Delay analysis method for avb traffic of tsn network facing incremental scenario

Abstract

This invention discloses an AVB traffic delay analysis method for TSN networks in incremental scenarios. Based on network calculus, this method overcomes the shortcomings of traditional analysis methods that rely on the global GCL scheduling table. First, the method characterizes the inhibitory effect of TT traffic on AVB flow transmission. Without parsing the full scheduling table, it constructs a fast TT arrival curve using only the maximum number and length of TT traffic slots. Second, it combines the hop-by-hop reshaping characteristics of the ATS shaper and the credit constraint mechanism of the CBS shaper to construct the arrival and service curve models of the AVB flow at the output port. Finally, it calculates the worst-case delay time at the port using the maximum horizontal distance between the arrival and service curves of the AVB flow, and accumulates the values ​​hop-by-hop along the AVB flow transmission path to obtain the end-to-end worst-case delay time. Compared with traditional network calculus methods, the incremental delay analysis method proposed in this invention does not rely on the global traffic information of the TSN network and can reduce computational complexity.

Description

Technical Field

[0001] This invention relates to the field of Time-Sensitive Networking (TSN) technology, and more particularly, to an AVB traffic delay analysis method for TSN networks in incremental scenarios. Background Technology

[0002] In airborne networks, TSN enables various traffic types to coexist on the same link and achieve analyzable Quality of Service (QoS) by overlaying queuing, gating, and shaping rules on standard Ethernet. A common practice is to stratify traffic according to its deterministic and bandwidth guarantee requirements: TT (Time-Triggered) flows require periodic end-to-end time slots; AVB (Audio Video Bridging) flows emphasize bounded delay and bandwidth reservation; and BE (Best-Effort) flows do not require hard real-time and only utilize remaining capacity. To address these differences, the IEEE 802.1 family of standards defines several shapers: Time-Aware Shaper (TAS, 802.1Qbv) relies on network-wide clock synchronization (such as 802.1AS) and uses a gated list (GCL) to open and close queue gates according to a schedule, drawing a predictable transmission window for TT traffic on the timeline, thereby achieving extremely low and deterministic latency and jitter; Asynchronous Traffic Shaper (ATS, 802.1Qcr) does not rely on coordinating a unified periodic table for the entire network, but instead uses local clocks at each node to "reshape" each hop's traffic, constraining bursts to a leaky bucket envelope to suppress cross-hop burst cascading and facilitate the provision of worst-case bounds; Credit-Based Shaper... Shaper (CBS, 802.1Qav) maintains credit values ​​for AVB-type queues and allocates link bandwidth over time using idle slope and sending slope, freeing up channels for high-priority services while avoiding blocking low-priority information flows.

[0003] Network calculus provides an abstract description of Time-Sensitive Networking (TSN), treating switches and end systems in a TSN network as network elements. These network elements share the common characteristic of providing some service to the data flow in the network, and can also be called servers. In the network calculus process, mathematical functions are used to describe the traffic characteristics of the data flow in the network and the service capabilities of network elements. The function describing the traffic characteristics of the data is called the flow arrival curve, and the function describing the server's service capability is called the server service curve. The arrival and service curves of TT flows were published in "Multi-Window Partition Scheduling Analysis Based on Network Calculus," He Feng et al., Acta Aeronautica Sinica, 2023, 44(2), January 25, 2023.

[0004] With the increasing demands of airborne network applications, networks need to flexibly add or remove traffic during operation to meet requirements such as temporary task loading, task mode switching, and system expansion. This makes incremental scenarios increasingly important. Network calculus constructs traffic arrival curves and node service curves, and uses the horizontal / vertical deviations of these curves to calculate latency and backlog upper bounds, respectively, thereby enabling computable analysis of worst-case network performance. However, in incremental scenarios, traditional real-time analysis methods of network calculus, due to their reliance on global network information and computational complexity, struggle to meet the controller's rapid response requirements to dynamic traffic. How to perform calculations as quickly as possible remains a pressing issue. Summary of the Invention

