Time-sensitive network-oriented transmission path determination method, device and storage medium

Abstract

The application provides a transmission path determination method, device and storage medium for a time-sensitive network. For each to-be-transmitted data stream, the method comprises: selecting a first set number of paths from the time-sensitive network as candidate transmission paths corresponding to the to-be-transmitted data stream; if each candidate transmission path meets a redundancy condition, updating path redundancy, and selecting a second set number of new paths from the time-sensitive network; for each new path, if the utility corresponding to the candidate transmission path is greater than the utility corresponding to the new path, accepting the new path as the candidate transmission path corresponding to the to-be-transmitted data stream based on an acceptance probability, otherwise accepting the new path as the candidate transmission path corresponding to the to-be-transmitted data stream; and repeating the selection of the new path and the updating of the candidate transmission path and the current path redundancy until a set stop condition is reached. The application can dynamically select the path redundancy and the transmission path of each data stream.

Description

Technical Field

[0001] This invention relates to the field of vehicle communication technology, and in particular to a method and device for determining transmission paths for time-sensitive networks. Background Technology

[0002] With the rapid development of intelligent transportation systems and autonomous driving technologies, in-vehicle networks have become a critical infrastructure of modern transportation systems, and the communication needs within vehicles are increasing daily. For example, autonomous driving requires processing large amounts of environmental perception information, which necessitates that in-vehicle networks provide low latency, low jitter, and high reliability transmission guarantees. To overcome the shortcomings of traditional Ethernet in providing deterministic transmission, the IEEE 802.1 standard introduced Time-Sensitive Networking (TSN), an enhanced solution based on Ethernet. The core objective of TSN is to achieve deterministic minimum time latency in non-deterministic Ethernet, ensuring that data can be transmitted in real-time, deterministically, and reliably. Specifically, TSN can use the Time-Aware Shaper (TAS) in the IEEE 802.1Qbv standard to achieve low-latency, low-jitter transmission. TAS reduces network latency and jitter by adjusting the Gate Control List (GCL) to ensure that critical traffic is transmitted within a predetermined time window. TSN can also use the Frame Replication and Elimination for Reliability (FRER) mechanism in the IEEE 802.1CB standard to achieve high-reliability transmission. The FRER mechanism selects multiple non-overlapping paths for redundant transmission based on redundancy. In the event of link errors or packet loss, redundant frames can be used to improve the reliability of data transmission.

[0003] Choosing the appropriate redundancy level is a crucial issue in FRER routing scheduling. Higher redundancy reduces latency and increases reliability for the data streams being transmitted, but it also makes the transmission process more complex and increases bandwidth overhead. Therefore, the FRER mechanism needs to select an appropriate redundancy level based on the transmission characteristics of each data stream (such as reliability, latency, and required bandwidth); otherwise, it may affect network transmission performance. However, existing FRER routing scheduling research mainly focuses on multi-path redundancy scheduling strategies within FRER, lacking methods to consider the impact of adaptive redundancy changes on network performance. For example, researchers have combined time-variable routing and FRER with heuristic optimizers to improve data stream support and transmission efficiency while reducing resource utilization. Another example is the proposed Reliability-enhanced Flow Routing and Scheduling Algorithm (RFRSA), which uses FRER to improve transmission reliability, dynamically manages limited bandwidth and buffer resources, and maximizes network resource availability. In other words, existing methods do not consider the impact of redundancy selection strategies on network performance when scheduling routes in FRER, leading to a decline in communication performance.

[0004] Furthermore, how to address packet out-of-order delivery caused by latency differences and other factors when transmitting data streams through the transmission path in FRER is an unresolved issue (in this application, packet out-of-order delivery refers to the disordered arrangement of data packets in the data packet queue received by the receiver). Existing research mainly focuses on the design and optimization of the receiver's buffer, attempting to reduce the resource overhead caused by out-of-order delivery by improving the reordering mechanism. Specific methods include optimizing the receiver's buffer reordering strategy, adjusting the size and structure of the buffer, and reducing the waste of processing resources by setting the length of the historical sequence. These methods can improve the system's resource utilization efficiency to some extent and alleviate the additional burden caused by out-of-order delivery. However, the core of these methods is to post-process the out-of-order delivery problem that has already occurred, and they do not fundamentally reduce the probability of packet out-of-order delivery. That is, existing solutions cannot alleviate packet out-of-order delivery caused by the FRER mechanism. Summary of the Invention

[0005] In view of this, embodiments of the present invention provide a transmission path determination method and device for TSN, which can adaptively select redundancy and corresponding transmission paths for each data stream to be transmitted when implementing FRER routing scheduling, thereby dynamically adapting to changes in network status and service requirements.

[0006] One aspect of the present invention provides a transmission path determination method for TSN, which, for each data stream to be transmitted, includes the following steps:

[0007] Candidate transmission paths are selected: Based on pre-stored TSN topology information, TSN performance information, and information of the data stream to be transmitted, a first set number of paths are selected from the TSN as candidate transmission paths corresponding to the data stream to be transmitted, and the utility of each candidate transmission path is measured based on a set utility function; wherein, the first set number is determined based on a set initial path redundancy, and the data stream to be transmitted is a TT stream.

[0008] Selecting new paths: If all candidate transmission paths meet the redundancy condition, the initial path redundancy is updated using the redundancy step size, a second set number of new paths are selected from the TSN, and the utility of each new path is measured based on the set utility function; wherein, the size of the second set number is equal to the set redundancy step size; all paths selected by selecting candidate transmission paths and selecting new paths meet the network constraints and are mutually exclusive.

[0009] Update candidate transmission paths and current path redundancy: For each new path, if the utility of each candidate transmission path is greater than the utility of the new path, calculate the acceptance probability based on the Monte Carlo criterion, and determine whether to accept the new path as a candidate transmission path for the data stream to be transmitted based on the acceptance probability; otherwise, directly accept the new path as a candidate transmission path for the data stream to be transmitted; and update the current path redundancy for the data stream to be transmitted based on the number of new paths accepted.

[0010] Stop decision: Repeatedly select new paths and update candidate transmission paths and current path redundancy until the set stop condition is met, to obtain the transmission path and path redundancy corresponding to the data stream to be transmitted, so that TSN can realize the transmission of the data stream to be transmitted based on frame duplication and frame elimination mechanisms.

[0011] In some embodiments of the present invention, the TSN performance information includes the transmission delay, bandwidth and reliability information of each link in the TSN, and the information of the data stream to be transmitted includes the maximum delay, required bandwidth and minimum reliability information of the data stream to be transmitted, as well as the source node and destination node corresponding to the data stream to be transmitted.

[0012] Network constraints include latency constraints, bandwidth constraints, reliability constraints, and data flow constraints. Among them, the latency constraint is that the latency of each data flow to be transmitted on each corresponding path meets its maximum latency requirement. The bandwidth constraint is that the total bandwidth consumed by the data flows to be transmitted on each link in the TSN does not exceed the remaining bandwidth of that link. The reliability constraint is that the minimum reliability requirements of each data flow to be transmitted are met, and the minimum reliability provided by the TSN for each data flow to be transmitted is the product of the reliability of each minimum cut set. The data flow constraint is that the data flows to be transmitted input at the input ports of each transmission node in the path are the same as the data flows to be transmitted output at the output ports, and each data flow to be transmitted is transmitted on the corresponding link.

[0013] In some embodiments of the present invention, the stopping condition is set as follows: the product of the current time threshold and the set parameter does not exceed the set minimum time threshold; or

[0014] The stopping condition is set as follows: the utility of each test path does not exceed the utility of the currently updated candidate transmission path. The test path is a third set number of paths selected from the TSN based on pre-stored TSN topology information, TSN performance information, and information of the data stream to be transmitted. Each test path satisfies network constraints and is mutually exclusive.

[0015] In some embodiments of the present invention, after obtaining the transmission path corresponding to the data stream to be transmitted, the method further includes:

[0016] Based on the TSN performance information and the latency information in the data stream to be transmitted, the latency of the data stream to be transmitted on each corresponding transmission path is calculated, thereby determining the maximum latency corresponding to the data stream to be transmitted.

[0017] Based on the maximum delay, determine the time slot allocation strategy for each transmission path corresponding to the data stream to be transmitted, and obtain the candidate time slots corresponding to each transmission path corresponding to the data stream to be transmitted, so that the delay of the data stream to be transmitted is aligned on each corresponding transmission path.

