A time-sensitive network scheduling method and device based on network calculus

Abstract

This invention discloses a time-sensitive network scheduling method and apparatus based on network calculus, relating to the field of time-sensitive network communication technology. The method includes: constructing a centralized control architecture for TSN based on SDN; sending service information and TSN switch GCL configuration information stored in the switch database module to the topology discovery module for statistical analysis, and forwarding them to the GCL scheduling calculation module; calculating the upper bound of switch latency using a constructed arrival-service curve model based on network calculus; constructing a cost function by comparing the upper bound of latency with the acquired user service flow latency QoS requirements; analyzing the schedulability of the current GCL configuration information based on the cost function and obtaining the analysis results; optimizing the current GCL configuration information using a designed incremental PID control algorithm, outputting the optimal switch GCL configuration information, and sending it to the network configuration module for configuration of the switches, thus completing service scheduling decisions and achieving dynamic adjustment and efficient operation of the network.

Description

Technical Field

[0001] This invention relates to the field of time-sensitive network communication technology, and in particular to a time-sensitive network scheduling method and apparatus based on network computation. Background Technology

[0002] Time-Sensitive Networking (TSN) has deterministic latency guarantee and multi-service carrying capability, solving the problem of data transmission in the same network in the Industrial Internet, and has become a research hotspot in industrial field networks. However, the standard TSN protocol only defines the data forwarding and processing methods, and does not regulate the networking of TSN in industrial environments, resulting in the following problems: (1) Service flows are blocked in the queue due to resource conflicts; (2) For services that require strict real-time performance, end-to-end latency and jitter are uncontrollable; (3) The configuration workload of hundreds or thousands of service flows in the network is difficult to complete manually and is prone to errors; moreover, any configuration error of a switch will cause the performance degradation of the entire service flow, and troubleshooting is extremely difficult; (4) Poor network expansion performance, unable to adapt to changes in services. Therefore, in complex industrial network environments, establishing an end-to-end latency analysis model for time-sensitive networks (TSNs) with multi-service QoS constraints such as latency, jitter, packet loss rate, and rate, and quantitatively analyzing the latency performance of the network under extreme conditions, is a current challenge for TSN applications.

[0003] For TSN (Time-Sensitive Networking), its primary task is to guarantee the end-to-end latency and jitter requirements of time-sensitive traffic flows. Therefore, the boundedness of end-to-end latency is the primary characteristic of TSN transmission determinism. Network calculus, a deterministic queuing theory based on minimum addition algebra, has been widely adopted in recent years to characterize network performance boundaries, including characterizing the upper bound of end-to-end latency for individual network nodes or the entire network. Current methods utilize the inherent characteristics of Ethernet AVB services to analyze the worst-case end-to-end latency of services under non-frame preemption strategies. Other existing methods analyze the factors contributing to network congestion under credit-based shapers and asynchronous traffic shapers, and evaluate the latency performance of network nodes at different rates.

[0004] Current latency upper bound analysis based on network calculus in TSN has several limitations. First, traffic characterization typically assumes simultaneous arrival of packets at nodes, leading to significant approximation errors in arrival curve modeling. Second, service curve modeling usually constructs service models independently for each switch, neglecting the connectivity between nodes during traffic transmission, thus limiting the analysis of overall network performance. Furthermore, the application of network calculus in TSN is mainly limited to its use as a tool for evaluating QoS guarantee capabilities through gating settings, guiding the design of traffic shapers. Therefore, improving the traditional arrival-service curve modeling method using network calculus to obtain a tight end-to-end latency upper bound, and designing an autonomous gating adjustment mechanism accordingly to provide deterministic service guarantees for time-triggered flows, is of significant research value. Summary of the Invention

[0005] To address the problems in existing technologies, such as the exponential expansion of the solution space for scheduling problems as network size and traffic volume increase, making it difficult to obtain scheduling solutions with high resource utilization and resulting in severe waste of network resources, and the difficulty in effectively adapting to dynamic changes in network conditions due to the fact that TSN scheduling is usually based on offline or static assumptions, this invention provides a time-sensitive network scheduling method and apparatus based on network calculus. The technical solution is as follows:

[0006] On the one hand, a time-sensitive network scheduling method based on network calculus is provided. This method is implemented by a time-sensitive network scheduling device based on network calculus, and includes:

[0007] S1. The switch database module sends the stored service information and TSN switch GCL configuration information to the topology discovery module.

[0008] S2. The topology discovery module collects business information and TSN switch queue gating status parameters in the current industrial automation and intelligent manufacturing time-sensitive network, and forwards them to the GCL scheduling calculation module.

[0009] The S3 and GCL scheduling calculation modules calculate the upper bound of switch latency using a network calculus-based arrival-service curve model. They then compare this upper bound with the acquired user service flow latency QoS requirements and construct a cost function. Based on this cost function, they analyze the schedulability of the current switch GCL configuration information and obtain the analysis results. Based on these results, they use a designed incremental PID control algorithm to optimize the current switch GCL configuration information, outputting the optimal switch GCL configuration information and sending it to the network configuration module.

[0010] S4. The network configuration module configures the optimal switch GCL configuration information to the switch to complete the service scheduling decision.

[0011] On the other hand, a time-sensitive network scheduling apparatus based on network computation is provided. This apparatus is applied to a time-sensitive network scheduling method based on network computation. The apparatus includes:

[0012] The switch database module is used to store the collected service information and TSN switch GCL configuration information;

[0013] The topology discovery module is used to collect statistics on service information and TSN switch queue gating status parameters in the current industrial automation and intelligent manufacturing time-sensitive network.

[0014] The GCL scheduling calculation module is used to calculate the upper bound of switch latency by constructing an arrival-service curve model based on network calculus; compare the upper bound of latency with the obtained user service flow latency QoS requirements, and construct a cost function; analyze the schedulability of the current switch GCL configuration information based on the cost function, and obtain the analysis results; based on the analysis results, optimize the current switch GCL configuration information using a designed incremental PID control algorithm, output the optimal switch GCL configuration information, and send it to the network configuration module.

[0015] The network configuration module is used to configure the optimal switch GCL configuration information to the switch to complete the service scheduling decision.