[0005] With the increasing demands of airborne network applications, networks need to flexibly add or delete traffic during operation to meet requirements such as temporary task loading, task mode switching, and system expansion. This makes incremental scenarios increasingly important. Traditional network calculus and analysis methods for incremental scenarios require acquiring all network information, resulting in high computational complexity. To address this issue, this invention proposes a latency analysis method for TSN network AVB traffic in incremental scenarios, where TT traffic (… Changes in AVB streams ( ) thus affect the AVB stream ( ) During transmission, this method does not require obtaining global traffic information of the TSN network, but only needs to... Calculation of the maximum number of time slots This invention addresses the worst-case latency of AVB traffic in incremental TSN networks, specifically for hybrid shaping systems employing TAS, ATS, and CBS. Based on network calculus, this method overcomes the limitations of traditional methods relying on global GCL scheduling tables. First, it characterizes the inhibitory effect of TT traffic on AVB flow transmission. Without parsing the full scheduling table, it constructs a fast TT arrival curve using only the maximum number and length of TT slots. Second, combining the hop-by-hop reshaping characteristics of the ATS shaper and the credit constraint mechanism of the CBS shaper, it constructs arrival and service curve models for AVB flows at the output port. Finally, it calculates the worst-case latency at the port using the maximum horizontal distance between the arrival and service curves of the AVB flow, and accumulates this hop-by-hop along the AVB flow transmission path to obtain the end-to-end worst-case latency.

[0006] The TSN network AVB traffic delay analysis method for incremental scenarios described in this invention includes the following steps:

[0007] S1, read and organize network and traffic information: obtain node information, topology connection relationship and port rate; read traffic information, including the maximum number of time slots and the maximum time slot length of TT traffic, the path, priority, frame length and other information of any AVB stream, and count the traffic set flowing through each node in the TSN network.

[0008] S2, Construct a fast arrival curve model for TT traffic: Based on the maximum number of TT traffic time slots and the maximum time slot length (including cases where the time slot lengths are the same or different), establish a fast arrival curve for TT traffic occupancy at the output port.

[0009] S3. Construct AVB flow arrival and service curve models: For arrival curves, combining the hop-by-hop reshaping characteristics of the ATS shaper, multiple AVB flows in the same priority shared queue are aggregated according to the shaped leaky bucket parameters to obtain class-level arrival curves; For service curves, considering the segmentation of available transmission time by TAS scheduling, the upper and lower bound constraints of the CBS shaper credit mechanism, and the blocking effect caused by different priority frame lengths, the service curves obtainable by the shared queue of AVB flows at the port are established, thus forming a unified curve pair that can be used for network computation and solution.

[0010] S4, calculate the end-to-end worst-case latency of the AVB stream: at each output port, the worst-case latency of that port is obtained from the maximum horizontal distance between the AVB arrival curve and the AVB service curve; then, the worst-case latencies of each hop output port are summed along the AVB stream transmission path to obtain the end-to-end worst-case latency, which is used for real-time analysis in incremental scenarios.

[0011] The analysis results obtained after processing steps S1-S4 satisfy the following properties in an engineering sense: Under the premise that the input parameters and curve construction rules are consistent, the given upper bound of the end-to-end delay of the AVB flow is strictly conservative; Under the premise that the path and port mapping is clear, the delay of each hop port can be calculated separately and cumulatively, which is convenient for locating the bottleneck port; In the scenario of incremental traffic changes, when the maximum number of time slots of TT traffic does not change, the existing results can be reused without recalculating the delay of the AVB flow in the entire TSN topology network.

[0012] The technical effects of this invention are as follows:

[0013] (1) Compared with traditional network calculus analysis methods, this method does not require global traffic information, has extremely low computational complexity, and requires significantly less computation time than traditional analysis methods.

[0014] (2) This method is for incremental scenarios. No matter how the TT traffic changes, as long as the maximum number of TT traffic slots remains unchanged, there is no need to recalculate the delay of the AVB stream.

[0015] (3) When calculating the delay of the AVB stream using this method, the delay of each hop port can be calculated separately and accumulated, which makes it easier to locate the bottleneck port. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of a network topology for a time-sensitive network consisting of 4 switches and 8 end systems.

[0017] Figure 2 This is a flowchart of a delay analysis method for AVB streams in a TSN network for incremental scenarios, according to the present invention.