[0018] In some embodiments of the present invention, after obtaining the candidate time slots, the method further includes:

[0019] If the candidate time slots corresponding to each transmission path of the data to be transmitted are available, the information of the corresponding candidate time slots is sent to the transmission nodes in the TSN to configure the time slots based on the time-aware shaping mechanism. Otherwise, while ensuring delay alignment, the delay of the data stream to be transmitted on each transmission path is adjusted, and the candidate time slots of each transmission path corresponding to the data to be transmitted are adaptively adjusted so that the candidate time slots corresponding to each transmission path are available.

[0020] In some embodiments of the present invention, accepting the new path as a candidate transmission path corresponding to the data stream to be transmitted includes: replacing the candidate transmission path with the least utility with the new path to update the candidate transmission path corresponding to the data stream to be transmitted.

[0021] In some embodiments of the present invention, before accepting the new path as a candidate transmission path corresponding to the data stream to be transmitted, the method further includes: sorting the candidate transmission paths corresponding to the data stream to be transmitted in descending order according to the magnitude of the utility of each candidate transmission path.

[0022] In some embodiments of the present invention, the utility function is used to balance the importance of latency, bandwidth, and reliability to network transmission performance;

[0023] Redundancy conditions include that the remaining bandwidth in the TSN is sufficient to meet the bandwidth required for the data stream to be transmitted on the candidate transmission path, and that the path redundancy corresponding to the data stream to be transmitted does not exceed the set empirical value.

[0024] Another aspect of the present invention provides a transmission path determination device for TSN, including a processor, a memory, and a computer program / instructions stored in the memory, wherein the processor is configured to execute the computer program / instructions, and when the computer program / instructions are executed, the device implements the steps of the method described in any of the above embodiments.

[0025] Another aspect of the present invention provides a computer-readable storage medium having a computer program / instructions stored thereon, which, when executed by a processor, implement the steps of the method described in any of the above embodiments.

[0026] The present invention proposes a transmission path determination method and device for time-sensitive networks. For each data stream to be transmitted, candidate transmission paths are updated by comparing the utility of corresponding paths and updating path redundancy, thereby realizing the transmission and reception of data streams based on the FRER mechanism. This method can adaptively select the corresponding path redundancy and transmission path for each data stream to be transmitted, thus dynamically adapting to changes in network conditions and service requirements.

[0027] Additional advantages, objects, and features of the invention will be set forth in part in the description which follows, and will also become apparent in part to those skilled in the art upon studying the description, or may be learned by practice of the invention. The objects and other advantages of the invention can be realized and obtained by means of the structures specifically pointed out in the description and drawings.

[0028] Those skilled in the art will understand that the objectives and advantages achievable with the present invention are not limited to those specifically described above, and that the above and other objectives achievable with the present invention will become clearer from the following detailed description. Attached Figure Description

[0029] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, are not intended to limit the scope of the invention. In the drawings:

[0030] Figure 1 This is a schematic diagram illustrating the implementation of data stream transmission using an overall algorithm framework in one embodiment of the present invention.

[0031] Figure 2 This is a flowchart illustrating a method for determining a transmission path for TSN according to an embodiment of the present invention.

[0032] Figure 3 This is a schematic diagram illustrating the use of the RAOP algorithm to adjust time slot allocation in one embodiment of the present invention.

[0033] Figure 4 This is a schematic diagram of the topology of an in-vehicle network in one embodiment of the present invention.

[0034] Figure 5 This is a simulation result diagram of the data stream scheduling success rate in one embodiment of the present invention.

[0035] Figure 6 This is a simulation result diagram of bandwidth utilization and data flow reliability in one embodiment of the present invention.

[0036] Figure 7 This is a simulation result diagram of packet disorder in one embodiment of the present invention. Detailed Implementation

[0037] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the embodiments and accompanying drawings. Here, the illustrative embodiments and descriptions of this invention are used to explain the invention, but are not intended to limit the invention.

[0038] It should also be noted that, in order to avoid obscuring the invention with unnecessary details, only the structures and / or processing steps closely related to the solution according to the invention are shown in the accompanying drawings, while other details that are not closely related to the invention are omitted.

[0039] It should be emphasized that the term "including / comprises" as used herein refers to the presence of a feature, element, step, or component, but does not exclude the presence or addition of one or more other features, elements, steps, or components.

[0040] It should also be noted that, unless otherwise specified, the term "connection" in this article can refer not only to a direct connection, but also to an indirect connection involving an intermediary.

[0041] In the following description, embodiments of the invention will be illustrated with reference to the accompanying drawings. In the drawings, the same reference numerals represent the same or similar parts, or the same or similar steps.

[0042] The FRER routing scheduling in this application mainly includes adaptive selection of path redundancy and redundant transmission based on FRER. The process of redundant data stream transmission based on the FRER mechanism is as follows: When the sender transmits data, the FRER function copies the data packets and transmits them along different transmission paths; when the receiver receives data, the FRER function deletes duplicate data packets, retaining only one corresponding data packet. The data stream can be transmitted within a TSN, where both the sender and receiver are transmission nodes within the TSN. The parameters of the FRER mechanism include the path redundancy K (hereinafter referred to as redundancy) corresponding to the data stream to be transmitted. The redundancy of the data stream represents the number of redundant paths required to transmit the data stream within the TSN. The transmission path corresponding to the data stream refers to the path within the TSN that can be used to transmit the data stream, and the number of transmission paths corresponding to the data stream is the sum of the redundancy of the data stream and 1. Furthermore, the data stream, data packet, and data frame mentioned in this application can be used to indicate the same data to be transmitted; stream, data packet, and frame are simply different names used for data at different protocol layers.

[0043] Current research on FRER routing and scheduling mainly focuses on multi-path redundancy scheduling strategies, lacking strategies for adaptively selecting redundancy. For example, Sun et al. (W. Sun, Y. Zou, N. Guan, X. Zhang, J. Fan, and Y. Meng, “Irfs: Acqf scheduling method integrating queue resources and flow features in time-sensitive networking,” IEEE Transactions on Vehicular Technology, pp. 1-11, 2024) in the RFRSA algorithm scenario comprehensively considered the latency, reliability requirements, network bandwidth, and buffer resource overhead of time-sensitive flows. They dynamically adjusted the allocation strategy according to the priority of time-sensitive flows, using a series of constraints to ensure low latency and high reliability requirements of time-sensitive flows, and used heuristic algorithms to dynamically calculate the optimal routing and scheduling strategy for time-sensitive flows. Although this scheme has several advantages: it reduces computational complexity through dimensionality reduction and heuristic methods, making it suitable for large-scale network deployment; it balances the latency and reliability requirements of time-sensitive flows and reduces network bandwidth and buffer resource overhead; it has dynamic adaptability, supports real-time calculation of routing and scheduling strategies, and can dynamically adapt to changes in network status and service requirements. However, regarding redundancy, the algorithm only points out that too many redundant paths will consume too many network resources, and suggests that the redundancy should not exceed 3 (that is, each data stream to be transmitted should correspond to a maximum of 4 transmission paths).

[0044] Furthermore, when using the FRER mechanism to transmit data streams through a transmission path, packet out-of-order issues can increase the resource overhead of the receiver's buffer. However, existing methods primarily address how to restore the sequence after out-of-order delivery occurs, without considering solutions from the perspective of the causes of packet out-of-order delivery. For example, existing solutions include: resolving packet out-of-order issues through receiver reordering mechanisms in FRER and optimizing buffer resource overhead using network calculus theory. However, receiver reordering mechanisms can only serve as a passive remedial measure and cannot actively prevent packet out-of-order delivery; or, by correctly setting the historical sequence length, reducing the additional resource overhead caused by out-of-order delivery.

[0045] Based on this, this application proposes an overall algorithm framework, which includes a Bandwidth-Aware for FRER (BA-FRER) algorithm and a Redundant-Aware Order Preservation (RAOP) algorithm. Specifically, the BA-FRER algorithm dynamically selects the redundancy and transmission path corresponding to each data stream to be transmitted based on a defined utility function. This algorithm can balance the current network bandwidth resource status, reliability, and latency transmission performance indicators through the defined utility function, adaptively selecting an appropriate redundancy for each data stream to be transmitted, thereby selecting a suitable transmission path to improve network performance. Further, the RAOP algorithm aligns the latency of all transmission paths based on the maximum latency of the corresponding transmission path for each data stream to be transmitted, thereby configuring the time slot resources of each data stream to be transmitted based on the TAS mechanism. This algorithm can reduce the probability of packet out-of-order delivery, thereby reducing the resource overhead of receiver buffer reordering.