[0016] On the other hand, a time-sensitive network scheduling device based on network computation is provided, the time-sensitive network scheduling device based on network computation includes: a processor; a memory, the memory storing computer-readable instructions, which, when executed by the processor, implement any of the methods in the above-described time-sensitive network scheduling method based on network computation.

[0017] On the other hand, a computer-readable storage medium is provided, wherein at least one instruction is stored in the storage medium, the at least one instruction being loaded and executed by a processor to implement any of the above-described time-sensitive network scheduling methods based on network computation.

[0018] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:

[0019] This invention utilizes an improved model based on arrival-service curves from traditional network calculus. For multi-node network applications, it enhances the accuracy of traditional network calculus in latency upper bound analysis by precisely characterizing arrival and service curves. Based on the latency estimation from the improved model, it compares it with a preset latency QoS threshold to determine the schedulability of service flows. Using this as the scheduling optimization objective, an IPO algorithm is designed. Under constraints on key parameters, window opening time, and window length, the time window is optimized to maximize network schedulability, thereby obtaining the optimal GCL configuration scheme.

[0020] This invention constructs a centralized TSN control architecture based on SDN, utilizing a centralized network controller and a centralized user configuration unit to achieve unified management of the entire TSN domain. The controller uses collected service flow characteristics, including period, packet length, priority, and end-to-end latency requirements, to perform scheduling planning based on a time window scheduling problem analysis model. Under window parameter constraints, an IPO optimization algorithm is used to adaptively adjust the scheduling results, generating the final GCL configuration for the switch network. The configuration is then distributed to the switch system through a standardized southbound interface, enabling dynamic network adjustment and efficient operation. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 This is a flowchart of a time-sensitive network scheduling method based on network calculus provided in an embodiment of the present invention;

[0023] Figure 2 This is a schematic diagram of the structure of a centralized TSN controller and functional modules based on SDN provided in an embodiment of the present invention;

[0024] Figure 3 This is a block diagram of GCL optimization for an IPO algorithm provided in an embodiment of the present invention;

[0025] Figure 4 This is a flowchart illustrating the implementation of a TSN GCL scheduling mechanism based on SDN provided in an embodiment of the present invention.

[0026] Figure 5 This is a diagram illustrating the target optimization effect of an IPO algorithm provided in an embodiment of the present invention;

[0027] Figure 6This is a block diagram of a time-sensitive network scheduling device based on network calculus provided in an embodiment of the present invention;

[0028] Figure 7 This is a schematic diagram of the structure of a time-sensitive network scheduling device based on network computation provided in an embodiment of the present invention. Detailed Implementation

[0029] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0030] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0031] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0032] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0033] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0034] This invention provides a time-sensitive network scheduling method based on network calculus. This method can be implemented by a time-sensitive network scheduling device based on network calculus, which can be a terminal or a server. Figure 1 The flowchart shown is for a time-sensitive network scheduling method based on network calculus. The processing flow of this method may include the following steps:

[0035] S1. The switch database module sends the stored service information and TSN switch GCL configuration information to the topology discovery module.

[0036] Among them, GCL (Group Control List) configuration information.

[0037] One feasible implementation method is, for example Figure 2 The diagram illustrates a centralized TSN controller and functional module structure based on SDN, as provided in an embodiment of the present invention. This embodiment employs a fully centralized control architecture model, constructing the TSN control architecture based on the software-defined network (SDN) concept. SDN provides efficient resource scheduling capabilities by centrally managing the network through a global view of the network. Under the premise of constant synchronization across the entire network, the fully centralized control architecture model delegates user requirements to a centralized user configuration manager. This centralized user configuration is centrally collected and uniformly converted to enhance QoS assurance capabilities. After conversion, the configuration is sent to the centralized network controller via the northbound interface. The centralized network controller, possessing network functions such as calculating topology paths, performs optimal calculations based on the global network view and then sends updated routing paths and GCL configuration information to the TSN switching network via the southbound interface. In one feasible implementation, a switch database module is used to store parameters of all switches in the current time-sensitive network for industrial automation and intelligent manufacturing, including switch port bandwidth and GCL (Global Chaining). A topology discovery module is used to collect service information in the network, including packet length, its generation time at the sending end, and transmission period, and forward it to the GCL scheduling calculation module. The GCL scheduling calculation module is used to determine whether network configuration needs to be adjusted based on the proposed traffic scheduling method, using the acquired switch information and latency information as input parameters. By adjusting the GCL configuration information, the newly calculated GCL configuration information is sent to the network configuration module to further configure the GCL of the switches and update the information in the switch database. The network configuration module is used to configure the GCL configuration information that meets the requirements to the switches, and finally complete the service scheduling decision.

[0038] S2. The topology discovery module collects business information and TSN switch queue gating status parameters in the current time-sensitive network of industrial automation and intelligent manufacturing, and forwards them to the GCL scheduling calculation module.

[0039] The S3 and GCL scheduling calculation modules calculate the upper bound of switch latency using a network calculus-based arrival-service curve model. They then compare this upper bound with the acquired user service flow latency QoS requirements and construct a cost function. Based on this cost function, they analyze the schedulability of the current switch GCL configuration information and obtain the analysis results. Based on these results, they use a designed incremental PID control algorithm to optimize the current switch GCL configuration information, outputting the optimal switch GCL configuration information, which is then sent to the network configuration module.

[0040] In one feasible implementation, the delay upper bound analysis in the GCL scheduling calculation module adopts a basic network calculus algebra method, namely Total Traffic Analysis (TFA), which calculates the switch delay upper bound hop-by-hop along the path and sums them to obtain the end-to-end delay. The arrival curve and service curve, as two basic analytical tools of network calculus, are important bases for system evaluation.

[0041] Optionally, S3's network calculus-based arrival-service curve model calculates the upper bound of switch latency, including:

[0042] S31. Based on periodic time-triggered flows, a ladder function is used to construct an arrival curve representation of the target flow; based on the arrival curve representation of the target flow, flow groups in the same queue are aggregated to construct an aggregated arrival curve representation.

[0043] The process of constructing the arrival curve of the target flow using a step function is represented by the following formula (1):

[0044] (1)

[0045] in, f Represents the target flow rate; in the step function, Indicates the length of the data frame; This represents the service flow sending cycle; it takes into account the data frame injection offset. So that it is between [0, Between [], no earlier than the start of transmission, and limited to the range of the stream cycle.