[0018] Figure 3 This is an end-to-end delay distribution diagram of the AVB stream calculated using the method of this invention.

[0019] Figure 4 This is a comparison chart showing the delay calculation time of AVB streams when the number of TT streams increases, using the method of this invention and the refined time analysis method. Detailed Implementation

[0020] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments. The examples of the parameters listed are merely preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

[0021] Construction of network topology for time-sensitive networks

[0022] Select the topology of the TSN network. For ease of explanation, an example is given, where the TSN network structure consists of 4 switches and 8 end systems, as follows: Figure 1 As shown in the diagram. The switches are Switch 1, Switch 2, Switch 3, and Switch 4; the end systems are End System 1, End System 2, End System 3, End System 4, End System 5, End System 6, End System 7, and End System 8. In a TSN network structure, both switches and end systems are referred to as network nodes. .

[0023] In this invention, the network topology in a time-sensitive network is denoted as... And network topology It contains a set of nodes and Connecting Edge Set ,Right now .

[0024] subscript The subscript indicates the identifier of a node in the network topology. This represents the total number of nodes in the network topology. Where: This represents the first node; Indicates the second node;

[0025] Indicates the first One node; This represents the last node. For ease of explanation, Also known as any node. .

[0026] In this invention, a node on a traffic path can also be referred to as a node.

[0027] In this invention, nodes can be distinguished using consecutive numerical identifiers. Network topology. Directed edges are represented using a directed connection matrix, denoted as . The A value of 1 indicates the existence of a directed edge, while a value of 0 indicates no directed edge. By... The value of is assigned to indicate whether there is a directed edge between two nodes. All edges form the edge set. .

[0028] exist In this context, the traffic characteristic information of any network traffic is denoted as... ,and .

[0029] Frame length, in bytes.

[0030] The flow period is measured in milliseconds (ms).

[0031] Priority is indicated by a value from 0 to 7. The traffic type classification is based on the article "A Review of Time-Sensitive Network Traffic Shaping Techniques" published in the January 2022 issue of *Microelectronics & Computer*, Volume 39, No. 1, by Zhang Lei and Wang Panpan.

[0032] This refers to the bandwidth used for data transmission, measured in Mbps.

[0033] The source node of the flow.

[0034] The destination node of the flow.

[0035] The routing of a flow also represents the set of nodes that the network traffic passes through on the transmission path.

[0036] This refers to the token bucket to which the priority queue of the flow belongs within the node. In this invention, the token bucket to which the computing node belongs employs the token bucket algorithm, which can be found in "Video Traffic Analysis and QoS Management," by Huang Tianyun, Chengdu: University of Electronic Science and Technology Press, March 2013, pp. 94-96. The token bucket algorithm is one of the most commonly used algorithms in network traffic shaping and rate limiting. Tokens are stored in a queue; input data enters the token bucket, and flows out of the token bucket after obtaining a sufficient number of tokens.

[0037] Under the IEEE 802.1 TSN standard, TSN performs precise network resource scheduling on a per-flow basis, such as reserving bandwidth and planning transmission time. In network topology... The set of TT streams in the dataset is denoted as Abbreviated as TT flow-flow set Also known as TT traffic In the Any TT flow in the data is denoted as subscript This is the identifier for the TT stream. The traffic characteristic information is denoted as (abbreviated as TT flow-flow characteristics), and TT flow-flow characteristics The information carried is represented as The aforementioned The maximum time slot length of the TT flow is denoted as .Should The fast arrival curve used to calculate TT flow.

[0038] This is the phase offset of the TT stream. This describes whether the time slots of multiple TT streams are aligned or staggered, thus affecting the interference pattern of AVB streams.

[0039] Under the IEEE 802.1 TSN standard, TSN performs precise network resource scheduling on a per-flow basis, such as reserving bandwidth and planning transmission time. In network topology... The stream set of AVB streams in the code is denoted as . AVB stream-stream set (abbreviated as AVB stream-stream set) Also known as AVB traffic In the Any AVB stream in the data is denoted as subscript This is the identifier for the AVB stream. The traffic characteristic information is denoted as (AVB stream - flow characteristics), and AVB stream - flow characteristics The information carried is represented as Regarding the aforementioned The maximum timeslot length of AVB streams is not set because the arrival and service curves of AVB streams are calculated based on traditional network calculus theory.