[0046] Furthermore, considering that data streams in a TSN typically include TT streams, AVB streams, and BE streams (with decreasing priority), to simplify the calculation process, this application assumes that all data streams to be transmitted in the TSN that require adaptive redundancy calculation have the same priority. Therefore, the data streams (or data streams) to be transmitted that require adaptive redundancy calculation mentioned in this application all refer to TT streams, i.e., pre-planned time-triggered data streams. Moreover, the redundant paths or transmission paths mentioned in this application are all composed of links (the path between adjacent transmission nodes can be called a link) and transmission nodes. Since the source node and destination node corresponding to each data stream to be transmitted are determined, the starting node of the transmission path corresponding to each data stream to be transmitted is the same (but may use different input ports and output ports).

[0047] In some embodiments of the present invention, the overall algorithm framework proposed in this application can be used to perform operations such as... Figure 1 The four steps are as follows: Step S100: For each data stream to be transmitted, dynamically select redundancy and transmission path. For example, a heuristic algorithm can be used to dynamically select the redundancy and transmission path for each data stream to be transmitted; Step S200: Calculate the latency of different transmission paths corresponding to the same data stream to be transmitted, and align the latency of the data stream on different transmission paths based on the maximum latency, thereby performing time slot management and configuration based on the TAS mechanism. This step can delay the transmission of data packets with smaller latency on the transmission path; Step S300: Using the FRER mechanism, the sending end sends the data stream to be transmitted to the receiving end through the determined transmission path, and the receiving end eliminates the duplicate data frames received.

[0048] As an example, before executing step S100, step S400 can also be executed using the overall algorithm framework proposed in this application: based on the pre-stored TSN topology information, TSN performance information, and information of the data stream to be transmitted, all candidate paths corresponding to each data stream to be transmitted are selected from the network, so that the candidate transmission path and transmission path corresponding to the data stream to be transmitted can be selected from the candidate paths in step S100. The candidate paths mentioned in this application refer to all paths in the TSN whose latency, bandwidth, and reliability can meet the requirements of the data stream to be transmitted. The selection process is as follows: if the transmission latency of a certain path can meet the maximum latency of the data stream to be transmitted, the remaining bandwidth resources can meet the bandwidth required by the data stream to be transmitted, and the reliability of the path can meet the minimum reliability of the data stream to be transmitted, then the path can be regarded as one of the candidate paths of the data stream to be transmitted. In addition, this application can construct a path set for each data stream to be transmitted to store the candidate paths, candidate transmission paths, and finally determined transmission paths corresponding to the data stream. For example, if data stream f i The corresponding k-th (k=0,1,…,K) max K max If the d-hop path is the adaptive redundancy obtained by the BA-FRER algorithm, then the data flow f in the path set... i The corresponding k-th path can be stored in the following format: {[h k1 ,h k2 ],[h k2 ,h k3 ],…,[h kd ,h kd+1 ]}.

[0049] In some embodiments of this invention, during data transmission, a utility function is typically needed to define the value of a path given the transmitted data and network predictions. However, current solutions mostly consider flow scheduling under network resource and Quality of Service (QoS) constraints when designing utility functions. For example, the overall utility function in the RFRSA algorithm mainly considers the latency, reliability, network bandwidth, and buffer resource overhead of time-sensitive traffic. Given that network scheduling and resource allocation problems typically have high combinatorial complexity, especially when considering multiple data flows, multi-hop transmission paths, bandwidth limitations, and QoS constraints, the path selection problem has been proven to be NP-hard. Therefore, this application, considering the importance of traffic access rate, defines a utility function that optimizes the target network performance to measure the reliability and latency of data flows and the impact of network bandwidth resources on the overall utility of the network system. That is, this application uses a utility function to quantitatively measure the network performance of a path when transmitting the corresponding data flow. In short, this application uses a utility function to comprehensively consider the current network bandwidth resource utility, latency utility, and reliability utility, and the utility function's measurement result can be expressed as utility T. UNP Digital representation is performed. This applies to the data stream f to be transmitted. i Its corresponding transmission path k The formula for the utility function can be expressed as:

[0050]

[0051] Among them, w r,ik w l,ik and w b,ik These represent the data stream f respectively. i path k Reliability, path k The weight of latency and network bandwidth resources, U r,ik U l,ik and U b,ik These represent the data stream f respectively. i path k Reliability, path k The utility value corresponding to latency and network bandwidth resources. That is, the utility function in formula (1) can be used to balance the importance of latency, bandwidth and reliability to network transmission performance.

[0052] As an example, w r,ik w l,ik and w b,ik This allows the overall system to adaptively adjust to balance the importance of reliability, latency, and bandwidth resources based on network and service requirements; for example, higher latency and bandwidth resources can be optimized for different network and service needs. r,ikThis indicates that reliability is a priority, and a higher w l,ik This indicates that low latency is prioritized, with a weight w. r,ik w l,ik and w b,ik If all values ​​are 1 / 3, it indicates a balanced consideration of reliability, latency, and bandwidth resources. The w in the utility function... r,ik w l,ik and w b,ik All parameters can be defined based on experience or determined using existing algorithms (such as the analytic hierarchy process or the entropy method). This invention does not impose specific limitations on the determination method.

[0053] As an example, when using utility functions to calculate the utility corresponding to the transmission path of each data stream, U r,ik U l,ik and U b,ik The calculation method can be represented as follows. The calculation method below is only an example, and other methods can also be used to calculate reliability utility, latency utility, and bandwidth utility. This invention is not limited thereto.

[0054] ① For the data stream f to be transmitted i Its corresponding transmission path k Reliability utility U r,ik (i.e., path) k It can be a data stream f i The reliability metric provided can be used to construct the transmission path. k The reliability of all nodes and links is measured and can be expressed by the formula:

[0055] U r,ik =R nodes,ik ×R links,ik (2)

[0056] Among them, R nodes,ik and R links,ik These represent paths. k The reliability that all nodes and links can provide for the data flow is calculated using the following formula:

[0057]

[0058] in, path k The reliability of the a-th node, R lb,ik path k The reliability of the b-th link, where A represents the path. k The total number of nodes on the path, B represents the path. k The total number of links on the network.

[0059] ② For the data stream f to be transmitted i Its corresponding transmission path k The delay effect U l,ik The expression is:

[0060]

[0061] Among them, D total,k For data stream f i In path k Total delay on D max,k For data stream f i The maximum possible total delay across all transmission paths, D min,k For data stream f i The minimum possible total delay among all corresponding transmission paths.

[0062] ③ Path k Bandwidth resource efficiency U b,k (i.e., data stream f) i In the corresponding path k The expression for bandwidth utilization can be:

[0063]

[0064] Among them, BU ik Represents data stream f i In path k The bandwidth resources consumed during uplink transmission reflect the bandwidth resources consumed by that specific stream f. i Available bandwidth, LB ik path k Total available bandwidth.

[0065] In some embodiments of the present invention, when data streams are transmitted in a network, they are usually subject to certain constraints due to network performance limitations. Therefore, in order to meet the data transmission requirements, this application designs the following network constraints based on a utility function, which can be expressed by the following formula:

[0066]

[0067] Formula (7) represents the delay constraint, used to ensure the data flow f i Delay on each transmission path Satisfying the flow f i Maximum latency requirement Req lat,i Formula (8) is a bandwidth constraint used to ensure that the total bandwidth consumed by all data streams to be transmitted on the link (u,v) does not exceed the remaining bandwidth R of the link. u,v In formula (8), x i,k(u,v,q) represents the data stream f i Transmission occurs on the path q from node u to node v (where data stream f) i The k-th path is path q), r i Represents data stream f i The bandwidth required for transmission on the link (u,v), R u,v I represents the remaining bandwidth of the link (u,v) between node u and node v. n This represents the number of all data flows along the link (u,v); Equations (9) and (10) are reliability constraints, with Equation (9) used to ensure the reliability of data flow f. i Reliability on each corresponding transmission path Satisfying data flow f i Minimum reliability requirement Req rel,i Formula (10) is used to represent the network as a data flow f i Minimum reliability R provided lower,fi For each minimal cut set R cut,m The product of reliability, where the minimum cut set is obtained as follows: all nodes in the TSN form an undirected graph G(V,E), and multiple minimum cut sets can be obtained by partitioning it according to reliability; Formulas (11) and (12) are data flow constraints, and Formula (11) is used to ensure that the data flow f to be transmitted is... i Transmission is performed on the corresponding link (u,v); Formula (12) is used to ensure that the data stream to be transmitted input at the input port of each transmission node in the transmission path is the same as the data stream to be transmitted output at the output port, that is, the input stream at each transmission node is equal to the output stream, and s in Formula (12) i Represents data stream f i The source node of the corresponding k-th path, d i Represents data stream f i The corresponding destination node of the k-th path, where x i,k (v,u,q) represents the data stream f i Transmission occurs on the path q from node v to node u (data stream f) i The k-th path is path q).