[0046] In one feasible implementation, traditional network calculus, when describing the arrival behavior of aggregated flows on a node's egress queue, sums the arrival curves of individual flows. This calculation rule assumes that all data packets can enter the same buffer queue simultaneously. In reality, different service flows cannot arrive at the same switching node simultaneously. Therefore, for service flows arriving at the same switch, their data packets are serialized to ensure that the queuing process occurs sequentially frame by frame, rather than simultaneously. For a flow group G converging in the same queue, its aggregated arrival curve is defined as denoted as... The aggregation arrival curve is represented by the following formula (2):

[0047] (2)

[0048] in, Indicates the maximum packet length in the shared queue input stream; Indicates the total number of service flows within the flow group; This represents the arrival curve of business flow i; The length of the data frame in service flow i is represented by C; the link bandwidth is represented by C; and the time is represented by t.

[0049] S32. Set the window service time slot length, calculate the maximum waiting time representation of the access switch and the parameter representation of the GCL period; based on the window service time slot length, the maximum waiting time representation of the access switch and the parameter representation of the GCL period, calculate the service curve representation of the switch.

[0050] To prevent transmission conflicts and ensure that the port / link is idle when data is sent, a guard band is set at the end of the time window, the length of which is the transmission time required for the maximum packet length (1500B). Due to the effect of the guard band, if the time window is too small, it will be difficult to carry the transmission needs of the data packets. The service slot length of window j meets the condition expressed by the following formula (3):

[0051] (3)

[0052] in, Indicates the window service slot length; Indicates the length of the k-time window of the switch; Indicates the length of the protective strip; C represents the minimum packet length; C represents the link bandwidth.

[0053] Among them, the maximum waiting time is set according to the position of the switching node in the link; (1) When the node is an access switch (AS), that is, directly connected to the service source, its maximum waiting time is determined only by its own queue status. The maximum waiting time of the access switch is expressed by the following formula (4):

[0054] (4)

[0055] in, Indicates the maximum waiting time of the access switch; Indicates the access switch gating period; Indicates the window service timeslot length of the access switch;

[0056] (2) For non-access switches (nAS), data is received from the previous switch (node ​​number k-1). Therefore, the maximum waiting time is determined by the time window offset of adjacent nodes based on the node connection relationship. This mechanism establishes the coupling relationship between multiple nodes in the topology network. The time interval for data frames to arrive at the node queue is defined with window j as a reference. The earliest and latest times are represented by the following formula (5):

[0057] (5)

[0058] in, This indicates the earliest time a data frame arrives at switch window j (k-th window). This indicates the opening time of window j of switch k-1; This represents the minimum transmission delay of switch k-1; Indicates a time delay constant. To delay the transmission time, To handle latency; This indicates the latest time a data frame arrives at switch window j.

[0059] The maximum waiting time of the access switch is further calculated by calculating the earliest and latest times, and is expressed by the following formula (6):

[0060] (6)

[0061] in, This represents the maximum waiting time for window j of switch k; Indicates the GCL period; Indicates a service slot; This indicates the closing time of window j in switch k.

[0062] In one feasible implementation, the service curve of switch k is represented by the following formula (7):

[0063] (7)

[0064] in, The service curve of switch k is represented by the following: GCL period represents the least common multiple of the periods of all service flows passing through the switch; This is a classic service model that follows the TDMA protocol; This indicates the maximum waiting time of the access switch.

[0065] S33. Combine the service curve representation of the switch and the aggregated arrival curve representation to construct an arrival-service curve model based on network calculus; analyze the upper bound of the latency based on the arrival-service curve model based on network calculus to obtain the maximum latency experienced by the service flow in the switch, which is expressed by the following formula (8):

[0066] (8)

[0067] in, This indicates the maximum latency experienced by the service flow in the switch; The arrival curve of business flow i is represented; The service curve of switch k is represented by ; i represents the service flow sequence number; and k represents the switch sequence number.

[0068] In one feasible implementation, the service flow typically passes through multiple switch nodes during end-to-end transmission, thus requiring consideration of the flow arrival status at each node. Between adjacent switches, the output arrival curve of the previous hop can serve as the input arrival curve of the next hop node, describing the arrival behavior of the service flow and initiating a new round of delay upper bound assessment. According to the network calculus output arrival curve theorem, the output arrival curve of the service flow after passing through the switch is represented by the following formula (9):

[0069] (9)

[0070] in, The output arrival curve represents the traffic flow after passing through the switch. This indicates the output arrival curve of the previous hop; This represents the arrival curve of the input for the next hop; This represents the impulse delay function, used to shift the network computation curve to the right as a whole. This indicates the upper bound of the queuing delay, which is... With transmission delay The difference, that is ;if , Otherwise it is .

[0071] Optionally, the upper bound of latency is compared with the obtained latency QoS requirements of user service flows, and a cost function is constructed; based on the cost function, the schedulability of the current switch GCL configuration information is analyzed to obtain the analysis results, including:

[0072] Based on network calculus-based latency upper bound analysis, the end-to-end total latency of the service flow from the input port to the output port is calculated. Under pre-set latency QoS constraints, boundary conditions for scheduling optimization are constructed, which are expressed by the following formula (10):

[0073] (10)

[0074] in, This indicates the upper bound of the end-to-end queuing delay; Indicates the number of switches traversed along path i of the service flow; Indicates the link propagation delay; Indicates transmission delay; Indicates the end-to-end transmission deadline for service flow i;

[0075] In one feasible implementation, when a service flow passes through a transmission node on the link and enters the switch through an input port, processing delay and queuing delay are generated inside the switch; when data is sent to the link through an output port, transmission delay is generated; in addition, the propagation of the service flow on the link also generates propagation delay. Therefore, the total end-to-end delay can be expressed by the following formula (11):

[0076] (11)

[0077] in, Indicates the total end-to-end latency of service flow i; Indicates processing delay; This indicates a delay in queuing; Indicates transmission delay; This indicates the link propagation delay.

[0078] Based on the boundary conditions for scheduling optimization, the schedulability of the current business flow is evaluated, and the evaluation results of schedulability are obtained.

[0079] Among them, the QoS requirements for user service flow latency are preset in advance based on the attributes of the service flow.

