequisite for data transmission and deterministic scheduling, making research on 5G-TSN network resource allocation urgent. Currently, research on 5G and TSN resource allocation has been conducted domestically and internationally, using various approaches, including SDN and NFV algorithms, maximum-minimum distance K-means clustering algorithms, on-demand scheduling algorithms, network calculus methods, and ILP algorithms. Research on heterogeneous network resource allocation after convergence has also been conducted using improved genetic algorithms, reinforcement learning, and latency constraints. While extensive research has been conducted on 5G-TSN resource allocation methods, due to the varying transmission requirements of 5G-TSN in different industrial scenarios, existing methods fail to incorporate the unique characteristics of 5G and TSN (data transmission protocols, real-time performance, reliability, and low latency). Directly adopting these approaches cannot meet the many demands of 5G-TSN.

[0004] In view of the current network resource allocation difficulties, it is urgent to study a new network resource allocation method for 5G-TSN to solve the above problems. Summary of the Invention

[0005] In view of this, the object of the present invention is to provide a network resource allocation method for 5G-TSN, which effectively allocates resources in 5G-TSN, thereby ensuring low-latency and high-reliability transmission of data streams in 5G-TSN.

[0006] In order to achieve the above object, the present invention provides the following technical solutions:

[0007] A network resource allocation method for 5G-TSN, specifically comprising the following steps:

[0008] S1: Build a network architecture for 5G-TSN network resource allocation;

[0009] S2: SDNC obtains network topology and data flow information of the entire network through CNC and TSN AF and manages the information;

[0010] S3: SDNC calculates the priority of data stream transmission in TSN based on the end-to-end priority input by the user;

[0011] S4: CNC uses an iterative resource allocation algorithm based on SMT to allocate network resources on the TSN side for data flows. SMT stands for Satisfiability Modulo Theory.

[0012] S5: SDNC calculates the transmission delay of each data stream on the TSN side;

[0013] S6: SDNC calculates the transmission delay of each data stream on the 5G side based on the end-to-end delay of each data stream and the transmission delay on the TSN side;

[0014] S7: The SDNC pre-configures a mapping table and calculates the packet delay budget, 5QI, and priority of each data flow on the 5G side.

[0015] S8: SDNC allocates network resources for data flows on the 5G side.

[0016] Furthermore, in step S1, the constructed network architecture includes a software-defined networking controller (SDNC), two centralized network configurations (CNCs), one 5G and two TSNs.

[0017] The 5G network includes user equipment (UE), radio access network (RAN), and 5G core network. The 5G core network includes unified data management (UDM), network exposure function (NEF), access and mobility management function (AMF), session management function (SMF), policy control function (PCF), time-sensitive network application function (TSN AF), user plane function (UPF), device-side TSN translator (DS-TT), and network-side TSN translator (NW-TT).

[0018] TSN includes TSN terminals and TSN switches.

[0019] The network architecture is divided into the control plane and the data plane, among which SDNC, CNC, UDM, NEF, AMF, SMF, PCF, and TSN AF are located on the control plane, and TSN terminals, TSN switches, DS-TT, UE, (R)AN, UPF, and NW-TT are located on the data plane.

[0020] On the control plane, SDNC centrally manages the entire network through CNC and TSN AF, discovers and updates 5G and TSN (hereinafter referred to as: 5G-TSN) topology, and formulates 5G-TSN resource allocation strategy.

[0021] CNC is mainly responsible for discovering the TSN physical topology, retrieving TSN switches, and implementing functions such as scheduling traffic; interacting with TSN AF to obtain 5GS bridge information; dividing the time slots of TSN switches and sending the calculation results to the TSN switches.

[0022] TSN AF is mainly responsible for interacting with CNC to realize the interaction and transmission of TSN flow transmission direction, flow cycle, transmission delay budget, data priority and other parameters with 5G.

[0023] On the data plane, 5G acts as a logical bridge in TSN (hereinafter referred to as: 5GS bridge). The 5GS bridge includes DS-TT, the data plane tunnel between the UE and UPF, and NW-TT; the ports on the NW-TT support connection with TSN, and the ports on the DS-TT side are associated with PDU sessions, thereby providing 5G connection with TSN.

[0024] Furthermore, step S2 specifically includes the following steps:

[0025] S21: The two CNCs discover the TSN network topology based on LLDP and forward the TSN network topology information to the SDNC. LLDP stands for Link Discovery Protocol.

[0026] S22: The terminal device sends information about each data stream, including the transmission period, data frame size, priority, and latency requirements, to the centralized user configurator through the user configuration protocol. The centralized user configurator sends the data stream information to the two CNCs through the user network interface. The two CNCs forward the data stream information to the SDNC.

[0027] S23: SDNC organizes the collected network topology information and data flow information, wherein the network topology information includes the link information in the first TSN, the link information from the first TSN to 5G, the link information in 5G, the link information from 5G to the second TSN, and the link information in the second TSN; the data flow information includes end-to-end delay requirements and data flow priority; the network topology consists of: a TSN terminal as a transmitter is connected to a TSN switch through DS-TT to enter the 5G network, then connected to another TSN switch through NW-TT, and finally connected to another TSN terminal as a receiver.

[0028] Furthermore, step S3 specifically includes the following steps:

[0029] S31: In a network composed of 5G and TSN, there are k data flows, and different data flows have different priorities, that is, data flow f k The corresponding end-to-end priority is priority k The end-to-end priority input by the user is designed to range from 1 to 15, with smaller values ​​giving higher priorities.

[0030] S32: SDNC obtains the priority of the TSN side based on the input end-to-end priority and the mapping rule between the established end-to-end priority and the TSN side priority, i.e., the priority of the data flow f k The corresponding TSN side priority is Among them, there are 8 priorities on the TSN side according to the standard, namely 0 to 7, and the larger the value, the higher the priority;

[0031] The mapping rule between the end-to-end priority and the TSN side priority is: the end-to-end priority is 1 to 15, and the corresponding TSN side priority is 7, 6, 6, 5, 5, 4, 4, 3, 3, 2, 2, 1, 1, 0, 0.

[0032] Furthermore, in step S4, the SMT-based iterative resource allocation algorithm specifically comprises the following steps: converting the TSN data stream slot planning problem into a constraint satisfaction problem; based on the characteristics of SMT, using the transmission slot and the gating cycle waiting quantity to be solved as variables, using the network status and data stream information as constants, and converting the data stream delay requirements into constraints; finally, using the SMT solver to solve the constraints to obtain the slot planning result;

[0033] The SMT-based iterative resource allocation algorithm specifically includes the following steps:

[0034] S41: The two CNCs obtain information about all data flows in this resource allocation, including the data flow transmission cycle, data frame size, priority, TSN switch transmission rate, TSN switch processing delay, etc., and use it as the input of the algorithm; the transmission delay of each data flow on the first TSN side is The demand calculation formula is shown in (1), where n = 33%:

[0035]

[0036] The transmission delay of each data stream on the second TSN side The demand calculation formula is shown in (2), where n = 33%:

[0037]

[0038] Where F represents the data flow set;

[0039] S42: The two CNCs add constants and variables to the SMT solver. The constants include the size and priority of each flow's data frame, the sending period, the sending rate of the TSN switch, and the processing delay of the TSN switch. This information is mainly used to model the network and data flow. The variables include the gating period of the TSN switch, the number of transmission time slots of each flow within the gating period, the start time of each transmission time slot of the data flow, the arrival time and departure time of the data flow in each time slot, the sending delay of the data flow on the TSN switch, and the transmission delay requirement. This information needs to be solved according to the constraints and is the output of the planning algorithm.