[0068] As an example, when selecting a path from a TSN, in addition to being constrained by the aforementioned network constraints, it is also necessary to utilize pre-configured and stored TSN topology information, TSN performance information (which can be considered as including the performance information of links and transmission nodes in the TSN), and information of the data stream to be transmitted (including data stream performance information and starting node information). Specifically, the TSN topology information can be the connection relationship between links and nodes in the TSN, the starting node information is the source node and destination node corresponding to the data stream to be transmitted, and the information contained in the TSN performance information and the performance information of the data stream to be transmitted can be adjusted according to the design of the utility function and network constraints. For example, if the utility function and network constraints in this application are designed according to formula (1) and formula (7)-(12) respectively, then the TSN performance information can include the transmission delay, remaining bandwidth, and reliability information of each link in the TSN, and the information of each data stream to be transmitted can include the maximum delay, required bandwidth, and minimum reliability information of the data stream to be transmitted. For example, if the utility function and network constraints are designed using the three performance metrics of bandwidth, latency, and channel rate, then the TSN performance information can include the transmission latency, remaining bandwidth, and information utilization information of each link in the TSN, and the information of each data stream to be transmitted can include the maximum latency, required bandwidth, and minimum channel utilization of the data stream to be transmitted.

[0069] This application uses utility functions to measure bandwidth resource utility, delay utility, and reliability utility as examples only. Utility functions can also be used to quantify and measure other performance indicators. The network constraints in the above formulas (7)-(12) are also just examples. Furthermore, the TSN performance information and the performance information of the data stream to be transmitted mentioned above are also just examples. The network constraints, utility functions, TSN performance information, and the performance information of the data stream to be transmitted can be customized according to transmission requirements. This invention is not limited to these. For example, this application can use utility functions to measure delay utility, reliability utility, jitter utility, and throughput utility. In this case, the network constraints, TSN performance information, and the performance information of the data stream to be transmitted can be designed for delay, reliability, jitter, and throughput. For another example, this application can use utility functions to measure bandwidth utility, delay utility, and channel utilization utility. In this case, the network constraints, TSN performance information, and the performance information of the data stream to be transmitted can be designed for bandwidth, delay, and channel rate.

[0070] To optimize multiple objectives, this application proposes a TSN-oriented transmission path determination method using an overall algorithm architecture, which is implemented by a computational center. For step S100, this method can utilize the BA-FRER algorithm to balance T... UNPRegarding traffic access rate, the core of the BA-FRER algorithm is an adaptive cooling algorithm (ACA) proposed in this application. This ACA algorithm is derived from an improved heuristic algorithm. Based on the results of the ACA algorithm, the BA-FRER algorithm can adaptively select appropriate redundancy and transmission paths for each TT flow, thereby optimizing the T flow. UNP This improves network performance and resource utilization efficiency by increasing traffic access rate. Furthermore, the transmission path determination method proposed in this application transforms a multidimensional problem into multiple one-dimensional problems: determining the corresponding redundancy and transmission path for a single data stream to be transmitted, thereby determining the redundancy and transmission path for all data streams to be transmitted in the TSN network. That is, the transmission path determination method for Time-Sensitive Networks (TSN) proposed in this application can dynamically determine the redundancy and transmission path for each data stream to be transmitted using a heuristic algorithm. i The corresponding redundancy K and transmission path.

[0071] Figure 2 This is a flowchart illustrating a method for determining a transmission path for TSN according to an embodiment of the present invention. Figure 1 As shown, for a certain data stream f to be transmitted i The method includes the following steps:

[0072] Step S110, the step of selecting candidate transmission paths: Based on the pre-stored TSN topology information, TSN performance information, and information of the data stream to be transmitted, a first predetermined number of paths are selected from the TSN as candidate transmission paths corresponding to the data stream to be transmitted, and the utility of each candidate transmission path is measured based on a predetermined utility function. The first predetermined number is determined based on a predetermined initial path redundancy (first predetermined number = initial path redundancy K0 + 1).

[0073] For example, a first set number of paths can be randomly selected from the TSN based on the TSN topology information. However, to meet the data transmission requirements, it is also necessary to determine whether the paths selected in this step meet the network constraints based on the TSN performance information and the data stream information to be transmitted. In addition, when paths intersect, there is a more complex transmission process that requires consideration of many factors. Therefore, to simplify the data stream transmission, this application requires that the paths selected in this step are mutually exclusive.

[0074] As an example, the TSN in this application uses the FRER mechanism to transmit data streams. The prerequisite for this mechanism is that there are redundant paths in the transmission of data streams in the network (i.e., K≥1). Therefore, the initial path redundancy is usually not set to 0 (but in this application, k = 0, 1, ..., K is set to describe candidate transmission paths). maxFurthermore, to accurately select the transmission path corresponding to the redundancy and data flow, this application can set the initial path redundancy to 1, and the redundancy step size can also be 1. The custom values ​​of the initial path redundancy and redundancy step size mentioned above are merely examples, and this invention is not limited thereto.

[0075] Step S120, the step of selecting new paths: If all candidate transmission paths meet the redundancy condition, the initial path redundancy is updated using the redundancy step size, a second predetermined number of new paths are selected from the TSN, and the utility corresponding to each new path is measured based on a predetermined utility function. The second predetermined number is equal to the predetermined redundancy step size. Furthermore, to distinguish between the paths selected in steps S110 and S120, in the following description, the paths selected in step S110 are referred to as first candidate paths, and the new paths selected in step S120 are referred to as second candidate paths.

[0076] More specifically, if all selected first candidate paths meet the redundancy condition, the redundancy can be adaptively adjusted to select more paths from the TSN that can be used to transmit the data stream to be transmitted. Therefore, the initial path redundancy is updated using the redundancy step size, including updating the redundancy corresponding to the data stream to be transmitted to the sum of the initial path redundancy and the redundancy step size. Furthermore, when calculating the utility of each first candidate path and each second candidate path, the computing center can calculate the utility corresponding to the path using pre-configured parameters or by obtaining real-time data from the TSN. This can be achieved using existing methods and will not be elaborated upon here.

[0077] Similarly, to simplify data flow transmission, this application also requires that the selected second candidate paths are mutually exclusive, and that the first candidate paths do not intersect with the second candidate paths. That is, each path selected through steps S110 and S120 satisfies the network constraints and is mutually exclusive, including: the first candidate paths are mutually exclusive, the second candidate paths are mutually exclusive, and the first candidate paths do not intersect with the second candidate paths (since the first candidate paths and second candidate paths do not intersect, this also means they are different paths). The path non-intersection mentioned in this application refers to the fact that the links in the path are different, and the channels formed by the input ports and output ports of each node in the path are different.

[0078] As an example, redundancy conditions may include that the remaining bandwidth in the TSN is sufficient to allow the data stream to be transmitted on the first candidate path, and that the redundancy of the data stream does not exceed a set empirical value. This set empirical value can be set manually, for example, by referring to the RFRSA algorithm and setting it to 3. The above redundancy conditions are merely examples, and this invention does not specifically limit them.

[0079] Step S130, updating the candidate transmission path and the redundancy of the current path: For each second candidate path, if the utility corresponding to each first candidate path is greater than the utility corresponding to the second candidate path, then the second candidate path is accepted as the candidate transmission path corresponding to the data stream to be transmitted with a certain probability; otherwise, the second candidate path is directly accepted as the candidate transmission path corresponding to the data stream to be transmitted; and, the redundancy corresponding to the data stream to be transmitted is updated according to the number of received second candidate paths.