[0040] In this invention, the topology of the TSN network is... The traffic characteristic information is formed into a text file ( The text file contains information that serves as input for delay analysis of the AVB stream.

[0041] See Figure 2 The present invention provides an AVB traffic delay analysis method for TSN networks in incremental scenarios, comprising the following steps:

[0042] S1, Read the association information of the TSN network topology;

[0043] In this invention, step S1 is used to establish the basic dataset for delay analysis of AVB flows in the TSN network. Its goal is to uniformly map network topology, link parameters, and service flow attributes into a computable data structure. By standardizing the reading of traffic and node information, the path relationships, priority relationships, and rate constraints of each AVB and TT flow in the TSN network can be clearly defined, providing consistent input for constructing the arrival and service curves of AVB traffic and avoiding analytical biases caused by inconsistent data sources.

[0044] in accordance with Figure 1 The information file (txt file) showing the TSN network topology includes the following associated information:

[0045] Read and organize TSN network and traffic information to obtain node information, topology connections, and port speeds;

[0046] Read TT traffic information, including at least the maximum number of TT traffic time slots and the maximum time slot length;

[0047] Read AVB traffic information, as well as the path, priority, and frame length of any AVB stream, and count the AVB traffic flowing through each node in the TSN network.

[0048] During the process of reading TT and AVB traffic information, the node output port rate C, the maximum number of TT traffic time slots n, and the maximum TT traffic time slot length W are specified. TT The AVB stream ID, frame length (in bytes), priority, and propagation path are specified. Among these, the maximum number of time slots and the maximum time slot length for TT traffic are key data points for rapid calculation in incremental scenarios.

[0049] Based on the input txt file, read the switch and end system information, including the switch number, node name, and node connection status. Construct the network topology as follows: Figure 1 As shown, the hop-by-hop output port sequence of each AVB stream is parsed based on the topology relationship to obtain the set of paths for delay summation of the stream.

[0050] S2, Construct a fast arrival curve model for TT traffic;

[0051] The present invention constructs a fast arrival curve model for TT traffic in incremental scenarios to address the periodic time slot occupancy characteristics caused by TAS gating. Without obtaining a complete global gating list, the arrival curve of TT traffic occupancy at the node output port is established based on the maximum number of time slots and the maximum time slot length of TT traffic.

[0052] When constructing a TT arrival curve in an incremental scenario, it is not necessary to use complete TT traffic information. It can be calculated quickly, hence it is called a fast arrival curve.

[0053] TAS gating refers to "Quantitative Performance Comparison of Various TrafficShapers in Time-Sensitive Networking", L. Zhao, P. Pop and S. Steinhorst, Section III-A, IEEE Transactions on Network and Service Management, vol. 19, no. 3, pp. 2899-2928, Sept. 2022.

[0054] Read File; and from The file contains the maximum number of time slots belonging to the TT stream (denoted as n) and the maximum time slot length of the TT stream (denoted as n). ), the period of the gating list (denoted as g), and the phase offset (denoted as s).

[0055] In this invention, the maximum number of time slots for continuously arriving TT streams is [number]. ,and .

[0056] In this invention, the shortest time required to reach a time slot with k TT flows within one period g is: ,and .

[0057] Within one period g, there is The shortest time required for a TT stream time slot is .

[0058] Within one period g, there is The shortest time required for a TT stream time slot is .

[0059] The fast arrival curve model refers to the model where, at time t, any network node... The arrival curve of the TT flow at the output port is represented as follows: When the remainder of time t divided by period g falls within the interval ( When ) in, that is , When the remainder of time t divided by period g falls within the interval ( When ) in, that is , .