[0040] S43: The two CNCs add constraints to the SMT solver to constrain the value range of the variables so that they meet the network resource limitations and data flow transmission quality requirements. The specific constraints are as follows:

[0041] 1) The TSN switch gating cycle Schedcycle is the lowest common multiple of the transmission cycles of all data flows flowing through the TSN switch, and needs to satisfy formula (3):

[0042]

[0043] Among them, F p Represents the set of data flows flowing through the TSN switch, T k represents the period of data stream k;

[0044] 2) The number of transmission time slots Scount of data stream k in the TSN switch gating cycle k It is equal to the TSN switch gating period divided by the data flow sending period, as shown in formula (4):

[0045]

[0046] 3) The start time of the transmission time slot of the data flow in the TSN switch gating and the transmission delay need to satisfy the relationship as shown in formula (5) and formula (6); the transmission delay is equal to the time required for the amount of data transmitted in one transmission cycle of the data flow (one data frame is sent in one transmission cycle) to be sent from the TSN switch; the start time of the time slot must be within the gating cycle;

[0047]

[0048] 0≤Sstart k,i ≤Schedcycle-SendDelay k,i (5)

[0049]

[0050] Among them, Sstart k,i SendDelay represents the start time of the i-th transmission time slot of data stream k in the TSN switch gating cycle, k,i Fsize represents the sending delay of the i-th transmission time slot of data stream k in the TSN switch gating cycle; k Indicates the size of the data frame of data stream k, and Sendrate indicates the sending rate of the TSN switch;

[0051] 4) Within the gating cycle of the TSN switch, the interval between different transmission time slots of the same data stream must be greater than the sending cycle of the data stream, as shown in formula (7):

[0052]

[0053] Sstart k,i +T k ≤Sstart k,i+1 (7)

[0054] 5) The transmission time slots within the gating cycle of the TSN switch cannot overlap, that is, the start time of a transmission time slot of data stream k must be greater than or equal to the end time of a transmission time slot of data stream g, or the end time must be less than or equal to the start time of a transmission time slot of data stream g, as shown in formula (8):

[0055]

[0056] (Sstart k,i +SendDelay k,i ≤Sstart g,j ) || (Sstart g,j +SendDelay g,j ≤ Sstart k,i (8)

[0057] 6) According to the characteristics of the transmission time slot, the arrival and departure time of the data frame of each data stream at different transmission time slots can be obtained; the arrival time of the data frame to the TSN switch is earlier than the departure time, and the relationship is shown in formula (9), formula (10), and formula (11):

[0058]

[0059]

[0060] Tleave k,i =Sstart k,i +SendDelay k,i (10)

[0061] Tarrival k,i <Tleave k,i (11)

[0062] Among them, Tarrival k,i represents the time when the data of the i-th transmission time slot of data stream k arrives at the TSN switch, Tleave k,i Indicates the time when the data of the i-th transmission time slot of data stream k leaves the TSN switch, Hsend k represents the sending time of data stream k at the sending end;

[0063] 7) The end-to-end delay of different time slots of the same data stream must be less than the upper limit of the delay, as shown in formula (12):

[0064]

[0065]

[0066] S44: The two CNCs run the SMT solver. The solver will try all possible values ​​of the variables in the solution space in turn. If a feasible solution is found, the transmission delay requirement of each data stream is iterated. The calculation formulas are shown in (13), (14), and (15). After the calculation, the process returns to step S42.

[0067] n=n-1% (13)

[0068]

[0069]

[0070] If no feasible solution is found after traversing all the values, the transmission delay requirement of each data stream is iterated again, and the calculation formulas are shown in (16), (17), and (18). After the calculation, return to step S42; if a feasible solution is found, the time slot division result is output;

[0071] n=n+0.1% (16)

[0072]

[0073]

[0074] The two CNCs encapsulate the calculation results into XML files and use the NETCONF protocol to configure the XML-based gating scheduling table to the TSN switch.

[0075] Furthermore, step S5 specifically includes the following steps:

[0076] S51: The first CNC obtains the transmission delay of each data stream on the first TSN side through step S4. The second CNC obtains the delay of each data stream when it is transmitted on the second TSN side through step S4 The two CNCs forward the delay information to the SDNC;

[0077] S52: The transmission delay of a data stream on the TSN side consists of two parts: the transmission delay of the data stream on the first TSN side and the transmission delay on the second TSN side. The sum of the transmission delay on the first TSN side and the transmission delay on the second TSN side is the transmission delay on the TSN side. SDNC calculates the transmission delay of each data stream on the TSN side based on the transmission delay of each data stream on the TSN1 side and the transmission delay on the second TSN side.

[0078] The transmission delay of each data stream on the TSN side The calculation formula is shown in (19):

[0079]

[0080] Furthermore, in step S6, the transmission delay of each data stream on the 5G side is The calculation formula is shown in (20):

[0081]

[0082] Further, in step S7, the mapping table pre-configured by the SDNC includes a 5QI value, which includes a resource type, priority, and PDB, etc., wherein the resource type is divided into a guaranteed bit rate (GBR) resource type, a non-guaranteed bit rate (Non-GBR) resource type, and a delay critical GBR (Delay critical GBR) resource type. The 5G-TSN network only includes GBR and delay critical GBR resource types, and GBR represents a guaranteed bit rate; the priority is used to distinguish data streams of the same UE, indicating the order of scheduling resources between data streams. The lowest priority level value corresponds to the highest priority, so that a data stream with a priority value of N takes precedence over a data stream with a higher priority value (i.e., N+1, N+2, etc.); PDB represents a packet budget delay, which refers to the pre-configured delay between the UE and the UPF / NW-TT;

[0083] Calculating the packet delay budget, 5QI, and priority of each data flow on the 5G side includes the following steps:

[0084] S71: The priority value is set by itself, and the PDB value is set by itself. The pre-configured mapping table on the TSN AF includes three parameters: 5QI, resource type, and PDB. The standardized 5QI and QoS feature mapping table is optimized to a 5QI and QoS feature mapping table adapted to the 5G-TSN network.

[0085] S72: The SMF retrieves the DS-TT to UE dwell time by establishing a PDU session between the UE and the UPF, and provides the DS-TT to the UE dwell time to the TSN AF through the PCF, which then sends it to the SDNC.

[0086] According to step S6, SDNC calculates the transmission delay of the data stream on the 5G side It is known that the transmission delay of data flow on the 5G side is the residence delay from DS-TT to UE Pre-configured delay PDB between UE and UPF / NW-TT k Composition; each data flow pre-configured delay PDB between UE and UPF / NW-TT k The calculation formula is shown in (21):

[0087]

[0088] S73: SDNC calculates the PDB of each data flow using formula (21) k , according to the packet delay budget column of the 5QI and QoS feature mapping table adapted to the 5G-TSN network, find the 5QI value corresponding to each data flow;

[0089] S74: SDNC formulates data flow priority mapping rules based on the end-to-end priority and the transmission latency of the data flow on the 5G side. According to the standard, the smaller the 5G side priority value, the higher the priority.

[0090] 1) The sum of the delays of data stream transmission on the 5G side is Sumdelay, as shown in formula (22):

[0091]

[0092] 2) The ratio of the transmission delay of each data stream on the 5G side k It is equal to the transmission delay of the data stream on the 5G side divided by the sum of the transmission delays of the data stream on the 5G side, as shown in formula (23):

[0093]

[0094] 3) The priority factor r of each data stream transmitted in 5G k It is equal to the end-to-end priority of the data flow multiplied by the proportion of the transmission delay of the data flow on the 5G side, as shown in formula (24):

[0095] r k =priority k ×Ratio k (twenty four)

[0096] SDNC calculates the priority factor of each data stream transmitted on the 5G side through formulas (22), (23), and (24) and sorts the priority factors from small to large. If the priority factors of the data streams are the same, the end-to-end delay of the data streams is compared based on the priority factors and sorted from small to large. The priority of the data stream transmitted on the 5G side is assigned according to the sorting of the priority factors of each data stream. The data stream with the smallest priority factor has a 5G side transmission priority of 1. The 5G side transmission priority of the data streams corresponding to the 5G side transmission priority is 2, 3, ..., N respectively.

[0097] S75: SDNC sends the 5QI, priority, and PDB parameters corresponding to each data flow to TSN AF. TSN AF creates a TSC auxiliary container with the branch 5QI, priority, and PDB information and provides it to PCF. PCF forwards it to SMF as part of the PCC rule. SMF derives TSC auxiliary information from the TSC auxiliary container and configures it to (R)AN. (R)AN prepares communication resources for the data flow in advance to ensure that communication resources do not need to be requested through scheduling when TSN data frames arrive.