[0080] In some embodiments of the present invention, accepting the second candidate path as a candidate transmission path corresponding to the data stream to be transmitted with a certain probability includes: calculating the acceptance probability based on the Monte Carlo (Metropolis) criterion, and determining whether to accept the second candidate path as a candidate transmission path corresponding to the data stream to be transmitted based on the acceptance probability. The Metropolis criterion mentioned above is only an example, and its purpose is to accept new second candidate paths with probability, rather than using completely deterministic rules. Therefore, other probability acceptance algorithms can also be used, and the present invention is not limited to this. Moreover, since the redundancy is a deterministic value in this step, if a certain second candidate path needs to be used as a candidate transmission path corresponding to the data stream, then accepting the second candidate path as a candidate transmission path corresponding to the data stream to be transmitted without updating the redundancy includes: replacing the first candidate path with the lowest corresponding utility with the second candidate path to update the candidate transmission path corresponding to the data stream to be transmitted.

[0081] More specifically, for each second candidate path, the process of updating the candidate transmission path in step S130 is as follows: if the utility corresponding to the first candidate path is greater than the utility corresponding to the second candidate path, then the acceptance probability is calculated, and based on the acceptance probability, it is determined whether to use the second candidate path to replace the first candidate path with the smallest utility, and to use it as one of the candidate transmission paths corresponding to the data stream to be transmitted; otherwise, the second candidate path is used to replace the first candidate path with the smallest utility, so that the candidate transmission paths corresponding to the data stream to be transmitted include the second candidate path.

[0082] Further, updating the redundancy of the data stream to be transmitted based on the number of accepted second candidate paths includes: subtracting the number of accepted second candidate paths from the current redundancy to obtain the updated redundancy of the data stream to be transmitted. That is, the updated redundancy in step S130 = the redundancy obtained in step S120 (i.e., the current redundancy) - the number of accepted second candidate paths.

[0083] As an example, before accepting the second candidate path as the candidate transmission path corresponding to the data stream to be transmitted, the method further includes: sorting the candidate transmission paths corresponding to the data stream to be transmitted in descending order according to the utility of each candidate transmission path, so as to replace the first candidate path with the smaller utility.

[0084] Step S140, the step of determining the stop: if the set stop condition is not met, then repeat steps S120 and S130; otherwise, stop updating the candidate transmission path and obtain the transmission path and path redundancy corresponding to the data stream to be transmitted, so that TSN can realize the transmission of the data stream to be transmitted based on the frame duplication and frame elimination mechanism.

[0085] More specifically, in step S140, steps S120 and S130 are repeated to update the candidate transmission paths until a set stopping condition is met. After obtaining the final updated transmission paths, the computing center can send the information of the final determined transmission paths (e.g., the identification information of the links in each transmission path and the identification information of the input and output ports of the nodes) to the TSN, so that the corresponding data stream to be transmitted can be sent from the sending end to the receiving end based on the FRER mechanism and the transmission path information.

[0086] As an example, the stopping condition in this application can be that the utility of each test path (also known as a third candidate path) does not exceed the utility of the currently updated candidate transmission path. In this case, the specific process of stopping in step S140 can be considered as follows: based on pre-stored TSN topology information, TSN performance information, and information about the data stream to be transmitted, a third set number of third candidate paths are selected from the TSN. The utility of each third candidate path is measured based on a set utility function. The utility of each third candidate path is compared with the utility of each candidate transmission path corresponding to the data stream updated in step S130. If the utility of each third candidate path does not exceed the utility of any candidate transmission path updated in step S130, then the updating of the candidate transmission path corresponding to the data stream is stopped; otherwise, steps S120 and S130 are repeated until the set stopping condition is met.

[0087] Unlike the first and second predetermined quantities, the third predetermined quantity in this application can be randomly defined and is not required to be based on the initial path redundancy or redundancy step size. The third candidate path can be selected in the same way as the first and second candidate paths, or other methods can be used. In this application, the method for selecting candidate paths can be a heuristic algorithm or random selection. Similarly, to simplify the calculation process, each third candidate path selected in this application must satisfy network constraints and be mutually exclusive (since the third candidate path in this application is only used as an auxiliary test path to determine whether the stopping condition has been met, this application does not require the third candidate path to be mutually exclusive with the first or second candidate path).

[0088] Since finding the global optimum for all TT streams can be considered an NP-hard problem, this application improves the ant colony algorithm based on simulated annealing. This not only finds a local optimum for each TT stream but also uses the ant colony algorithm to narrow the search space, allowing the redundancy and transmission path selection results to gradually approach the global optimum. In other words, this application can determine the redundancy and transmission path for each data stream to be transmitted based on the simulated annealing algorithm.

[0089] As an example, besides comparing the utility of the third candidate path with the candidate transmission path, referring to the simulated annealing algorithm, the setting stop condition mentioned in this application can also be that the product of the current time threshold and the setting parameter does not exceed the set minimum time threshold. Specifically, after the first round of execution steps S120-S130, if the product of the initial time threshold T0 and the setting parameter α exceeds the set minimum time threshold (otherwise, the candidate transmission path is not updated), then the second round of execution steps S120-S130 continues. If the product of the current time threshold T1 (T1 = T0 * α) and the parameter α exceeds the set minimum time threshold (otherwise, the candidate transmission path is not updated), then the third round of execution steps S120-S130 continues. If the product of the current time threshold T2 (T2 = T1 * α) and the setting parameter α exceeds the set minimum time threshold (otherwise, the candidate transmission path is not updated), and so on, until the setting stop condition is met. In this application, the initial time threshold and the setting parameter α can be set with reference to the simulated annealing algorithm, and the current time threshold refers to the time threshold updated after the previous execution step S130 or the initial time threshold (the initial time threshold only indicates the current time threshold when the first execution step S140 is executed).

[0090] The two stop conditions mentioned above are merely examples and can be customized according to specific needs. This invention does not impose any specific limitations on them.

[0091] In some embodiments of the present invention, if the ACA algorithm is obtained by improving the simulated annealing algorithm, the algorithm for adaptively determining redundancy and transmission path using the BA-FRER algorithm in this application is shown in Table 1 (i.e., this application obtains the transmission path corresponding to each data stream to be transmitted based on the simulated annealing algorithm). The topology information, TSN performance information, and data stream information to be transmitted are used as inputs to the BA-FRER algorithm, and the transmission path information of the data stream to be transmitted can be output. Table 1 is an example of the BA-FRER algorithm in one embodiment of the present invention. In Table 1, for each TT stream f... i The specific process for determining the corresponding redundancy and transmission path is as follows:

[0092] The initial path redundancy K0 is set to 1, the redundancy step size is 1, and the initial time threshold T0 and parameter α are set (line 1). Since the redundancy is 1 at this time, two first candidate paths can be selected from the TSN as the data streams to be transmitted f. i The corresponding candidate transmission paths, and these two first candidate paths do not intersect and satisfy network constraints (7)-(12) (these two first candidate paths can be represented in the form of a path set as follows) That is, the algorithm in Table 1 (Line 2). If all first candidate paths meet the redundancy condition (in Table 1, the redundancy condition is that the bandwidth meets QoS and the redundancy is ≤3), update the initial path redundancy using the redundancy step size, so that the current redundancy K... new Updated to 3 (lines 5-6). Due to the change in redundancy, a new path (i.e., the second candidate path) needs to be selected from the TSN. (thus updating the path set) The second candidate path must not intersect with any of the first candidate paths, and the second candidate path must satisfy network constraints (7)-(12) (line 7). The utility of the second candidate path is measured based on the set utility function (line 8). If the utility of the second candidate path is greater than the utility of any of the first candidate paths, then the second candidate path is accepted as the data stream to be transmitted f. i The corresponding candidate transmission path; otherwise, based on the Monte Carlo criterion, with a certain probability (acceptance probability is...). Accept the second candidate path as the data stream to be transmitted. i The corresponding candidate transmission paths (lines 9-12). Based on the path set. The utility of each candidate transmission path in the data stream f iThe corresponding candidate transmission paths are sorted in descending order (line 13), and the time threshold is updated (line 14). If the product of the current time threshold (which can be the initial time threshold T0) and the set parameter α (i.e., the updated time threshold is the product of the current time threshold and α) does not exceed the set minimum time threshold T, then the process is repeated. min If the candidate transmission path is not found, then stop updating (lines 15-16); otherwise, repeat lines 6-16 of the algorithm. After the set stopping condition is finally met, the algorithm can return the set of paths determined by the BA-FRER algorithm.