[0080] This involves comparing the upper bound of the latency with the pre-set QoS requirements for user service flow latency, and using this numerical comparison to determine... The Boolean value is used for evaluation. Specifically, if the upper limit of latency is greater than the QoS value of the user service flow latency, then... =1 indicates that the current business flow i has failed to be scheduled; otherwise... =0. Passed. Construct the system cost function.

[0081] The system cost function cost(s) is used to determine the schedulability of the service flow. This indicates whether flow i is schedulable in the network; if it is schedulable, the value is 0, otherwise it is 1. This can be expressed by the following formula (12):

[0082] (12)

[0083] The cost function for determining the schedulability of a service flow is constructed and expressed by the following formula (13):

[0084] (13)

[0085] in, Indicates the total number of business flows; This indicates whether flow i is schedulable in the network; the value is 0 if it is schedulable, and 1 otherwise.

[0086] Based on the schedulability evaluation results, the cost function of service flow schedulability is calculated. When the cost function is 0, the current switch GCL configuration information meets the scheduling requirements of all service flows, and the current switch GCL configuration information is directly output. When the cost function is not 0, the current switch GCL configuration information is optimized.

[0087] Optionally, based on the analysis results, S3 uses a designed incremental PID control algorithm to optimize the current switch GCL configuration information and outputs the optimal switch GCL configuration information, including:

[0088] Using queue time windows as the scheduling object and maximizing network schedulability as the objective, a multi-constraint traffic scheduling optimization problem model is constructed; wherein, the multi-constraint traffic scheduling optimization problem model uses the cost function as the minimization objective function;

[0089] In this embodiment of the invention, the goal is to maximize network schedulability. The queue time window is used as the scheduling object, and a Global Time Limit (GCL) is generated by optimizing the time slot allocation of the queue. This ensures that more service flows are successfully scheduled under limited available bandwidth conditions. This allows for precise planning of the time window used for service flow scheduling.

[0090] Optionally, the multi-condition constraints include: time window size constraint, time window offset constraint, and window opening time constraint;

[0091] Among these, compared to latency requirements, background traffic, due to its low priority, primarily focuses on throughput. Therefore, while ensuring stable throughput, it is necessary to effectively coordinate bandwidth allocation among different traffic flows. The time window size constraint is expressed by the following formula (14):

[0092] (14)

[0093] in, Indicates the scaling factor; C represents the link bandwidth; This indicates the minimum throughput requirement for the background flow;

[0094] The time window length must not be less than 10% of the GCL period, i.e. This is to avoid resource waste or failure to meet performance requirements due to improper time slot allocation. The scaling factor should be limited to the range specified by the following formula (15):

[0095] (15)

[0096] The window offset should cover all possible positions where a window may appear in the queue, thus ensuring that the window can receive data packets from window forwarding. The constraint range of the time window offset is expressed by the following formula (16):

[0097] (16)

[0098] in, This represents the offset between window j on queue m of switches a and b; This represents the length of window j on queue m of switch b; This indicates the gating period of queue a on switch; This indicates the gating period of queue b on switch;

[0099] The offset is defined as the difference at the moment the window opens, and is expressed by the following formula (17):

[0100] (17)

[0101] in, This represents the offset between windows on queues m of switches a and b. This indicates the window opening time of queue m in switch b; This represents the window opening time for queue m in switch a.

[0102] Among them, based on the time window offset constraint of adjacent switches, the variation range of queue gating timing is analyzed to determine the coupling relationship between multiple nodes.

[0103] The window opening time constraint is expressed by the following formula (18):

[0104] (18)

[0105] Where j represents the time window number; Indicates the gating period of switch b; Indicates the length of window j in switch b; This indicates the opening time of window j on switch b.

[0106] Based on the multi-constraint traffic scheduling optimization problem model, the designed incremental PID control algorithm is used to iteratively calculate the GCL window allocation that meets the constraints. By optimizing the cost function, the iterative calculation stops when the cost function converges to the minimum cost, and the optimal switch GCL configuration information corresponding to the minimum cost is output.

[0107] In this invention, a novel metaheuristic algorithm, namely the incremental PID (IPO) control algorithm, is designed. This algorithm starts with a random solution and optimizes the solution quality by continuously adjusting the system deviation, gradually guiding the entire population to converge to the optimal solution.

[0108] Optionally, the calculation process of the incremental PID control algorithm includes:

[0109] Initialize the population; calculate the adjusted output value according to the incremental PID control law;

[0110] In one feasible implementation, for an N-dimensional decision variable, its initial population can be represented by the following formula (19):

[0111] (19)

[0112] in, Represents an individual The d-th dimension; Indicates the upper bound of the variable; Indicates the lower bound of the variable; This represents the first random number between (0, 1).

[0113] Set a zero-output condition factor; based on the adjusted output value and the zero-output condition factor, iteratively update the population. When the iteration reaches time t+1, the updated population is obtained.

[0114] Alternatively, the adjusted output value can be expressed by the following formula (20):

[0115] (20)

[0116] in, The first PID parameter; This is the second PID parameter; This is the third PID parameter; Indicates iteration deviation; This represents a second random number between (0,1); This represents a third random number between (0, 1); This represents the fourth random number between (0,1); Indicates adjusting the output value;

[0117] Among them, iterative deviation The calculation process is expressed by the following formula (21):

[0118] (twenty one)

[0119] in, This represents the global optimal solution for the population at time t; unlike the traditional PID control framework, in the IPO algorithm, the user-defined value is... .definition However, the extreme values ​​after each iteration are not always the same, therefore To reduce the time complexity of the algorithm, It can be equivalently replaced by the following formula (22):

[0120] (twenty two)

[0121] in, This represents the overall bias of the previous iteration when the number of iterations is t; This represents the overall deviation in the (t-1)th iteration; This represents the best individual at iteration number t-1.

[0122] Among them, for different iteration numbers, the following relationship exists as shown in formula (23):

[0123] (twenty three)

[0124] in, This represents the total deviation in the t-th iteration; This represents the overall deviation between the first two iterations; This represents the overall deviation from the previous iteration.

[0125] To prevent the population from getting trapped in local optima and causing the algorithm to converge prematurely, a zero-output condition factor is added; the zero-output condition factor is expressed by the following formula (24):

[0126] (twenty four)

[0127] in, Indicates zero output factor; Indicates the upper bound of the algorithm iteration; Indicates the adjustment factor; Represents the fifth random number between (0,1); L represents the Levy flight function; This represents the overall deviation of the iteration.