[0098] Further, step S8 specifically includes the following steps:

[0099] S81: SDNC abstracts the channel into a two-dimensional resource matrix consisting of time resources and frequency resources and initializes all time and frequency resources of the channel to idle. The channel scheduling period is 1ms. The basic scheduling unit is the physical resource block (PRB). The time domain resource corresponding to each PRB is a time slot; the frequency domain resource corresponding to each PRB is a resource block (RB). The number of time slots in the scheduling period is determined by the subcarrier spacing (SCS).

[0100] Data frames are the basic data units in 5G transmission. Considering data frames as items, time-frequency resources as boxes, with time-domain resources abstracted as the box's height and frequency-domain resources abstracted as the box's width, the resource allocation problem can be transformed into a two-dimensional bin packing problem: pack all rectangular items into the box while minimizing the box's height.

[0101] S82: The set of data frames to be transmitted in the channel is P, and there are h data frames in the set. The data frames are sorted from low to high according to the priority value of each data frame and numbered as p1, p2, ..., p h , that is, the set of data frames is P = {p1, p2, ..., p h}, where the priority relationship of data frames in obtaining time-frequency resources is that data frame p1 has the highest priority, followed by data frame p2, and data frame p h Minimum; The two-dimensional packing problem can be expressed as a linear programming problem. In the two-dimensional coordinates formed by time and frequency (x h ,y h) indicates the position of the data frame; when the data frame enters the channel, the time domain and frequency domain resources of the data frame correspond one-to-one with the time and resource block number coordinates;

[0102] 1) Delay time for all data frame transmission to complete UE-RAN is the maximum value of the sum of the start time of the data frame on the x-axis and the transmission time on the channel, as shown in formula (25):

[0103]

[0104] 2) The number of ending resource blocks of data frame h on the y-axis Equal to the starting resource block number y of data frame h on the y-axis h The number of resource blocks n occupied by data frame h on the channel h The sum of , as shown in formula (26):

[0105]

[0106] 3) The number of resource blocks N that can be allocated to a channel with a channel bandwidth of BW RB It is equal to the channel bandwidth BW divided by 12 times the subcarrier spacing SCS, as shown in formula (27):

[0107]

[0108] 4) The number of resource blocks at the end of the data frame P on the y-axis should not be greater than the number of resource blocks N that can be allocated to the channel RB , as shown in formula (28):

[0109]

[0110] 5) Different data frames cannot overlap in two-dimensional space, that is, if two or more data frames occupy the same time, they cannot occupy the same resource block; similarly, if two or more data frames occupy the same resource block, they cannot occupy the same time; assuming that data frame p h The data frame p that has entered the channel and then enters the channel j The occupied space range must satisfy formula (29):

[0111]

[0112] S83: When a rectangular object arrives, the 5G resource allocation algorithm based on data stream priority finds the rectangle with the highest priority from the set of rectangles, creates the first layer of the box, and places the rectangle with the highest priority among the remaining rectangles in this layer. If there is no space on the current layer to place a rectangle, the next layer is created, and the height of the layer is the height of the first rectangle placed. This cycle continues until all rectangles are placed. The 5G resource allocation algorithm based on data stream priority follows the following rules:

[0113] Rule 1: Different data frames cannot overlap in two-dimensional space. That is, if two or more data frames occupy the same time, they cannot occupy the same resource block. Similarly, if two or more data frames occupy the same resource block, they cannot occupy the same time.

[0114] Rule 2: If the data frame p h The number of ending resource blocks on the y-axis is not greater than the number of resource blocks N that can be allocated to the channel. RB , this layer can store data frame p h ; If the data frame p h The number of ending resource blocks on the y-axis is greater than the number of resource blocks N that can be allocated to the channel. RB , then open up the next layer to store the data frame p h ;

[0115] Rule 3: In a 5G network, a data frame can only exist in one form during transmission. The original form of the data frame during transmission is known. When allocating resources to the data frame, it can be transformed between various states as needed. That is, the length and height of the data frame (a rectangular object) can change, but the area remains unchanged.

[0116] S84: SDNC calculates the start and end positions of the time slots and resource blocks occupied by each data frame and sends the data frame information to TSN AF. TSN AF sends the start and end position information of the time slots and resource blocks occupied by each data frame to (R)AN through PCF and SMF. (R)AN allocates corresponding resource blocks and time slots to the data frames in the channel based on the data frame information, thereby realizing resource allocation on the 5G side.

[0117] The beneficial effects of the present invention are as follows: the present invention takes the network resources of 5G-TSN as the research object, studies the resource allocation requirements of data streams on the TSN side, combines the end-to-end transmission requirements of data streams and the transmission requirements on the TSN side in different scenarios, studies the transmission requirements of data streams on the 5G side, and then obtains the resource allocation requirements of data streams on the 5G side, effectively allocates resources in 5G-TSN, thereby ensuring low-latency and high-reliability transmission of data streams in 5G-TSN.

[0118] Other advantages, objects, and features of the present invention will be described in part in the following description and, in part, will be apparent to those skilled in the art upon examination of the following description or may be learned from practice of the present invention. The objects and other advantages of the present invention may be realized and obtained through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0119] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention will be described in detail below with reference to the accompanying drawings, in which:

[0120] Figure 1 This is a diagram of the network resource allocation architecture for 5G-TSN;

[0121] Figure 2 A schematic diagram of a network resource allocation method for 5G-TSN;

[0122] Figure 3 Flowchart of the network resource allocation algorithm for 5G-TSN;

[0123] Figure 4 This is the flow chart of network resource allocation on the TSN side;

[0124] Figure 5 It is a schematic diagram of the two-dimensional resource matrix. DETAILED DESCRIPTION

[0125] The following describes the embodiments of the present invention by means of specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present invention, and the following embodiments and features in the embodiments can be combined with each other without conflict.

[0126] Among them, the accompanying drawings are only for illustrative purposes and represent only schematic diagrams rather than actual pictures, and should not be understood as limiting the present invention. In order to better illustrate the embodiments of the present invention, some parts of the accompanying drawings may be omitted, enlarged or reduced, and do not represent the dimensions of actual products. For those skilled in the art, it is understandable that some well-known structures and their descriptions may be omitted in the accompanying drawings.

[0127] The same or similar numbers in the drawings of the embodiments of the present invention correspond to the same or similar parts; in the description of the present invention, it should be understood that if there are terms such as "upper", "lower", "left", "right", "front", "back", etc. indicating directions or positional relationships, they are based on the directions or positional relationships shown in the drawings. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific direction, be constructed and operate in a specific direction. Therefore, the terms describing the positional relationship in the drawings are only used for illustrative purposes and cannot be understood as limiting the present invention. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to specific circumstances.

[0128] See also Figures 1 to 5 This paper takes the network resources of 5G-TSN as the research object, studies the resource allocation requirements of data flow on the TSN side, combines the end-to-end transmission requirements of data flow and the transmission requirements of the TSN side in different scenarios, studies the transmission requirements of data flow on the 5G side, and then obtains the resource allocation requirements of data flow on the 5G side.

[0129] 1. Network resource allocation architecture for 5G-TSN

[0130] With the continuous integration and development of industrialization and informatization, in actual factory networks, the network that carries data stream transmission urgently needs to be composed of different networks. In such heterogeneous networks, the premise for network data transmission and deterministic scheduling is to clearly allocate network resources. Figure 1 As shown, the present invention proposes a network resource allocation architecture for 5G-TSN, which consists of a software-defined network centralized controller (SDNC), a centralized network configurator (CNC), 5G and TSN.

[0131] 5G includes user equipment (UE), radio access network (R)AN) and 5G core network, among which the 5G core network consists of unified data management (UDM), network exposure function (NEF), access and mobility management function (AMF), session management function (SMF), policy control function (PCF), time-sensitive networking application function (TSN AF), user plane function (UPF), device-side TSN translator (DS-TT) and network-side TSN translator (NW-TT).

[0132] TSN includes TSN1 and TSN2, where TSN1 consists of TSN1 terminals and TSN1 switches, and TSN2 consists of TSN2 terminals and TSN2 switches.

[0133] The architecture is divided into a control plane and a data plane, where SDNC, CNC, UDM, NEF, AMF, SMF, PCF, and TSN AF are located on the control plane, and TSN terminals, TSN switches, DS-TT, UE, (R)AN, UPF, and NW-TT are located on the data plane.