[0093] Table 1 Examples of the BA-FRER algorithm

[0094]

[0095] Path delay differences are a key factor leading to packet out-of-order delivery. However, finding redundant paths with the same delay is challenging. Therefore, this application proposes the RAOP algorithm to address the packet out-of-order problem at its source. The core idea of ​​the RAOP algorithm is to delay the transmission of paths with lower delays to align the delays of transmission paths. Specifically, the computing center can determine the time slot allocation strategy for different transmission paths corresponding to the same data stream based on the core objective of the RAOP algorithm, namely delay alignment. Thus, during data stream transmission, time slots are dynamically allocated based on the TAS mechanism, and time slot offsets are applied at the sending end. This ensures that each transmission path corresponding to the same data stream is time-aligned with the path with the longest delay among all transmission paths corresponding to that data stream, thereby alleviating packet out-of-order delivery and reducing the sorting pressure on the receiving end's buffer.

[0096] In some embodiments of the present invention, after determining in step S140 that the set stopping condition has been met and the candidate transmission path update has been stopped (i.e., the transmission path corresponding to the data stream to be transmitted is obtained), step S200 can be continued, including: calculating the latency of the data stream to be transmitted on each corresponding transmission path based on the latency information in the TSN performance information and the information of the data stream to be transmitted, thereby determining the maximum latency corresponding to the data stream to be transmitted; determining the time slot allocation strategy of each transmission path corresponding to the data stream to be transmitted according to the maximum latency, obtaining the candidate time slots corresponding to each transmission path corresponding to the data stream to be transmitted, so that the latency of the data stream to be transmitted on each corresponding transmission path is aligned.

[0097] Specifically, for each data stream f to be transmitted i The specific calculation process of the RAOP algorithm is as follows:

[0098] Step S210: Determine the data stream f i The corresponding maximum latency. This represents the delay of the k-th path. Represents data stream f i The maximum delay across all corresponding transmission paths. This step can be expressed by the formula:

[0099]

[0100] As an example, the latency of the data stream on each transmission path can be calculated based on the link latency information in the TSN performance information, the maximum latency information in the performance information of the data stream to be transmitted, and the queuing time determined based on TAS. The process of calculating the latency in step S210 can be implemented using existing calculation schemes, which will not be elaborated here.

[0101] Step S220: The computing center can determine the maximum latency. Generate data stream f i The time slot allocation strategy for each corresponding transmission path is used to determine the data flow f. i Candidate time slots on each corresponding transmission path. Furthermore, time slots can be allocated to the GCL of each transmission path via the TAS mechanism. To ensure that the data stream f i Data packets from all corresponding transmission paths can arrive at the receiving end simultaneously. In step S220, data stream f i Corresponding transmission path k The size of the allocated time slot resources can be expressed as:

[0102] For data stream f i For each corresponding transmission path, the time slot allocation strategy is used to indicate the time slot resources that need to be allocated to the input and output ports of each transmission node in that transmission path, so as to align the delay of the data stream to be transmitted on each corresponding transmission path. Furthermore, the candidate time slots in this application do not represent actual time slot resources in the network, but rather are used to indicate the time slot resources that will be allocated to the ports (including input and output ports) of each transmission node in each transmission path corresponding to the data stream.

[0103] As an example, after step S210, the method further includes step S230: for data stream f i The corresponding transmission paths k Calculate the data stream f i Delay and maximum delay on each corresponding transmission path The delay difference between the data streams is used to determine candidate time slots in step S220, ensuring that the data packets corresponding to the data stream can simultaneously reach the receiving end through the transmission path. The formula for calculating the delay difference is:

[0104]

[0105] Where, Δ d,k Represents data stream f i In path k The data stream f needs to be delayed during transmission to allow the data to be transmitted. i Delay alignment on each corresponding transmission path.

[0106] In some embodiments of the present invention, after obtaining the candidate time slots, the method further includes step S240:

[0107] If the candidate time slots corresponding to each transmission path of the data to be transmitted are available, the information of the corresponding candidate time slots is sent to the TSN for time slot configuration based on the time-aware shaping mechanism. Otherwise, while ensuring delay alignment, the delay of the data stream to be transmitted on each transmission path is adjusted, and the candidate time slots of each transmission path corresponding to the data to be transmitted are adaptively adjusted so that the candidate time slots corresponding to each transmission path are available.

[0108] Since the core of the RAOP algorithm is delay alignment, after determining candidate time slots in step S220, if a candidate time slot is unavailable, it is impossible to adjust the candidate time slots of each transmission path corresponding to the data stream by reducing time slot resource allocation. Instead, the time slot allocation strategy needs to be adjusted by increasing time slot resource allocation to make the adjusted candidate time slots available. The resource amount corresponding to the adjusted available candidate time slots can be expressed as:

[0109]

[0110] in, To ensure that data stream f is available when time slots are available i Corresponding transmission path k The amount of time slot resources to be allocated, Δ t This indicates the amount of time delay adjusted to make the time slot available.

[0111] This application does not impose specific limitations on the method of adjusting candidate time slots; for example, multiple Δ slots can be set based on experience. t Furthermore, the time slot adjustment amount and time slot allocation scheme can be automatically output through methods such as model construction. In addition, in this application, "candidate time slot available" means that the time slot resources allocated to the ports of each transmission node in the transmission path are not occupied by other data streams, that is, each transmission node can use the allocated time slot resources to transmit data packets. As can be seen from the above steps S210 to S240, the RAOP algorithm proposed in this application can reduce the probability of packet out-of-order delivery during data packet transmission by increasing the delay.

[0112] Figure 3This diagram illustrates the use of the RAOP algorithm to adjust time slot allocation in this application. Assume a data stream to be transmitted has two corresponding transmission paths: Node 1 → Node 2 → Node 3 (path a) and Node 1 → Node 3 (path b). Data packet a corresponds to transmission path a, and data packet b corresponds to transmission path b. Furthermore, data packets a and b are data packets containing the same data. Figure 3 As shown in (a), if data packets a and b start transmitting at the same time on node 1, then node 3, as the receiving end, receives data packets a and b at different times. If the time slots are adjusted using the RAOP algorithm proposed in this application, as shown in (a), Figure 3 As shown in (b), increasing the delay of data packet b at node 1 (i.e., data packets a and b do not start transmission at the same time on node 1) ensures that data packets a and b arrive at the receiver simultaneously. In other words, the RAOP algorithm ensures that data packets transmitted along the transmission path corresponding to a data stream arrive at the receiver simultaneously, thereby reducing packet out-of-order delivery and improving network reliability and efficiency.

[0113] After aligning the delay using the RAOP algorithm for multiple data packets, the data flow process is as follows: Figure 1 As shown. The sender's buffer contains data stream 2 and its copy, which are transmitted via paths A and B respectively. At time t0, data packets on path B begin transmission; at delay... path At time t2, path A begins transmitting data packets. Assume that during transmission, data packets with IDs 1 and 2 on path A are lost, as is data packet with ID 4 on path B. At time t2, the receiver simultaneously receives data packets from paths A and B. Due to time delay alignment, the receiver's buffer can receive data packets with IDs 1, 2, and 3 from path A, and data packet with ID 4 from path B, in sequence.