[0128] In one feasible implementation, when the iteration reaches time t+1, the updated population is obtained based on the zero-output condition factor and the adjusted output value; wherein, the updated population is represented by the following formula (25):

[0129] (25)

[0130] in, Indicates the weighting coefficient. This represents the sixth random number between (0,1).

[0131] One feasible implementation method is, for example Figure 3 The diagram shown is a block diagram of GCL optimization using the IPO algorithm provided in an embodiment of the present invention. This embodiment utilizes the IPO algorithm to iteratively calculate the GCL window allocation that satisfies the constraints. The quality of each feasible solution is evaluated by using the schedulability of the service flow as a cost function. Finally, through cost function optimization, the optimal GCL synthesis scheme for the entire network is obtained.

[0132] S4. The network configuration module configures the optimal switch GCL configuration information to the switch to complete the service scheduling decision.

[0133] Among them, such as Figure 4 The diagram illustrates a flowchart of a TSN GCL scheduling mechanism based on SDN provided in this embodiment of the invention. Based on statistical traffic and switch GCL configuration information, the upper bound of latency is calculated using a network calculus arrival-service curve model, and compared with the latency QoS requirements of service flows to construct a cost function cost(s). When cost(s) = 0, it indicates that the current GCL can meet the scheduling needs of all service flows, and the current GCL can be directly returned as the output; otherwise, the IPO algorithm is invoked to adjust the GCL, gradually reducing the cost function until it converges to the minimum cost or reaches the preset iteration limit. The final output is the minimum cost and its corresponding optimal GCL network configuration result. The embodiments of the present invention can also be applied to: (1) industrial automation and industrial networks, specifically including: motion control processes, safety systems (i.e., safety-related I / O signals are transmitted through a guaranteed low-latency channel) and automation and monitoring; (2) automotive in-vehicle networks, specifically including: autonomous driving domain fusion (i.e., multi-sensor data are synchronously uploaded to the central computing unit within a strict time window), high-bandwidth infotainment and vehicle domain control (i.e., the issuance of key control commands from the backbone network); (3) aerospace airborne networks, specifically including: avionics system interconnection, cabin entertainment systems (i.e., centralized management and distribution of audio and video service content) and aircraft status monitoring and transmission.

[0134] In one feasible implementation, this invention uses Matlab software for simulation experiments. The bandwidth allocation scheme is determined using Matlab software, running on a computer equipped with a 3GHz Intel i9-13900K 24-core CPU and 128GB of memory, with Windows 10 operating system. The optimal solution for global network gating configuration under algorithm-driven analysis is analyzed. Table 1 shows the basic simulation parameter settings. The experimental network consists of 4 terminals and 9 switches, where the switches are divided into access nodes and core nodes. Traffic is sent from left to right. For the definition of service flows, the frame length and period of each flow are randomly selected from the set {500, 600, 800, 1000}, ranging from 50 to 1000 bytes. The number of input service flows in the experimental network is 600. In path selection, each service flow outputs from sending node 1 or 2, and a random path is assigned to receiving node 10 or 11. The experimental network can be a satellite network, a vehicle network, an avionics system communication network, or an industrial automation network, etc.

[0135] Table 1

[0136]

[0137] Table 2 shows the network parameter settings.

[0138] Table 2

[0139]

[0140] In this case, based on the parameter setting requirements of the IPO optimization algorithm, a PID control parameter of K is adopted. p = 1, K i = 0.5, K d =1.2. Additionally, the population size was set to 15, and the scheduling model was optimized through 20 iterations. Table 3 shows the parameter settings for the IPO algorithm.

[0141] Table 3

[0142]

[0143] Among them, such as Figure 5 The figure shows the optimization effect of the IPO algorithm provided in this embodiment of the invention. It can be seen from the figure that the objective function value gradually converges with the iteration process, and the minimum cost value obtained is 0, indicating that after optimization, all service flows have been successfully scheduled. Therefore, by deploying the GCL scheduling module in the centralized network controller, the GCL configuration of multiple switches in the network can be calculated.

[0144] To address the TAS mechanism widely used in industrial TSN scenarios, this invention proposes a TSN queue window scheduling mechanism based on an improved network calculus method. This mechanism analyzes the schedulability of service flows in multi-node networking scenarios based on latency upper bound assessment, and uses an IPO intelligent algorithm to iteratively optimize schedulability, thereby achieving optimal GCL configuration. This invention also establishes a centralized TSN control architecture based on SDN, which completes GCL control decisions and network configuration through the collaborative cooperation of functional modules.

[0145] This invention utilizes an improved model based on arrival-service curves from traditional network calculus. For multi-node network applications, it enhances the accuracy of traditional network calculus in latency upper bound analysis by precisely characterizing arrival and service curves. Based on the latency estimation from the improved model, it compares it with a preset latency QoS threshold to determine the schedulability of service flows. Using this as the scheduling optimization objective, an IPO algorithm is designed. Under constraints on key parameters, window opening time, and window length, the time window is optimized to maximize network schedulability, thereby obtaining the optimal GCL configuration scheme.

[0146] This invention constructs a centralized TSN control architecture based on SDN, utilizing a centralized network controller and a centralized user configuration unit to achieve unified management of the entire TSN domain. The controller uses collected service flow characteristics, including period, packet length, priority, and end-to-end latency requirements, to perform scheduling planning based on a time window scheduling problem analysis model. Under window parameter constraints, an IPO optimization algorithm is used to adaptively adjust the scheduling results, generating the final GCL configuration for the switch network. The configuration is then distributed to the switch system through a standardized southbound interface, enabling dynamic network adjustment and efficient operation.

[0147] Figure 6 This is a block diagram of a time-sensitive network scheduling device based on network calculus provided in an embodiment of the present invention. This device is used in a time-sensitive network scheduling method based on network calculus. (Refer to...) Figure 6 The device includes a switch database module 610, a topology discovery module 620, a GCL scheduling calculation module 630, and a network configuration module 640. Among them:

[0148] The switch database module 610 is used to store the collected service information and TSN switch GCL configuration information.