[0134] On the control plane, SDNC centrally manages the entire network through CNC and TSN AF, discovers and updates 5G and TSN (hereinafter referred to as: 5G-TSN) topology, and formulates 5G-TSN resource allocation strategy.

[0135] CNC is mainly responsible for discovering the TSN physical topology, retrieving TSN switches, and implementing functions such as scheduling traffic; interacting with TSN AF to obtain 5GS bridge information; dividing the time slots of TSN switches and sending the calculation results to the TSN switches.

[0136] TSN AF is mainly responsible for interacting with CNC to realize the interaction and transmission of TSN flow transmission direction, flow cycle, transmission delay budget, data priority and other parameters with 5G.

[0137] On the data plane, 5G acts as a logical bridge within TSN (hereafter referred to as the 5GS Bridge). The 5GS Bridge consists of the following components: DS-TT, the data plane tunnel between the UE and UPF, and the NW-TT. The ports on the NW-TT support connections to TSN, and the ports on the DS-TT side are associated with PDU sessions, thus providing 5G connectivity to TSN.

[0138] 2. Network resource allocation method for 5G-TSN

[0139] 2.1 Overview of Network Resource Allocation Methods

[0140] When 5G-TSN is integrated, how to allocate 5G-TSN network resources is one of the key factors to ensure real-time and reliable network transmission. How to allocate 5G-TSN network resources mainly depends on how to divide the latency requirements on the 5G side and the TSN side. The latency requirement of the data stream on the 5G side is obtained by the end-to-end latency requirement of the data stream and the latency requirement of the data stream on the TSN side. The schematic diagram of the network resource allocation method for 5G-TSN of the present invention is shown in the figure below. Figure 2 shown.

[0141] Therefore, the present invention combines the transmission priority, latency requirements and other information of the data stream, based on the 802.1Qcc protocol and the new network element functions proposed in 3GPP R17, to analyze the allocation of 5G-side resources and TSN-side resources when different industrial data streams are transmitted in 5G-TSN.

[0142] First, based on 5G-TSN cross-network clock synchronization, CNC1 and CNC2 calculate and allocate resources for TSN1 and TSN2, optimizing the transmission latency of data streams on TSN1 and TSN2 to minimize the latency on TSN1 and TSN2. Second, SDNC calculates the maximum transmission latency of data streams on 5G based on the end-to-end latency parameters of different data streams in industrial application scenarios and the maximum transmission latency on TSN obtained in the previous step, thereby clarifying the latency requirements for data streams on 5G. Finally, SDNC allocates resources for 5G and optimizes the transmission latency of data streams on 5G.

[0143] The execution steps of the network resource allocation algorithm for 5G-TSN are as follows:

[0144] Step 1: SDNC obtains network topology and data flow information of the entire network through CNC1, CNC2, and TSN AF and manages the information;

[0145] Step 2: SDNC calculates the priority of data flow transmission in TSN based on the end-to-end priority input by the user;

[0146] Step 3: CNC1 and CNC2 allocate network resources on the TSN side for data flows;

[0147] Step 4: SDNC calculates the transmission delay of each data stream on the TSN side;

[0148] Step 5: SDNC calculates the transmission latency of each data stream on the 5G side based on the end-to-end latency of each data stream and the transmission latency on the TSN side;

[0149] Step 6: SDNC pre-configures a mapping table and calculates the packet delay budget, 5QI, and priority of each data flow on the 5G side;

[0150] Step 7: SDNC allocates network resources on the 5G side for data flows.

[0151] 2.2 Network Resource Allocation Method

[0152] The network resource allocation method for 5G-TSN specifically includes the following steps:

[0153] Step 1: SDNC obtains network topology and data flow information of the entire network through CNC1, CNC2, and TSN AF and manages the information.

[0154] 1) CNC1 and CNC2 discover the TSN network topology based on the Link Layer Discovery Protocol (LLDP) and forward the TSN network topology information to SDNC.

[0155] 2) The terminal device sends information such as the sending period, data frame size, priority, and delay requirements of each data stream to the centralized user configurator through the user configuration protocol. The centralized user configurator sends the data stream information to CNC1 and CNC2 through the user network interface. CNC1 and CNC2 forward the data stream information to SDNC.

[0156] 3) SDNC organizes the collected network topology and data flow information. Network topology information includes information about links within TSN1, links from TSN1 to 5G, links within 5G, links from 5G to TSN2, and links within TSN2. Data flow information includes end-to-end latency requirements and data flow priority. The network topology consists of a TSN terminal (transmitter) connected to a TSN switch, entering the 5G network via DS-TT, then connecting to another TSN switch via NW-TT, and finally connecting to another TSN terminal (receiver).

[0157] The explanation of the parameters involved in the calculation process is shown in Table 1.

[0158] Table 1 Explanation of symbols

[0159]

[0160]

[0161] Step 2: SDNC calculates the priority of data flow transmission in TSN based on the end-to-end priority input by the user.

[0162] 1) In a network consisting of 5G and TSN, there are k data streams, and different data streams have different priorities, as shown in Table 2. The end-to-end priority input by the design user ranges from 1 to 15, with a smaller value indicating a higher priority.

[0163] Table 2 End-to-end priority of data flows

[0164] <![CDATA[Data stream (f k )]]> <![CDATA[End-to-end priority (priority k )]]> <![CDATA[f1]]> <![CDATA[priority1]]> <![CDATA[f2]]> <![CDATA[priority2]]> <![CDATA[f3]]> <![CDATA[priority3]]> <![CDATA[f4]]> <![CDATA[priority4]]> <![CDATA[f5]]> <![CDATA[priority5]]> … … <![CDATA[f k ]]> <![CDATA[priority k ]]>

[0165] 2) SDNC uses the input end-to-end priority and the mapping rules in Table 3 to obtain the TSN side priority (Table 4). The TSN side priority has eight levels according to the standard, ranging from 0 to 7. The larger the value, the higher the priority.

[0166] Table 3 Mapping rules between end-to-end priority and TSN side priority

[0167]

[0168]

[0169] Table 4 TSN side priority of data flow

[0170]

[0171] Step 3: CNC1 and CNC2 allocate network resources for data flow on the TSN side.

[0172] This invention patent proposes a progressive approximation (iterative) resource allocation algorithm based on Satisfiability Modulo Theories (SMT). The time slot planning problem of time-sensitive network data flow is converted into a constraint satisfaction problem. Based on the characteristics of SMT, the transmission time slot and the gate cycle waiting quantity are used as variables, the network status, data flow information, etc. are used as constants, and the delay requirements of the data flow are converted into constraints. Finally, the SMT solver is used to solve the constraint problem to obtain the result of the time slot planning. The process of network resource allocation on the TSN side is as follows: Figure 4 shown.

[0173] The specific design steps of the iterative resource allocation algorithm based on Satisfiability Modulo Theory (SMT) are as follows:

[0174] 1) CNC1 and CNC2 obtain the information of all data flows in this resource allocation, including the data flow transmission cycle, data frame size, priority, TSN switch transmission rate, TSN switch processing delay, etc., and use it as the input of the algorithm. The transmission delay of each data flow on the TSN1 side is The demand calculation formula is shown in (1), where n = 33%:

[0175]

[0176] The transmission delay of each data stream on the TSN2 side The demand calculation formula is shown in (2), where n = 33%:

[0177]

[0178] 2) CNC1 and CNC2 add constants and variables to the SMT solver. Constants include the size and priority of each flow's data frame, the transmission period, the TSN switch's transmission rate, and the TSN switch's processing latency. This information is primarily used to model the network and data flows. Variables include the TSN switch's gating period, the number of transmission slots per flow within the gating period, the start time of each data flow transmission slot, the arrival and departure times of data flows in each slot, the transmission latency of data flows on the TSN switch, and the transmission latency requirement. This information needs to be solved based on the constraints and is the output of the planning algorithm. The specific constants, variables, and their symbolic representations are shown in Table 5.