[0060] In the rapid arrival curve modeling of TT traffic in incremental scenarios, this invention does not directly rely on a complete global scheduling table. Instead, it characterizes the arrival characteristics of TT traffic based on the upper bound of available time slot resources under the time-triggered mechanism. Specifically, the arrival curve of TT traffic on the node output port is constructed using the maximum number of time slots and the maximum time slot length as key parameters. The core idea of ​​this construction is that the occupation of link and queue resources by TT traffic often manifests as a periodic or quasi-periodic "exclusive / high-priority" transmission window in engineering. When only the upper bound information of the time slot size is known (the maximum number of time slots that can be occupied and the longest duration of a single time slot), the strongest occupation pattern of TT traffic on the available service capacity of AVB traffic on the same port can be given. Compared with traditional methods that require parsing or enumerating a complete global scheduling table, this approach significantly reduces the modeling and data acquisition costs. It can still quickly obtain the worst-case description of TT traffic occupation intensity even when scheduling details are incomplete or the scheduling table is large, thus providing a calculable, reusable, and computationally inefficient input for subsequent deduction of AVB flow service curves (or equivalent residual service characterization).

[0061] S3, a network computation model for AVB streams is constructed based on the fast arrival curve of TT traffic;

[0062] This step characterizes the queuing behavior of AVB flows in incremental scenarios from two aspects: arrival and service. On the arrival side, an upper bound for aggregated arrival traffic is established based on the leaky bucket parameters after ATS shaping. On the service side, a lower bound for available service is established by combining the constraints of TAS gating scheduling on available transmission slots and the adjustment effect of CBS credit mechanism on bandwidth allocation. The complementary description of the arrival upper bound and service lower bound transforms the complex temporal coupling relationship under multi-service competition into a processable relationship between arrival and service curves in network computation, thus providing a unified and formally derivable mathematical framework for worst-case latency analysis.

[0063] Step 301: Construct an arrival curve model for the AVB stream;

[0064] This sub-step aggregates and models multiple AVB flows within the same priority queue. Since the AVB flows are constrained into a leaky bucket shape by ATS before entering the shared queue, the class-level arrival curve can be obtained by summing the leaky bucket parameters of a single flow. This approach preserves the bursty superposition effect while avoiding flow-by-flow temporal enumeration, making it suitable for incremental and rapid evaluation scenarios.

[0065] At time t, any network node The priority at the output port is (i is the AVB traffic priority identifier, with a value of...) The arrival curves for AVB flow aggregation are obtained by summing the arrival curves of individual flows. Since the flows have already been shaped into a standard leaky bucket model by ATS before reaching the shared queue, the rate parameter is... The burst parameters are Therefore, the arrival curve of the AVB stream is ,and .

[0066] It is an integer queue.

[0067] For nodes The priority at the output port is A shared queue.

[0068] For the integer queue Any one of the traffic flows.

[0069] The process by which ATS shapes traffic into a standard leaky bucket model is referenced in "Quantitative Performance Comparison of Various Traffic Shapers in Time-Sensitive Networking," L. Zhao, P. Pop and S. Steinhorst, Section IV-C, IEEE Transactions on Network and Service Management, vol. 19, no. 3, pp. 2899-2928, Sept. 2022.

[0070] Step 302: Construct a service curve model for AVB streams based on the fast arrival curve of TT streams;

[0071] This sub-step quantifies the minimum service capacity available to the AVB stream queue at the port. The model comprehensively considers TT stream gating occupancy, CBS credit upper and lower bounds, and the blocking effects caused by different priority frame lengths, thus obtaining a service lower bound expression that more closely resembles the actual scheduling mechanism. The service curve and arrival curve of this AVB stream jointly determine the upper limit of queue backlog and waiting time.

[0072] At the node port In this context, considering the TAS's occupation of the transmission window, the upper and lower bounds of the CBS credit mechanism, and the blocking effect of low-priority large frames, a class is constructed. The lower bound of the service curve. Specifically, first estimate the available service deduction items (such as...) based on the TT traffic fast arrival curve. Figure 2 (As shown), then introduce CBS's credit growth rate. and credit consumption rate The relevant constraints are used to calculate the upper and lower bounds of the credit, ultimately yielding a service curve expression that meets the requirements of worst-case analysis.

[0073] In priority , For ease of explanation, the current priority is denoted as... Located in the The subsequent priority is denoted as In this invention, the priority is lower than... The maximum frame length of the traffic is denoted as Priority is The maximum frame length of the traffic is denoted as .