[0149] Topology discovery module 620 is used to collect service information and TSN switch queue gating status parameters in the current time-sensitive network of industrial automation and intelligent manufacturing.

[0150] The GCL scheduling calculation module 630 is used to calculate the upper bound of switch latency by constructing an arrival-service curve model based on network calculus; compare the upper bound of latency with the obtained user service flow latency QoS requirements and construct a cost function; analyze the schedulability of the current switch GCL configuration information based on the cost function and obtain the analysis results; based on the analysis results, optimize the current switch GCL configuration information using a designed incremental PID control algorithm, output the optimal switch GCL configuration information, and send it to the network configuration module.

[0151] The network configuration module 640 is used to configure the optimal switch GCL configuration information to the switch to complete the service scheduling decision.

[0152] Optionally, the arrival-service curve model based on network calculus, which calculates the upper bound of switch latency, includes:

[0153] Based on periodic time-triggered flows, a ladder function is used to construct an arrival curve representation of the target flow; based on the arrival curve representation of the target flow, flow groups in the same queue are aggregated to construct an aggregated arrival curve representation.

[0154] Set the window service slot length, calculate the maximum waiting time representation of the access switch and the parameter representation of the GCL period; based on the window service slot length, the maximum waiting time representation of the access switch and the parameter representation of the GCL period, calculate the service curve representation of the switch;

[0155] The service curve representation of the switch and the aggregated arrival curve representation are combined to construct an arrival-service curve model based on network calculus. The arrival-service curve model based on network calculus is used to analyze the upper bound of the delay to obtain the maximum delay experienced by the service flow in the switch, which is expressed by the following formula (1):

[0156] (1)

[0157] in, This indicates the maximum latency experienced by the service flow in the switch; The arrival curve of business flow i is represented; The service curve of switch k is represented by ; i represents the service flow sequence number; and k represents the switch sequence number.

[0158] Optionally, the step of comparing the upper bound of latency with the acquired user service flow latency QoS requirements and constructing a cost function; and analyzing the schedulability of the current switch GCL configuration information based on the cost function to obtain the analysis results, including:

[0159] Based on network calculus-based latency upper bound analysis, the end-to-end total latency of the service flow from the input port to the output port is calculated. Under pre-set latency QoS constraints, boundary conditions for scheduling optimization are constructed, which are expressed by the following formula (2):

[0160] (2)

[0161] in, This indicates the upper bound of the end-to-end queuing delay; Indicates the number of switches traversed along path i of the service flow; Indicates the link propagation delay; Indicates transmission delay; Indicates the end-to-end transmission deadline for service flow i;

[0162] Based on the boundary conditions for scheduling optimization, the schedulability of the current business flow is evaluated, and the evaluation results of schedulability are obtained.

[0163] The cost function for determining the schedulability of a business flow is constructed and expressed by the following formula (3):

[0164] (3)

[0165] in, Indicates the total number of business flows; This indicates whether flow i is schedulable in the network; the value is 0 if it is schedulable, and 1 otherwise.

[0166] Based on the schedulability evaluation results, the cost function of service flow schedulability is calculated. When the cost function is 0, the current switch GCL configuration information meets the pre-set scheduling requirements of all service flows, and the current switch GCL configuration information is directly output. When the cost function is not 0, the current switch GCL configuration information is optimized.

[0167] Optionally, based on the analysis results, the designed incremental PID control algorithm is used to optimize the current switch GCL configuration information, and the optimal switch GCL configuration information is output, including:

[0168] Using queue time windows as the scheduling object and maximizing network schedulability as the objective, a multi-constraint traffic scheduling optimization problem model is constructed; wherein, the multi-constraint traffic scheduling optimization problem model uses the cost function as the minimization objective function;

[0169] Based on the multi-constraint traffic scheduling optimization problem model, the designed incremental PID control algorithm is used to iteratively calculate the GCL window allocation that meets the constraints. By optimizing the cost function, the iterative calculation stops when the cost function converges to the minimum cost, and the optimal switch GCL configuration information corresponding to the minimum cost is output.

[0170] Optionally, the multi-condition constraints include: time window size constraint, time window offset constraint, and window opening time constraint;

[0171] The time window size constraint is expressed by the following formula (4):

[0172] (4)

[0173] in, Indicates the scaling factor; C represents the link bandwidth; This indicates the minimum throughput requirement for the background flow;

[0174] The constraint range of the time window offset is expressed by the following formula (5):

[0175] (5)

[0176] in, This represents the offset between window j on queue m of switches a and b; This represents the length of window j on queue m of switch b; This indicates the gating period of queue a on switch; This indicates the gating period of queue b on switch;

[0177] The window opening time constraint is expressed by the following formula (6):

[0178] (6)

[0179] Where j represents the gated window number; Indicates the gating period of queue b of switch; This indicates the length of window j in switch b; This indicates the time when window j of switch b was opened.

[0180] Optionally, the calculation process of the incremental PID control algorithm includes:

[0181] Initialize the population; calculate the adjusted output value according to the incremental PID control law;

[0182] Set a zero-output condition factor; based on the adjusted output value and the zero-output condition factor, iteratively update the population. When the iteration reaches time t+1, the updated population is obtained.

[0183] Optionally, the adjusted output value is represented by the following formula (7):

[0184] (7)

[0185] in, The first PID parameter; This is the second PID parameter; This is the third PID parameter; Indicates iteration deviation; This represents a second random number between (0,1); This represents a third random number between (0, 1); This represents the fourth random number between (0,1); Indicates adjusting the output value;

[0186] The zero-output condition factor is expressed by the following formula (8):

[0187] (8)

[0188] in, Indicates zero output factor; Indicates the upper bound of the algorithm iteration; Indicates the adjustment factor; Represents the fifth random number between (0,1); L represents the Levy flight function; This represents the overall deviation of the iteration.

[0189] This invention utilizes an improved model based on arrival-service curves from traditional network calculus. For multi-node network applications, it enhances the accuracy of traditional network calculus in latency upper bound analysis by precisely characterizing arrival and service curves. Based on the latency estimation from the improved model, it compares it with a preset latency QoS threshold to determine the schedulability of service flows. Using this as the scheduling optimization objective, an IPO algorithm is designed. Under constraints on key parameters, window opening time, and window length, the time window is optimized to maximize network schedulability, thereby obtaining the optimal GCL configuration scheme.