[0179] Table 5 Symbolic explanation of constants and variables

[0180]

[0181] 3) CNC1 and CNC2 add constraints to the SMT solver to constrain the value range of the variables so that they meet the limitations of network resources and the transmission quality requirements of the data flow. The specific constraints are as follows:

[0182] 3.1) The TSN switch gating cycle Schedcycle is the lowest common multiple of the transmission cycles of all data flows flowing through the TSN switch and needs to satisfy formula (3):

[0183]

[0184] 3.2) The number of transmission time slots of a data stream within the TSN switch gating cycle is equal to the TSN switch gating cycle divided by the data stream sending cycle, as shown in formula (4):

[0185]

[0186] 3.3) The start time of the transmission time slot of the data flow in the TSN switch gating and the transmission delay must satisfy the relationship as shown in Formula (5) and Formula (6). The transmission delay is equal to the time required for the data volume transmitted in one transmission cycle of the data flow (one data frame is sent in one transmission cycle) to be sent from the TSN switch; the start time of the time slot must be within the gating cycle.

[0187]

[0188] 0≤Sstart k,i ≤Schedcycle-SendDelay k,i (5)

[0189]

[0190] 3.4) Within the gating cycle of the TSN switch, the interval between different transmission time slots of the same data stream must be greater than the sending cycle of the data stream, as shown in formula (7):

[0191]

[0192] Sstart k,i +T k ≤Sstart k,i+1 (7)

[0193] 3.5) The transmission time slots within the gating cycle of the TSN switch cannot overlap, that is, the start time of a transmission time slot of data stream k must be greater than or equal to the end time of a transmission time slot of data stream g, or the end time must be less than or equal to the start time of a transmission time slot of data stream g, as shown in formula (8):

[0194]

[0195] (Sstart k,i +SendDelay k,i ≤Sstart g,j ) || (Sstart g,j +SendDelay g,j ≤ Sstart k,i (8)

[0196] 3.6) Based on the characteristics of the transmission time slot, the arrival and departure times of the data frames of each data stream at different transmission time slots can be obtained. The arrival time of the data frame at the TSN switch is earlier than the departure time. The relationship is shown in formulas (9), (10), and (11):

[0197]

[0198]

[0199] Tleave k,i =Sstart k,i +SendDelay k,i (10)

[0200] Tarrival k,i <Tleave k,i (11)

[0201] 3.7) The end-to-end delay of the same data stream in different time slots must be less than the upper limit of the delay, as shown in formula (12):

[0202]

[0203]

[0204] 4) CNC1 and CNC2 run the SMT solver. The solver will try all possible values ​​of the variables in the solution space in turn. If a feasible solution is found, the transmission delay requirement of each data stream will be iterated. The calculation formulas are shown in (13), (14), and (15). After the calculation, return to step 2 of step 3).

[0205] n=n-1% (13)

[0206]

[0207]

[0208] If no feasible solution is found after traversing all the values, the transmission delay requirement of each data stream is iterated again, and the calculation formulas are shown in (16), (17), and (18). After the calculation, return to step 2 of step 3. If a feasible solution is found, the time slot allocation result is output.

[0209] n=n+0.1% (16)

[0210]

[0211]

[0212] 5) CNC1 and CNC2 encapsulate the calculation results into an XML file and use the NETCONF protocol to configure the XML-based gating schedule to the TSN switch.

[0213] Step 4: SDNC calculates the transmission delay of each data stream on the TSN side.

[0214] 1) CNC1 obtains the transmission delay of each data stream on the TSN1 side through step 3 CNC2 obtains the transmission delay of each data stream on the TSN2 side through step 3. CNC1 and CNC2 forward the delay information to SDNC.

[0215] 2) The data stream transmission latency on the TSN side consists of two parts: the transmission latency on TSN1 and the transmission latency on TSN2. The sum of the transmission latency on TSN1 and TSN2 is the transmission latency on TSN. SDNC calculates the transmission latency of each data stream on TSN based on the transmission latency on TSN1 and TSN2.

[0216] The transmission delay of each data stream on the TSN side The calculation formula is shown in (19):

[0217]

[0218] Step 5: SDNC calculates the transmission latency of each data stream on the 5G side based on the end-to-end latency of each data stream and the transmission latency on the TSN side.

[0219] CNC1 and CNC2 calculate the data stream's transmission latency over TSN1 and TSN2, and then superimpose the latency from TSN1 and TSN2 on both sides to 5G to obtain the end-to-end latency of the data stream. Given the end-to-end latency of different data streams in 5G-TSN convergence, the latency of the data stream during TSN1 and TSN2 transmission is subtracted step by step to obtain the data stream's latency during 5G transmission.

[0220] The transmission delay of each data stream on the 5G side The calculation formula is shown in (20):

[0221]

[0222] Step 6: SDNC pre-configures a mapping table and calculates the packet delay budget, 5QI, and priority of each data flow on the 5G side.

[0223] The SDNC preconfigured mapping table includes 5QI values ​​(resource type, priority, and packet delay budget). Resource types are categorized as guaranteed bit rate (GBR), non-guaranteed bit rate (Non-GBR), and delay-critical GBR. 5G-TSN networks only include GBR and delay-critical GBR resource types. Priority is used to differentiate data flows from the same UE, indicating the order in which resources are scheduled. The lowest priority level corresponds to the highest priority, so that a data flow with a priority value of N takes precedence over data flows with higher priority values ​​(i.e., N+1, N+2, etc.). Packet delay budget (PDB) refers to the preconfigured delay between the UE and the UPF / NW-TT. Table 6 shows the standardized mapping of 5QI and QoS characteristics.

[0224] Table 6 Standardized 5QI and QoS feature mapping table

[0225]

[0226] 1) According to 3GPP 23.501, the known priority can adopt the default priority in the standardized 5QI, or the priority value can be set by itself according to the actual needs of the data flow, and the default priority is replaced by the set value; the PDB can adopt the static PDB in the standardized 5QI, or the PDB can be set by itself according to the actual needs of the data flow, and the static value is replaced by the set value, but the self-set PDB is not greater than the static PDB in the standardized 5QI. The priority designed by the present invention adopts the self-set priority value, and the PDB adopts the self-set PDB value. The pre-configured mapping table on the TSN AF includes three parameters: 5QI, resource type, and PDB. The standardized 5QI and QoS feature mapping table is optimized to a 5QI and QoS feature mapping table adapted to the 5G-TSN network, as shown in Table 7.

[0227] 5QI and QoS feature mapping table adapted to 5G-TSN network

[0228]

[0229] 2) SMF retrieves the DS-TT to UE residence time by establishing a PDU session between UE and UPF, and provides the DS-TT to UE residence time to TSN AF through PCF, which then sends it to SDNC. According to step 5, SDNC calculates the transmission delay of the data flow on the 5G side. It is known that the transmission delay of data flow on the 5G side is the residence delay from DS-TT to UE Pre-configured delay PDB between UE and UPF / NW-TT kThe pre-configured delay PDB for each data flow between UE and UPF / NW-TT k The calculation formula is shown in (21):

[0230]

[0231] 3) SDNC calculates the PDB of each data flow using formula (21) k ,According to the packet delay budget column in Table 7, find the 5QI value corresponding to each data stream.

[0232] 4) SDNC formulates data flow priority mapping rules based on the end-to-end priority and the transmission delay of the data flow on the 5G side. According to the standard, the smaller the value of the 5G side priority, the higher the priority.

[0233] 4.1) The sum of the delays of data streams transmitted on the 5G side is Sumdelay, as shown in formula (22):

[0234]

[0235] 4.2) The ratio of the transmission delay of each data stream on the 5G side k It is equal to the transmission delay of the data stream on the 5G side divided by the sum of the transmission delays of the data stream on the 5G side, as shown in formula (23):

[0236]

[0237] 4.3) Priority factor r for each data stream transmitted in 5G k It is equal to the end-to-end priority of the data flow multiplied by the proportion of the transmission delay of the data flow on the 5G side, as shown in formula (24):

[0238] r k =priority k ×Ratio k (twenty four)

[0239] SDNC calculates the priority factor of each data stream transmitted on the 5G side using formulas (22), (23), and (24) and sorts the priority factors from small to large. If the priority factors of the data streams are the same, the end-to-end delay of the data streams is compared based on the priority factors and sorted from small to large. The priority of the data stream transmitted on the 5G side is assigned based on the sorting of the priority factors of each data stream. The data stream with the smallest priority factor corresponds to a 5G side transmission priority of 1. The 5G side transmission priority of the data streams corresponding to the 5G side transmission priority is accumulated by 1, and the 5G side transmission priority of the data streams corresponds to 2, 3, ..., N respectively.