[0114] As an example, as shown in Table 2, for each TT stream f i The transmission path P determined by the BA-FRER algorithm i Using the TSN performance information and the latency information from the data stream to be transmitted as input, the transmission path P can be obtained through the RAOP algorithm. i Corresponding candidate time slot set Among them, y i,k This indicates that for TT stream f i The corresponding k-th transmission path will be allocated time slot resources. For each data stream f to be transmitted... i The RAOP algorithm is as follows:

[0115] Calculate the latency of the data stream to be transmitted on each corresponding transmission path. This determines the maximum delay corresponding to the data stream to be transmitted. (Line 2). For data stream f i The corresponding transmission paths k Initialize the candidate time slot to the time slot with the maximum delay on the path. (Line 4). Determining the data flow f i If all candidate time slots for all transmission paths are available, the corresponding candidate time slot information is sent to the transmission nodes in the TSN for time slot configuration based on TAS; if candidate time slots are unavailable (if data flow f i If any candidate time slot for any corresponding transmission path is unavailable, iterate through each hop on the path, reset the candidate time slot (line 8), and update the size of the time slot resource with the maximum delay on the path to the sum of the current delay and the time slot adjustment (lines 9-10). Repeat this process until the data flow f... i All corresponding time slots are available (line 11). Finally, the algorithm generates the TT stream f. i The corresponding candidate time slot set (line 12) is returned. (Line 13)

[0116] Table 2 Examples of the RAOP algorithm

[0117]

[0118] The complexity of the BA-FRER algorithm in Table 1 of this application can be expressed as O(K). max ·(|V|+|E|)+O(K max ·logK max Among them, K max This represents the adaptive redundancy determined by the BA-FRER algorithm, where V represents the total number of nodes in the TSN, and E represents the total number of links in the TSN. Furthermore, the time complexity of the RAOP algorithm is primarily determined by the data flow f. i The corresponding number of transmission paths and the number of hops in the path determine: for each TT stream f i The RAOP algorithm requires calculating the delay of each redundant path, with a complexity of O(K). max ·L)(where L is the number of hops in each redundant path), with O(K max The complexity of RAOP is O(K) as it allocates time slots for each path. Therefore, the total complexity of the RAOP algorithm can be expressed as O(K). max ·2L). Since the number of redundant paths in TSN networks is usually small and the path length is limited to 7 hops, the RAOP algorithm can run efficiently in real-time networks and is suitable for large-scale time-sensitive network environments.

[0119] In a specific embodiment of the present invention, the method proposed in this application is evaluated through simulation experiments, and the results are obtained. Figures 5 to 7 The simulation results are as follows. The simulation conditions and results are as follows:

[0120] ① Experimental parameters and environment of the scheme

[0121] The in-vehicle network (IVN) topology, built using BAIC's ARCFOX centralized electrical and electronic architecture, is as follows: Figure 4 As shown, the topology includes 5 switches and 26 terminal system nodes (such as sensors, controllers, and actuators). This topology reflects the actual network layout of ARCFOX vehicles and can simulate the communication needs between different modules within the vehicle.

[0122] Furthermore, in order to evaluate the performance of the overall algorithmic framework of this application under a wider range of network conditions, based on Figure 4 The topology generated an Erdos-Renyi (ER) random graph, which consists of 50 TSN switches and 625 links. This random graph model is used to simulate a complex industrial network environment with a large number of network nodes and links. After connecting the network nodes, the link failure rate is evenly distributed between 0.01 and 0.1, and the link bandwidth is set to 1Gbps.

[0123] Furthermore, the data streams to be transmitted in this network are characterized as follows: each data stream in the network is generated by randomly selecting a pair of nodes as its source and destination nodes. The generation period of the random data stream can be randomly selected from the set {2ms, 4ms, 5ms, 10ms, 15ms}, and the deadline is equal to the period. The data packet size corresponding to each data stream is 1000B, the reliability of each data stream is between 0.9 and 0.9999, and the data streams are arbitrarily selected from the set {1, 2, 3}.

[0124] Furthermore, the experimental environment was set up as follows: Typical industrial network scenarios were simulated using Python 3.9 and the NetworkX 2.6.3 library. The performance of the BA-FRER algorithm was evaluated in terms of network bandwidth utilization, traffic access rate, and reliability. The effectiveness of the RAOP algorithm in redundant path alignment and reducing packet out-of-order delivery was also evaluated. A simulation environment for the TAS mechanism was built using the simpy 4.0.1 and numpy 1.21.2 packages in Python. Discrete event simulation was implemented using the simpy package in the TAS simulation, and a custom time window and GCL scheduling mechanism were used to simulate packet transmission and reception. Numerical calculations, including time window segmentation and delay calculation, were performed using the numpy package.

[0125] ② Simulation results of data stream scheduling success rate

[0126] The reliability-enhanced flow routing and scheduling (RFRSA) algorithm is used as a benchmark method and compared with the BA-FRER algorithm proposed in this application: RFRSA can operate with fixed redundancy, balancing bandwidth and buffer capacity constraints while ensuring low latency and high reliability; while this application adopts adaptive redundancy to balance network bandwidth resources, reliability, latency and traffic access rate.

[0127] In this simulation experiment, this application randomly generated and scheduled 50, 100, 150, 200, 250, 300, 350, and 400 TT flows. Under the network topology of ER random graph and IIVN, the randomly generated TT flows were routed and scheduled using a redundancy adaptive selection strategy, and the performance of the BA-FRER algorithm was evaluated using the TT flow scheduling success rate.

[0128] like Figure 5 As shown, regardless of the fixed redundancy selected by the RFRSA algorithm, the scheduling success rate of the BA-FRER algorithm is consistently higher than that of the RFRSA algorithm. When the number of TT flows is small, the difference in scheduling success rate between the two algorithms is not significant; however, as the number of TT flows increases, whether K=1, K=2, or K=3, the scheduling success rate of the RFRSA algorithm decreases significantly, especially when the redundancy K=3, the scheduling success rate is the lowest. This indicates that the BA-FRER algorithm proposed in this application, through optimized scheduling and path selection, can manage complex network traffic more efficiently and achieve a higher traffic transmission success rate. Furthermore, Figure 5 In the figure, (a) represents the simulation results of the scheduling success rate of the ER random topology. Figure 5 (b) in the figure represents the simulation results of the scheduling success rate of the IVN topology, which is compared with the results of the simulation. Figure 5 As shown in (a) and (b), under the same load, the scheduling success rate of the ER random topology is better than that of the IVN topology. This is because the connectivity of the ER random topology is significantly better than that of the IVN topology, and the redundant paths in the IVN topology often require the use of links between TSN switch nodes. This leads to the number of successfully scheduled traffic in the IVN topology gradually approaching the maximum network traffic capacity under high load conditions (more than 300 TT flows).

[0129] ③ Simulation results of bandwidth utilization and data stream reliability

[0130] In this simulation experiment, to evaluate the bandwidth resource utilization and reliability of the BA-FRER algorithm, both the BA-FRER and RFRAS algorithms were tested under ER random topology and IVN topology. The RFRAS algorithm was operated with a fixed redundancy K = 1, 2, or 3 to balance network bandwidth resource utilization and transmission reliability. Furthermore, in this simulation experiment, 50, 100, 150, 200, 250, 300, 350, and 400 TT flows were randomly generated and scheduled.

[0131] like Figure 6 As shown, the BA-FRER algorithm outperforms the RFRSA algorithm in both bandwidth resource utilization and transmission reliability. Figure 6 In the figure, (a) and (c) represent the bandwidth utilization and data flow reliability under the ER random topology, respectively. Figure 6 In the diagram, (b) and (d) represent bandwidth utilization and traffic reliability under the IVN topology, respectively.

[0132] In terms of bandwidth utilization, BA-FRER is not significantly different from RFRSA under low load conditions (e.g., 50 TT streams). However, as the load increases, BA-FRER's bandwidth utilization is almost twice that of RFRSA, especially when RFRSA's redundancy is K=3, where its bandwidth utilization drops significantly. Regarding reliability, BA-FRER's reliability under low load conditions is comparable to RFRSA's reliability under high redundancy (K=2 or 3). However, under high load conditions, RFRSA's reliability drops sharply when redundancy is K=1, and is significantly lower than BA-FRER's.

[0133] from Figure 6 As can be seen from the above, the BA-FRER algorithm proposed in this application can maintain high bandwidth utilization while avoiding the additional overhead caused by redundant data through more efficient scheduling and resource utilization strategies, especially under high load conditions. This shows that BA-FRER maintains high transmission reliability while avoiding bandwidth waste, demonstrating its better adaptability and scalability in complex network environments.

[0134] ④ Simulation results of packet disorder

[0135] In the process of measuring packet reordering in the RAOP algorithm through simulation experiments, each data stream was set to consist of 500 1000B data packets, with 800us as a time slot unit. Figure 7As shown, compared to the existing RFRSA algorithm, the number of out-of-order packets is significantly lower when combining RFRSA and RAOP algorithms. This indicates that the proposed RAOP algorithm can effectively reduce out-of-order issues during packet transmission, thereby reducing resource consumption at the receiver when processing packet reordering. Simulation results show that, under the same load, the RAOP algorithm can minimize out-of-order occurrences by optimizing the sending and transmission order of packets.