[0190] This invention constructs a centralized TSN control architecture based on SDN, utilizing a centralized network controller and a centralized user configuration unit to achieve unified management of the entire TSN domain. The controller uses collected service flow characteristics, including period, packet length, priority, and end-to-end latency requirements, to perform scheduling planning based on a time window scheduling problem analysis model. Under window parameter constraints, an IPO optimization algorithm is used to adaptively adjust the scheduling results, generating the final GCL configuration for the switch network. The configuration is then distributed to the switch system through a standardized southbound interface, enabling dynamic network adjustment and efficient operation.

[0191] Figure 7 This is a schematic diagram of the structure of a time-sensitive network scheduling device based on network computation provided in an embodiment of the present invention, as shown below. Figure 7 As shown, a time-sensitive network scheduling device based on network computation may include the above-mentioned... Figure 6 The illustrated time-sensitive network scheduling device based on network computation. Optionally, the time-sensitive network scheduling device 710 based on network computation may include a first processor 2001.

[0192] Optionally, the time-sensitive network scheduling device 710 based on network computation may also include a memory 2002 and a transceiver 2003.

[0193] The first processor 2001, memory 2002, and transceiver 2003 can be connected via a communication bus.

[0194] The following is combined Figure 7 A detailed introduction to each component of the network computation-based time-sensitive network scheduling device 710 is provided below:

[0195] The first processor 2001 is the control center of the time-sensitive network scheduling device 710 based on network computation. It can be a single processor or a collective term for multiple processing elements. For example, the first processor 2001 can be one or more central processing units (CPUs), application-specific integrated circuits (ASICs), or one or more integrated circuits configured to implement embodiments of the present invention, such as one or more digital signal processors (DSPs), or one or more field-programmable gate arrays (FPGAs).

[0196] Optionally, the first processor 2001 can perform various functions of the time-sensitive network scheduling device 710 based on network computation by running or executing software programs stored in the memory 2002 and calling data stored in the memory 2002.

[0197] In a specific implementation, as one example, the first processor 2001 may include one or more CPUs, for example... Figure 7 CPU0 and CPU1 are shown in the diagram.

[0198] In a specific implementation, as one example, the time-sensitive network scheduling device 710 based on network computation may also include multiple processors, for example... Figure 7 The first processor 2001 and the second processor 2004 are shown in the diagram. Each of these processors can be a single-core processor or a multi-core processor. Here, a processor can refer to one or more devices, circuits, and / or processing cores used to process data (such as computer program instructions).

[0199] The memory 2002 is used to store the software program that executes the present invention, and is controlled by the first processor 2001 to execute it. The specific implementation method can be referred to the above method embodiment, and will not be repeated here.

[0200] Optionally, the memory 2002 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto. The memory 2002 may be integrated with the first processor 2001 or may exist independently, and may be accessed through the interface circuit of the time-sensitive network scheduling device 710 based on network computation. Figure 7 (Not shown in the image) is coupled to the first processor 2001, and this embodiment of the invention does not specifically limit this.

[0201] The transceiver 2003 is used to communicate with network devices or with terminal devices.

[0202] Alternatively, transceiver 2003 may include a receiver and a transmitter. Figure 7 (Not shown separately). The receiver is used to implement the receiving function, and the transmitter is used to implement the transmitting function.

[0203] Optionally, the transceiver 2003 can be integrated with the first processor 2001 or exist independently, and can be connected via the interface circuit of the network computation-based time-sensitive network scheduling device 710. Figure 7 (Not shown in the image) is coupled to the first processor 2001, and this embodiment of the invention does not specifically limit this.

[0204] It should be noted that, Figure 7 The structure of the network computation-based time-sensitive network scheduling device 710 shown in the figure does not constitute a limitation on the router. Actual network computation-based time-sensitive network scheduling devices may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0205] Furthermore, the technical effects of the time-sensitive network scheduling device 710 based on network computation can be referred to the technical effects of the time-sensitive network scheduling method based on network computation described in the above method embodiments, and will not be repeated here.

[0206] It should be understood that the first processor 2001 in the embodiments of the present invention may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor, or it may be any conventional processor, etc.

[0207] It should also be understood that the memory in the embodiments of the present invention can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate synchronous DRAM (DDR SDRAM), enhanced synchronous DRAM (ESDRAM), synchronous linked DRAM (SLDRAM), and direct rambus RAM (DR RAM).

[0208] The above embodiments can be implemented, in whole or in part, by software, hardware (such as circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0209] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. A and B can be singular or plural. Additionally, the character " / " in this article generally indicates an "or" relationship between the preceding and following related objects, but it can also represent an "and / or" relationship. Please refer to the context for a more accurate understanding.

[0210] In this invention, "at least one" means one or more, and "more than one" means two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of a single item or a plurality of items. For example, at least one of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be a single item or multiple items.

[0211] It should be understood that, in various embodiments of the present invention, the order of the above-mentioned process numbers does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0212] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0213] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the devices, apparatuses, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0214] In the several embodiments provided by this invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0215] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0216] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0217] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0218] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A time-sensitive network scheduling method based on network calculus, characterized in that, The time-sensitive network scheduling based on network computation is implemented by a switch database module, topology discovery module, GCL scheduling calculation module, and network configuration module of a centralized control architecture based on SDN (Software-Defined Networking). The method includes: S1. The switch database module sends the stored service information and TSN switch GCL configuration information to the topology discovery module. S2. The topology discovery module collects business information and TSN switch queue gating status parameters in the current time-sensitive network of industrial automation and intelligent manufacturing, and forwards them to the GCL scheduling calculation module. The S3 and GCL scheduling calculation modules calculate the upper bound of switch latency using a network calculus-based arrival-service curve model. They then compare this upper bound with the acquired user service flow latency QoS requirements and construct a cost function. Based on this cost function, they analyze the schedulability of the current switch GCL configuration information and obtain the analysis results. Based on these results, they use a designed incremental PID control algorithm to optimize the current switch GCL configuration information, outputting the optimal switch GCL configuration information and sending it to the network configuration module. S4. The network configuration module configures the optimal switch GCL configuration information to the switch to complete the service scheduling decision.