[0240] SDNC sends the 5QI, priority, and PDB parameters corresponding to each data flow to TSN AF. TSN AF creates a TSC auxiliary container with the branch 5QI, priority, and PDB information and provides it to PCF. PCF forwards it to SMF as part of the PCC rule. SMF derives TSC auxiliary information from the TSC auxiliary container and configures it to (R)AN. (R)AN prepares communication resources for the data flow in advance to ensure that communication resources do not need to be requested through scheduling when TSN data frames arrive.

[0241] 5) SDNC sends the 5QI, priority, and PDB parameters corresponding to each data flow to TSN AF. TSN AF creates a TSC auxiliary container with the 5QI, priority, and PDB information and provides it to PCF. PCF forwards it to SMF as part of the PCC rule. SMF derives TSC auxiliary information from the TSC auxiliary container and configures it to (R)AN. (R)AN prepares communication resources for the data flow in advance to ensure that communication resources do not need to be requested through scheduling when TSN data frames arrive.

[0242] Step 7: SDNC allocates network resources on the 5G side for data flows.

[0243] SDNC allocates network resources on the 5G side by allocating time domain resources and frequency resources to each data frame to be transmitted, so that the data frame can be transmitted reliably in real time on specific network resources. The present invention proposes a 5G resource allocation algorithm based on data stream priority to allocate time domain resources and frequency domain resources on the channel, and abstract the time domain resources and frequency domain resources into a two-dimensional space, which is similar to the two-dimensional packing problem. When the UE or (R)AN needs to send data, resource allocation is performed in advance at the MAC layer to determine the size and position of the time and frequency resources occupied by each data block in the channel, and the transmission delay of the data stream in the 5G network is optimized through the 5G resource allocation algorithm based on data stream priority. The specific symbol explanation is shown in Table 8.

[0244] Table 8 Symbol explanation on the 5G side

[0245]

[0246] 1) SDNC abstracts the channel into a two-dimensional resource matrix consisting of time and frequency resources and initializes all time and frequency resources of the channel to idle. The channel scheduling period is 1ms, and the basic scheduling unit is the physical resource block (PRB). Each PRB corresponds to a time slot, and each PRB corresponds to a frequency block (RB). The number of time slots in the scheduling period is determined by the subcarrier spacing (SCS). The correspondence between the subcarrier spacing and the number of time slots is shown in Table 9.

[0247] Table 9 Correspondence between subcarrier spacing and number of time slots

[0248] SCS(kHz) Number of slots Slot length (ms) 15 1 1 30 2 0.5 60 4 0.25 120 8 0.125 240 16 0.0625

[0249] The scenario targeted by the present invention is the uRLLC scenario, which requires low transmission delay. This requirement can be met by configuring a larger subcarrier spacing and reducing the length of the corresponding time slot. Therefore, the SCS of the present invention is selected as 240kHz, the corresponding number of slots is 16, and the slot length is 0.0625ms. The two-dimensional resource matrix composed of time domain resources (time) and frequency domain resources (resource blocks) is as follows Figure 5 shown.

[0250] Data frames are the basic data units in 5G transmission. Considering data frames as objects, time-frequency resources as boxes, time-domain resources as the height of the box, and frequency-domain resources as the width of the box, the resource allocation problem can be transformed into a two-dimensional bin packing problem: pack all rectangular objects into the box while minimizing the occupied box height.

[0251] 2) The set of data frames to be transmitted in the channel is P, and there are h data frames in the set. The data frames are sorted from low to high according to the priority value of each data frame and numbered as p1, p2, ..., p h , that is, the set of data frames is P = {p1, p2, ..., p h}, where the priority relationship of data frames in obtaining time-frequency resources is that data frame p1 has the highest priority, followed by data frame p2, and data frame p h The two-dimensional packing problem can be expressed as a linear programming problem. In the two-dimensional coordinates formed by time and frequency (x h ,y h ) indicates the location of the data frame. When the data frame enters the channel, the time domain and frequency domain resources of the data frame correspond one-to-one with the time and resource block number coordinates.

[0252] 2.1) Delay: the time it takes to complete the transmission of all data frames UE-RAN is the maximum value of the sum of the start time of the data frame on the x-axis and the transmission time on the channel, as shown in formula (25):

[0253]

[0254] 2.2) The number of ending resource blocks of data frame h on the y-axis Equal to the starting resource block number y of data frame i on the y-axis h The number of resource blocks n occupied by data frame h on the channel h The sum of , as shown in formula (26):

[0255]

[0256] 2.3) The number of resource blocks N that can be allocated to a channel with a channel bandwidth of BW RB It is equal to the channel bandwidth BW divided by 12 times the subcarrier spacing SCS, as shown in formula (27):

[0257]

[0258] 2.4) The number of resource blocks at the end of the data frame P on the y-axis should not be greater than the number of resource blocks N that can be allocated to the channel RB , as shown in formula (28):

[0259]

[0260] 2.5) Different data frames cannot overlap in two-dimensional space. That is, if two or more data frames occupy the same time, they cannot occupy the same resource block. Similarly, if two or more data frames occupy the same resource block, they cannot occupy the same time. Assume that data frame p h The data block p that has entered the channel and then enters the channel j The occupied space range must satisfy formula (29):

[0261]

[0262] 3) When a rectangular object arrives, the 5G resource allocation algorithm based on data stream priority finds the rectangle with the highest priority from the set of rectangles, creates the first layer of the box, and places the rectangle with the highest priority among the remaining rectangles on this layer. If there is no space on the current layer to place a rectangle, the next layer is created, with a height equal to the height of the first rectangle placed. This cycle repeats until all rectangles are placed. The 5G resource allocation algorithm based on data stream priority follows the following rules:

[0263] Rule 1: Different data frames cannot overlap in two-dimensional space. That is, if two or more data frames occupy the same time, they cannot occupy the same resource block. Similarly, if two or more data frames occupy the same resource block, they cannot occupy the same time.

[0264] Rule 2: If the data frame p h The number of ending resource blocks on the y-axis is not greater than the number of resource blocks N that can be allocated to the channel. RB , this layer can store data frame p h ; If the data frame p h The number of ending resource blocks on the y-axis is greater than the number of resource blocks N that can be allocated to the channel. RB , then open up the next layer to store the data frame p h .

[0265] Rule 3: In a 5G network, a data frame can only exist in one form during transmission. The original form of the data frame during transmission is known. When resources are allocated to the data frame, it can be transformed between various states as needed. That is, the length and height of the data frame (rectangular object) can change, but the area remains unchanged.

[0266] 4) The SDNC calculates the start and end positions of the time slots and resource blocks occupied by each data frame and sends this information to the TSN AF. The TSN AF then sends this information to the (R)AN via the PCF and SMF. The (R)AN uses this information to allocate the appropriate resource blocks and time slots to the data frames in the channel, thus implementing 5G-side resource allocation.

[0267] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not limiting. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention can be modified or replaced by equivalents without departing from the purpose and scope of the technical solutions, which should all be included in the scope of the claims of the present invention.