[0136] The results obtained from the above simulation experiments show that the BA-FRER algorithm proposed in this application can improve network bandwidth resource utilization, traffic access rate and reliability, and the RAOP algorithm can reduce the probability of packet out-of-order occurrence.

[0137] The transmission path determination method for Time-Sensitive Networks (TSNs) proposed in this application has the following advantages:

[0138] ① It can balance network bandwidth resource utilization, traffic reliability, and traffic access rate. The overall algorithm architecture proposed in this solution can improve network bandwidth resource utilization, traffic reliability, and traffic access rate by dynamically selecting the redundancy of each data stream.

[0139] ② It provides packet ordering at the packet granularity level (also known as packet scheduling), improving buffer resource utilization. This scheme proposes the RAOP algorithm, which provides packet ordering at the packet granularity level in industrial control topologies, ensuring that the same data stream arrives at the receiving end in sequence through multiple transmission paths. Simultaneously, because the packets arrive in order, the degree of packet out-of-order delivery is reduced, significantly decreasing the reordering pressure on the receiving end's buffer and thus reducing the overhead of the receiving end's buffer resources.

[0140] ③ It has strong versatility. The scheme proposed in this application is compatible with the gating mechanism of time-sensitive networks, and the routing scheduling algorithm can be applied to all networks that conform to the IEEE 802.1Qbv standard based on existing software.

[0141] Corresponding to the above method, the present invention also provides a transmission path determination device for Time-Sensitive Networks (TSN). The device includes a computer device, which includes a processor and a memory. The memory stores computer programs / instructions, and the processor executes the computer programs / instructions stored in the memory. When the computer programs / instructions are executed by the processor, the device implements the steps of the method described above.

[0142] This invention also provides a computer-readable storage medium storing a computer program / instructions thereon, which, when executed by a processor, implements the steps of the aforementioned edge computing server deployment method. The computer-readable storage medium can be a tangible storage medium, such as random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, floppy disks, hard disks, removable storage disks, CD-ROMs, or any other form of storage medium known in the art.

[0143] Those skilled in the art will understand that the exemplary components, systems, and methods described in conjunction with the embodiments disclosed herein can be implemented in hardware, software, or a combination of both. Whether 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 implementation should not be considered beyond the scope of this invention. When implemented in hardware, it can be, for example, electronic circuits, application-specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc. When implemented in software, the elements of this invention are programs or code segments used to perform the desired tasks. The programs or code segments can be stored in a machine-readable medium or transmitted over a transmission medium or communication link via data signals carried in a carrier wave.

[0144] It should be clarified that the present invention is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of the present invention is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications, and additions, or change the order of steps, after understanding the spirit of the present invention.

[0145] In this invention, features described and / or illustrated for one embodiment may be used in the same or similar manner in one or more other embodiments, and / or combined with or in place of features of other embodiments.

[0146] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, various modifications and variations of the embodiments of the present invention are possible. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for determining transmission paths in Time-Sensitive Networks (TSNs), characterized in that, For each data stream to be transmitted, the method includes the following steps: Candidate transmission paths are selected: Based on pre-stored TSN topology information, TSN performance information, and the information of the data stream to be transmitted, a first set number of paths are selected from the TSN as candidate transmission paths corresponding to the data stream to be transmitted, and the utility of each candidate transmission path is measured based on a set utility function; wherein, the first set number is determined based on a set initial path redundancy, and the data stream to be transmitted is a TT stream. Selecting new paths: If all candidate transmission paths meet the redundancy condition, the initial path redundancy is updated using the redundancy step size, a second set number of new paths are selected from the TSN, and the utility of each new path is measured based on a set utility function; wherein, the size of the second set number is equal to the set redundancy step size; all paths selected through the selection of candidate transmission paths and the selection of new paths meet the network constraints and are mutually exclusive; Update candidate transmission paths and current path redundancy: For each new path, if the utility of each candidate transmission path is greater than the utility of the new path, calculate the acceptance probability based on the Monte Carlo criterion, and determine whether to accept the new path as a candidate transmission path for the data stream to be transmitted based on the acceptance probability; otherwise, directly accept the new path as a candidate transmission path for the data stream to be transmitted; and update the current path redundancy for the data stream to be transmitted based on the number of accepted new paths. Stop the decision: Repeat the process of selecting a new path and updating the candidate transmission path and the current path redundancy until the set stop condition is met, so as to obtain the transmission path and path redundancy corresponding to the data stream to be transmitted, so that TSN can realize the transmission of the data stream to be transmitted based on the frame duplication and frame elimination mechanism.

2. The method according to claim 1, characterized in that, The TSN performance information includes the transmission latency, bandwidth, and reliability information of each link in the TSN. The information of the data stream to be transmitted includes the maximum latency, required bandwidth, and minimum reliability information of the data stream to be transmitted, as well as the source node and destination node corresponding to the data stream to be transmitted. The network constraints include latency constraints, bandwidth constraints, reliability constraints, and data flow constraints. Specifically, the latency constraint ensures that the latency of each data flow to be transmitted on its corresponding path meets its maximum latency requirement; the bandwidth constraint ensures that the total bandwidth consumed by the data flows to be transmitted on each link in the TSN does not exceed the remaining bandwidth of that link; the reliability constraint ensures that the minimum reliability requirements of each data flow to be transmitted are met, and the minimum reliability provided by the TSN for each data flow to be transmitted is the product of the reliability of each minimum cut set; and the data flow constraint ensures that the data flows to be transmitted input at the input ports of each transmission node in the path are the same as the data flows to be transmitted output at the output ports, and that each data flow to be transmitted is transmitted on its corresponding link.

3. The method according to claim 1, characterized in that, The set stop condition is that the product of the current time threshold and the set parameter does not exceed the set minimum time threshold; or The set stopping condition is that the utility of each test path does not exceed the utility of the currently updated candidate transmission path; wherein, the test path is a third set number of paths selected from the TSN based on pre-stored TSN topology information, TSN performance information and the information of the data stream to be transmitted, and each test path satisfies network constraints and is mutually exclusive.

4. The method according to claim 2, characterized in that, After obtaining the transmission path corresponding to the data stream to be transmitted, the method further includes: Based on the TSN performance information and the latency information in the data stream to be transmitted, the latency of the data stream to be transmitted on each corresponding transmission path is calculated, thereby determining the maximum latency corresponding to the data stream to be transmitted. Based on the maximum delay, the time slot allocation strategy for each transmission path corresponding to the data stream to be transmitted is determined, and the candidate time slots corresponding to each transmission path corresponding to the data stream to be transmitted are obtained, so that the delay of the data stream to be transmitted is aligned on each corresponding transmission path.

5. The method according to claim 4, characterized in that, After obtaining the candidate time slots, the method further includes: If the candidate time slots corresponding to each transmission path of the data to be transmitted are available, the information of the corresponding candidate time slots is sent to the transmission nodes in the TSN to configure the time slots based on the time-aware shaping mechanism. Otherwise, while ensuring delay alignment, the delay of the data stream to be transmitted on each transmission path is adjusted, and the candidate time slots of each transmission path corresponding to the data to be transmitted are adaptively adjusted so that the candidate time slots corresponding to each transmission path are available.

6. The method according to claim 1, characterized in that, The step of accepting the new path as a candidate transmission path for the data stream to be transmitted includes: replacing the candidate transmission path with the least utility with the new path to update the candidate transmission path for the data stream to be transmitted.

7. The method according to claim 1, characterized in that, Before accepting the new path as a candidate transmission path for the data stream to be transmitted, the method further includes: sorting the candidate transmission paths for the data stream to be transmitted in descending order according to the utility of each candidate transmission path.

8. The method according to claim 1, characterized in that, The utility function is used to balance the importance of latency, bandwidth, and reliability to network transmission performance; The redundancy conditions include that the remaining bandwidth in the TSN is sufficient to meet the bandwidth required for the data stream to be transmitted on the candidate transmission path, and that the path redundancy corresponding to the data stream to be transmitted does not exceed a set empirical value.

9. A transmission path determination device for Time-Sensitive Networks (TSN), comprising a processor, a memory, and a computer program / instructions stored in the memory, characterized in that, The processor is configured to execute the computer program / instructions, and when the computer program / instructions are executed, the device implements the steps of the method as described in any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program / instructions stored thereon, characterized in that, When the computer program / instructions are executed by the processor, they implement the steps of the method as described in any one of claims 1 to 8.

Images