2. The time-sensitive network scheduling method based on network calculus according to claim 1, characterized in that, The arrival-service curve model based on network calculus in S3 calculates the upper bound of switch latency, including: S31. Based on periodic time-triggered flows, a ladder function is used to construct an arrival curve representation of the target flow; based on the arrival curve representation of the target flow, flow groups in the same queue are aggregated to construct an aggregated arrival curve representation. S32. Set the window service time slot length, calculate the maximum waiting time representation of the access switch and the parameter representation of the GCL period; based on the window service time slot length, the maximum waiting time representation of the access switch and the parameter representation of the GCL period, calculate the service curve representation of the switch. S33. Combine the service curve representation of the switch and the aggregated arrival curve representation to construct an arrival-service curve model based on network calculus; analyze the upper bound of the delay based on the arrival-service curve model based on network calculus to obtain the maximum delay experienced by the service flow in the switch, which is expressed by the following formula (1): （1） in, This indicates the maximum latency experienced by the service flow in the switch; The arrival curve of business flow i is represented; The service curve of switch k is represented by ; i represents the service flow sequence number; k represents the switch sequence number.

3. The time-sensitive network scheduling method based on network calculus according to claim 1, characterized in that, The process involves comparing the upper bound of latency with the acquired QoS requirements for user service flow latency and constructing a cost function. The schedulability of the current switch GCL configuration information is analyzed based on the cost function, and the analysis results are as follows: Based on network calculus-based latency upper bound analysis, the end-to-end total latency of the service flow from the input port to the output port is calculated. Under pre-set latency QoS constraints, boundary conditions for scheduling optimization are constructed, which are expressed by the following formula (2): （2） in, This indicates the upper bound of the end-to-end queuing delay; Indicates the number of switches traversed along path i of the service flow; Indicates the link propagation delay; Indicates transmission delay; Indicates the end-to-end transmission deadline for service flow i; Based on the boundary conditions for scheduling optimization, the schedulability of the current business flow is evaluated, and the evaluation results of schedulability are obtained. The cost function for determining the schedulability of a business flow is constructed and expressed by the following formula (3): （3） in, Indicates the total number of business flows; This indicates whether flow i is schedulable in the network; the value is 0 if it is schedulable, and 1 otherwise. Based on the schedulability evaluation results, the cost function of service flow schedulability is calculated. When the cost function is 0, the current switch GCL configuration information meets the pre-set scheduling requirements of all service flows, and the current switch GCL configuration information is directly output. When the cost function is not 0, the current switch GCL configuration information is optimized.

4. The time-sensitive network scheduling method based on network calculus according to claim 1, characterized in that, Based on the analysis results, S3 uses a designed incremental PID control algorithm to optimize the current switch GCL configuration information and outputs the optimal switch GCL configuration information, including: Using queue time windows as the scheduling object and maximizing network schedulability as the objective, a multi-constraint traffic scheduling optimization problem model is constructed; wherein, the multi-constraint traffic scheduling optimization problem model uses the cost function as the minimization objective function; Based on the multi-constraint traffic scheduling optimization problem model, the designed incremental PID control algorithm is used to iteratively calculate the GCL window allocation that meets the constraints. By optimizing the cost function, the iterative calculation stops when the cost function converges to the minimum cost, and the optimal switch GCL configuration information corresponding to the minimum cost is output.

5. The time-sensitive network scheduling method based on network calculus according to claim 4, characterized in that, The multi-condition constraints include: time window size constraint, time window offset constraint, and window opening time constraint; The time window size constraint is expressed by the following formula (4): （4） in, Indicates the scaling factor; C represents the link bandwidth; This indicates the minimum throughput requirement for the background flow; The time window offset constraint range is expressed by the following formula (5): （5） in, This represents the offset between window j on queue m of switches a and b; This represents the length of window j on queue m of switch b; This indicates the gating period for queue a of switch; This indicates the gating period of queue b on switch; The window opening time constraint is expressed by the following formula (6): （6） Where j represents the gated window number; Indicates the gating period of queue b of switch; This indicates the length of window j in switch b; This indicates the time when window j of switch b was opened.

6. The time-sensitive network scheduling method based on network calculus according to claim 1, characterized in that, The calculation process of the incremental PID control algorithm includes: Initialize the population; calculate the adjusted output value according to the incremental PID control law; Set a zero-output condition factor; based on the adjusted output value and the zero-output condition factor, iteratively update the population. When the iteration reaches time t+1, the updated population is obtained.

7. The time-sensitive network scheduling method based on network calculus according to claim 6, characterized in that, The adjusted output value is expressed by the following formula (7): （7） in, The first PID parameter; This is the second PID parameter; This is the third PID parameter; Indicates iteration deviation; This represents a second random number between (0,1); This represents a third random number between (0, 1); This represents the fourth random number between (0,1); Indicates adjusting the output value; The zero-output condition factor is expressed by the following formula (8): （8） in, Indicates zero output factor; Indicates the upper bound of the algorithm iteration; Indicates the adjustment factor; Represents the fifth random number between (0,1); L represents the Levy flight function; This represents the overall deviation of the iteration.

8. A time-sensitive network scheduling device based on network computation, wherein the time-sensitive network scheduling device based on network computation is used to implement the time-sensitive network scheduling method based on network computation as described in any one of claims 1-7, characterized in that, The device includes: The switch database module is used to store service information and TSN switch GCL configuration information; The topology discovery module is used to collect statistics on service information and TSN switch queue gating status parameters in the current time-sensitive network of industrial automation and intelligent manufacturing. The GCL scheduling calculation module is used to calculate the upper bound of switch latency by constructing an arrival-service curve model based on network calculus; compare the upper bound of latency with the obtained user service flow latency QoS requirements, and construct a cost function; analyze the schedulability of the current switch GCL configuration information based on the cost function, and obtain the analysis results; based on the analysis results, optimize the current switch GCL configuration information using a designed incremental PID control algorithm, output the optimal switch GCL configuration information, and send it to the network configuration module. The network configuration module is used to configure the optimal switch GCL configuration information to the switch to complete the service scheduling decision.

9. A time-sensitive network scheduling device based on network calculus, characterized in that, The time-sensitive network scheduling device based on network calculus includes: processor; A memory storing computer-readable instructions that, when executed by the processor, implement the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium contains program code that can be invoked by a processor to execute the method as described in any one of claims 1 to 7.

Images