Claims

1. A network resource allocation method for 5G-TSN, characterized in that: The method specifically comprises the following steps: S1: Build a network architecture for 5G-TSN network resource allocation, including one SDNC, two CNCs, one 5G, and two TSNs. SDNC stands for Software Defined Network Centralized Controller, and CNC stands for Centralized Network Configurator. S2: The SDNC obtains and manages the network topology and data flow information of the entire network through the CNC and TSN AF. The network topology information includes the link information in the first TSN, the link information from the first TSN to 5G, the link information in 5G, the link information from 5G to the second TSN, and the link information in the second TSN. The data flow information includes end-to-end latency requirements and data flow priority. The network topology consists of a TSN terminal as the transmitter connected to a TSN switch through DS-TT to enter the 5G network, then connected to another TSN switch through NW-TT, and finally connected to another TSN terminal as the receiver. S3: SDNC calculates the priority of data stream transmission in TSN based on the end-to-end priority input by the user; S4: The CNC uses an iterative resource allocation algorithm based on SMT to allocate network resources for data flows on the TSN side. SMT stands for Satisfiability Modulo Theory. The SMT-based iterative resource allocation algorithm specifically transforms the TSN data flow slot planning problem into a constraint satisfaction problem. Based on the characteristics of SMT, the transmission slot and gating period are used as variables, the network status and data flow information are used as constants, and the data flow latency requirements are converted into constraints. Finally, the SMT solver is used to solve the constraints to obtain the slot planning results. S5: SDNC calculates the transmission delay of each data stream on the TSN side; S6: SDNC calculates the transmission delay of each data stream on the 5G side based on the end-to-end delay of each data stream and the transmission delay on the TSN side; S7: The SDNC pre-configures a mapping table, including 5QI values, which include resource type, priority, and PDB. PDB stands for packet delay budget, which refers to the pre-configured delay between the UE and the UPF / NW-TT. The SDNC also calculates the packet delay budget, 5QI, and priority of each data flow on the 5G side. The calculation of the packet delay budget, 5QI, and priority of each data flow on the 5G side includes the following steps: S71: The priority value is set by itself, and the PDB value is set by itself. The pre-configured mapping table on the TSN AF includes three parameters: 5QI, resource type, and PDB. The standardized 5QI and QoS feature mapping table is optimized to a 5QI and QoS feature mapping table adapted to the 5G-TSN network. S72: The SMF retrieves the DS-TT to UE dwell time by establishing a PDU session between the UE and the UPF, and provides the DS-TT to the UE dwell time to the TSN AF through the PCF, which then sends it to the SDNC. According to step S6, SDNC calculates the transmission delay of the data stream on the 5G side ; It is known that the transmission delay of data flow on the 5G side is the residence delay from DS-TT to UE , pre-configured delay between UE and UPF / NW-TT Composition; pre-configured delay of each data flow between UE and UPF / NW-TT The calculation formula is shown in (1): (1) S73: SDNC calculates the value of each data flow through formula (1) , according to the packet delay budget column of the 5QI and QoS feature mapping table adapted to the 5G-TSN network, find the 5QI value corresponding to each data flow; S74: SDNC formulates data flow priority mapping rules based on the end-to-end priority and the transmission latency of the data flow on the 5G side. According to the standard, the smaller the 5G side priority value, the higher the priority. 1) The sum of the transmission delays of data streams on the 5G side , as shown in formula (2): (2) 2) The proportion of transmission delay of each data stream on the 5G side It is equal to the transmission delay of the data stream on the 5G side divided by the sum of the transmission delays of the data stream on the 5G side, as shown in formula (3): (3) 3) Priority factor for each data stream transmitted in 5G It is equal to the end-to-end priority of the data flow multiplied by the proportion of the transmission delay of the data flow on the 5G side, as shown in formula (4): (4) SDNC calculates the priority factor of each data stream transmitted on the 5G side through formulas (2), (3), and (4) and sorts the priority factors from small to large. If the priority factors of the data streams are the same, the end-to-end delay of the data streams is compared based on the priority factors and sorted from small to large. The priority of the data stream transmitted on the 5G side is assigned according to the sorting of the priority factors of each data stream. The data stream with the smallest priority factor has a 5G side transmission priority of 1, and the priority factors are accumulated by 1 in sequence. The 5G side transmission priorities of the data streams correspond to N; S75: The SDNC sends the 5QI, priority, and PDB parameters corresponding to each data flow to the TSN AF. The TSN AF creates a TSC auxiliary container with the branch 5QI, priority, and PDB information and provides it to the PCF. The PCF forwards it to the SMF as part of the PCC rule. The SMF derives TSC auxiliary information from the TSC auxiliary container and configures it to the (R)AN. The (R)AN prepares communication resources for the data flow in advance to ensure that communication resources do not need to be requested through scheduling when the TSN data frame arrives. S8: SDNC allocates network resources based on the transmission priority of data flows on the 5G side.

2. The network resource allocation method according to claim 1, wherein: In step S1, The 5G includes UE, (R)AN and 5G core network, where UE represents user equipment and (R)AN represents radio access network; the 5G core network includes UDM, NEF, AMF, SMF, PCF, TSN AF, UPF, DS-TT and NW-TT, where UDM represents unified data management, NEF represents network exposure function, AMF represents access and mobility management function, SMF represents session management function, PCF represents policy control function, TSN AF represents time-sensitive network application function, UPF represents user plane function, DS-TT represents device side translator, and NW-TT represents network side translator; The TSN includes a TSN terminal and a TSN switch; The network architecture is divided into the control plane and the data plane, of which SDNC, CNC, UDM, NEF, AMF, SMF, PCF, and TSN AF are located in the control plane, and TSN terminals, TSN switches, DS-TT, UE, (R)AN, UPF, and NW-TT are located in the data plane; On the control plane, SDNC centrally manages the entire network through CNC and TSN AF, discovers and updates 5G and TSN topologies (referred to as 5G-TSN topologies), and formulates 5G-TSN resource allocation strategies. CNC is responsible for discovering the TSN physical topology, retrieving TSN switches, and scheduling traffic; Interact with TSN AF to obtain 5GS bridge information; divide the time slots of TSN switches and send the calculation results to TSN switches; The TSN AF is responsible for interacting with the CNC to implement the interaction and transmission of TSN flow transmission direction, flow cycle, transmission delay budget, data priority and 5G; On the data plane, 5G acts as a logical bridge in TSN, referred to as the 5GS bridge. The 5GS bridge includes the DS-TT, the data plane tunnel between the UE and the UPF, and the NW-TT. The ports on the NW-TT support connection with TSN, and the ports on the DS-TT side are associated with the PDU session, thereby providing 5G connectivity with TSN.

3. The network resource allocation method according to claim 2, wherein: Step S2 specifically includes the following steps: S21: The two CNCs discover the TSN network topology based on LLDP and forward the TSN network topology information to the SDNC. LLDP stands for Link Discovery Protocol. S22: The terminal device sends the transmission period, data frame size, priority, and latency requirements for each data stream to the centralized user configurator through the user configuration protocol. The centralized user configurator sends the data stream information to the two CNCs through the user network interface. The two CNCs forward the data stream information to the SDNC. S23: SDNC organizes the collected network topology information and data flow information.

4. The network resource allocation method according to claim 3, wherein: Step S3 specifically includes the following steps: S31: In a network composed of 5G and TSN, there are k data streams, different data streams have different priorities, that is, data streams The corresponding end-to-end priority is The end-to-end priority input by the user is designed to range from 1 to 15, with smaller values ​​giving higher priorities. S32: SDNC obtains the priority of the TSN side based on the input end-to-end priority and the mapping rules between the established end-to-end priority and the TSN side priority, i.e. the data flow The corresponding TSN side priority is ,Among them, there are 8 priorities on the TSN side according to the standard,,i.e., 0 to 7. The larger the value, the higher the priority; The mapping rule between the end-to-end priority and the TSN side priority is: the end-to-end priority is 1 to 15, and the corresponding TSN side priority is 7, 6, 6, 5, 5, 4, 4, 3, 3, 2, 2, 1, 1, 0, 0.

5. The network resource allocation method according to claim 4, wherein: In step S4, the SMT-based iterative resource allocation algorithm specifically includes the following steps: S41: The two CNCs obtain information about all data flows in this resource allocation, including the data flow transmission cycle, data frame size, priority, TSN switch transmission rate, TSN switch processing delay, etc., and use it as the input of the algorithm; the transmission delay of each data flow on the first TSN side is The demand calculation formula is shown in (5), where : (5) The transmission delay of each data stream on the second TSN side The demand calculation formula is shown in (6), where : (6) in, Represents a collection of data streams, Indicates the k End-to-end delay requirements for each data stream; S42: Two CNCs add constants and variables to the SMT solver. The constants include the size and priority of each flow's data frame, the sending period, the sending rate of the TSN switch, and the processing delay of the TSN switch, which are used to model the network and data flow. The variables include the gating period of the TSN switch, the number of transmission time slots of each flow within the gating period, the start time of each transmission time slot of the data flow, the arrival time and departure time of the data flow in each time slot, the sending delay of the data flow on the TSN switch, and the transmission delay requirement. These variables need to be solved according to the constraints and are the output of the planning algorithm. S43: The two CNCs add constraints to the SMT solver to constrain the value range of the variables so that they meet the network resource limitations and data flow transmission quality requirements. The specific constraints are as follows: 1) TSN switch gating cycle is the lowest common multiple of the transmission periods of all data flows passing through the TSN switch, and needs to satisfy formula (7): (7) in, Represents the set of data flows flowing through the TSN switch, Representing data flow cycle; 2) Data Flow k Number of transmission time slots within a TSN switch gating cycle It is equal to the TSN switch gating period divided by the data flow sending period, as shown in formula (8): (8) 3) The start time of the transmission time slot of the data flow in the TSN switch gating and the sending delay need to satisfy the relationship as shown in formula (9) and formula (10); the sending delay is equal to the time required for the amount of data transmitted in one sending cycle of the data flow to be sent from the TSN switch; the start time of the time slot must be within the gating cycle; (9) (10) in, Representing data flow In the TSN switch gating cycle The start time of a transmission time slot, Representing data flow In the TSN switch gating cycle The transmission delay of each transmission slot; Representing data flow The size of the data frame, Indicates the sending rate of the TSN switch; 4) Within the gating cycle of the TSN switch, the interval between different transmission time slots of the same data stream must be greater than the sending cycle of the data stream, as shown in formula (11): (11) 5) The transmission time slots within the gating cycle of the TSN switch cannot overlap, that is, the data flow k The start time of a transmission slot must be greater than or equal to the data flow The end time of a transmission time slot or the end time is less than or equal to the data flow The starting time of a transmission time slot is as shown in formula (12): (12) 6) According to the characteristics of the transmission time slot, the arrival and departure time of the data frame of each data stream at different transmission time slots of the TSN switch are obtained; the arrival time of the data frame to the TSN switch must be earlier than the departure time, and the relationship is shown in formula (13), formula (14), and formula (15): (13) (14) (15) in, Representing data flow No. The moment when the data of a transmission time slot arrives at the TSN switch, Representing data flow No. The moment when the data of a transmission time slot leaves the TSN switch, Representing data flow At the sending time of the sender, Indicates the processing delay of the TSN switch; 7) The end-to-end delay of different time slots of the same data stream must be less than the upper limit of the delay, as shown in formula (16): (16) S44: The two CNCs run the SMT solver. The solver tries all possible values ​​of the variables in the solution space in turn. If a feasible solution is found, the transmission delay requirement of each data stream is iterated. The calculation formulas are shown in (17), (18), and (19). After the calculation, the process returns to step S42. (17) (18) (19) If no feasible solution is found after traversing all the values, the transmission delay requirement of each data stream is iterated again, and the calculation formulas are shown in (20), (21), and (22). After the calculation, return to step S42; if a feasible solution is found, the time slot division result is output; (20) (21) (22) The two CNCs encapsulate the calculation results into XML files and use the NETCONF protocol to configure the XML-based gating scheduling table to the TSN switch.

6. The network resource allocation method according to claim 5, characterized in that: Step S5 specifically includes the following steps: S51: The first CNC obtains the transmission delay of each data stream on the first TSN side through step S4. The second CNC obtains the delay of each data stream when it is transmitted on the second TSN side through step S4 ;The two CNCs forward the delay information to the SDNC; S52: The transmission delay of a data stream on the TSN side consists of two parts: the transmission delay of the data stream on the first TSN side and the transmission delay on the second TSN side. The sum of the transmission delay on the first TSN side and the transmission delay on the second TSN side is the transmission delay on the TSN side. SDNC calculates the transmission delay of each data stream on the TSN side based on the transmission delay of each data stream on the TSN1 side and the transmission delay on the second TSN side. The transmission delay of each data stream on the TSN side The calculation formula is shown in (23): (23)。 7. The network resource allocation method according to claim 6, characterized in that: In step S6, the transmission delay of each data stream on the 5G side The calculation formula is shown in (24): (24)。 8. The network resource allocation method according to claim 7, characterized in that: In step S7, the resource types are divided into guaranteed bit rate resource types, non-guaranteed bit rate resource types, and delay-critical GBR resource types. The 5G-TSN network only includes GBR and delay-critical GBR resource types. GBR means guaranteed bit rate; the priority is used to distinguish data streams of the same UE, indicating the order of scheduling resources between data streams, and the lowest priority level value corresponds to the highest priority.

9. The network resource allocation method according to claim 8, characterized in that: Step S8 specifically includes the following steps: S81: SDNC abstracts the channel into a two-dimensional resource matrix consisting of time resources and frequency resources and initializes all time and frequency resources of the channel to idle. The channel scheduling period is 1ms. The basic scheduling unit is the physical resource block (PRB). The time domain resource corresponding to each PRB is a time slot, and the frequency domain resource corresponding to each PRB is a resource block (RB). The number of time slots in the scheduling period is determined by the subcarrier spacing (SCS). Data frames are the basic data units in 5G transmission. Considering data frames as items, time-frequency resources as boxes, the time domain resources are abstracted as the box's height, and the frequency domain resources are abstracted as the box's width, the resource allocation problem is transformed into a two-dimensional bin packing problem: pack all rectangular items into the box while minimizing the box's height. S82: The set of data frames to be transmitted in the channel is , there are Data frames are sorted from low to high according to the priority value of each data frame and numbered as follows: , that is, the set of data frames is , where the priority relationship of data frames obtaining time-frequency resources is data frames Highest priority, data frame Second, data frame Minimum; The two-dimensional packing problem is expressed as a linear programming problem, in the two-dimensional coordinates formed by time and frequency ( , ) indicates the location of the data frame; when the data frame enters the channel, the time domain and frequency domain resources of the data frame correspond one-to-one with the time and resource block number coordinates; 1) Time when all data frames are transmitted For data frame The maximum value of the sum of the start time of the axis and the transmission time on the channel is shown in formula (25): (25) 2) Data Frame exist The number of ending resource blocks for the axis Equal to data frame exist The starting resource block number of the axis With data frame The number of resource blocks occupied on the channel The sum of , as shown in formula (26): (26) 3) Channel bandwidth is The number of resource blocks that can be allocated to the channel Equal to the channel bandwidth Divide by 12 times the subcarrier spacing , as shown in formula (27): (27) 4) Data Frame exist The number of resource blocks that terminate the axis should not be greater than the number of resource blocks that can be allocated to the channel. , as shown in formula (28): (28) 5) Different data frames cannot overlap in two-dimensional space, that is, if two or more data frames occupy the same time, they cannot occupy the same resource block; similarly, if two or more data frames occupy the same resource block, they cannot occupy the same time; assuming that the data frame Data frames that have already entered the channel and enter the channel later The occupied space range must satisfy formula (29): （29） S83: When a rectangular object arrives, the 5G resource allocation algorithm based on data stream priority finds the rectangle with the highest priority from the set of rectangles, creates the first layer of the box, and places the rectangle with the highest priority among the remaining rectangles in this layer. If there is no space on the current layer to place a rectangle, the next layer is created, and the height of the layer is the height of the first rectangle placed. This cycle continues until all rectangles are placed. The 5G resource allocation algorithm based on data stream priority follows the following rules: Rule 1: Different data frames cannot overlap in two-dimensional space. That is, if two or more data frames occupy the same time, they cannot occupy the same resource block. Similarly, if two or more data frames occupy the same resource block, they cannot occupy the same time. Rule 2: If the data frame exist The number of resource blocks that terminate the axis is not greater than the number of resource blocks that can be allocated to the channel. , this layer can store data frames ; If the data frame exist The number of resource blocks that can be allocated to the channel is greater than the number of resource blocks that can be allocated to the channel. , then open up the next layer to store data frames ; Rule 3: In a 5G network, a data frame can only exist in one form during transmission. The original form of the data frame during transmission is known. When resources are allocated to the data frame, it can be transformed between various states as needed. That is, the length and height of the data frame can change, but the area remains unchanged. S84: SDNC calculates the start and end positions of the time slots and resource blocks occupied by each data frame and sends the data frame information to TSN AF. TSN AF sends the start and end position information of the time slots and resource blocks occupied by each data frame to (R)AN through PCF and SMF. (R)AN allocates corresponding resource blocks and time slots to the data frames in the channel based on the data frame information, thereby realizing resource allocation on the 5G side.

Images