[0074] The maximum frame length of the BE stream (excluding TT and AVB streams) in a Time-Sensitive Network is denoted as . .

[0075] Priority The lower bound of the credit value of AVB traffic is denoted as ,and Priority The upper bound of the credit value of AVB traffic is denoted as ,and . Priority The lower bound of the credit value of AVB traffic.

[0076] At time t, any network node At the output port, class The service curve of AVB traffic is ,and ,in, .

[0077] This step provides a unified characterization of the queuing behavior of the shared queue of AVB streams at the output port in an incremental scenario, considering both the arrival and service processes. On the arrival side, leveraging the leaky bucket characteristics formed by ATS shaping of each AVB stream before it enters the shared queue, the leaky bucket parameters of multiple AVB streams with the same priority are aggregated and summed to construct... The class-level arrival curve provides the upper bound of aggregated arrival traffic. On the service side, considering factors such as TT flow gating occupancy, TAS segmentation and constraints on available transmission windows, upper and lower bounds of the CBS credit mechanism, and non-preemptive blocking caused by the maximum frame length of different priorities, the available service capacity is deducted and constrained by combining the fast TT arrival curve obtained based on the upper bound parameter of TT traffic slots. This constructs the lower bound of the service curve obtainable by the AVB flow queue at the port. The above arrival upper bound and service lower bound are expressed in the form of arrival curves and service curves in network calculus, providing key input for subsequent calculation of the worst-case port delay by the maximum horizontal distance and hop-by-hop accumulation along the path to obtain the end-to-end worst-case delay upper bound.

[0078] S4, calculate the end-to-end delay of the AVB stream;

[0079] This step, based on the horizontal distance principle in network computation, transforms the aforementioned curve model into an interpretable latency metric in an incremental scenario. First, the worst-case latency at the port level is calculated, then the worst-case end-to-end latency is obtained by accumulating hop-by-hop along the path.

[0080] Step 401: Calculate the worst-case delay time of the AVB stream's output port;

[0081] In port-level analysis, worst-case latency is given by the maximum horizontal distance between the arrival curve and the service curve. Its physical meaning is "the maximum waiting time required for a service to go from entering the queue to receiving sufficient service under the most unfavorable competition conditions." This metric directly reflects the scheduling pressure and service stress of the current port and provides a standardized tiered latency input for subsequent end-to-end accumulation.

[0082] At time t, any network node The priority at the output port is (i is the AVB traffic priority identifier, with a value of...) Calculate the arrival curve of the AVB flow rate. With service curve Maximum horizontal distance between The Also a priority The worst-case latency of AVB traffic at the output port. It can reflect the degree of port congestion and scheduling contention, and can be used to locate network bottleneck ports.

[0083] Step 402: Calculate the worst-case end-to-end latency of the AVB stream;

[0084] This sub-step sums the worst-case latency times of the traffic at each output port along the path to obtain the end-to-end worst-case latency time. This upper bound is verifiable and comparable, and can be used for optimal solution selection under different topologies and traffic configurations. In incremental scenarios, it can also be used to quickly assess whether the access of a new flow violates existing latency constraints, thereby supporting online reconstruction and admission decisions.

[0085] According to network calculus theory, Assign to . This refers to the bandwidth of the AVB stream. The worst-case delay time.

[0086] Traffic The end-to-end delay is denoted as ,and .therefore, By traffic path The worst-case delay time of each node's output port is summed to obtain the result.

[0087] The core of this step is to convert the previously constructed curve model into a quantifiable latency metric: at each switch output port, the worst-case queuing / waiting time upper bound of the AVB flow at that port is calculated using the maximum horizontal distance in network calculation (the horizontal deviation of the arrival curve relative to the service curve); then, according to the actual forwarding path of the AVB flow, the port-level upper bounds of each hop output port on the path are added together hop by hop to obtain the end-to-end worst-case latency upper bound of the entire path in the incremental scenario.

[0088] Example 1

[0089] The case used in this experiment involves a TSN network topology consisting of 4 switches and 8 end systems, such as... Figure 1 As shown, the connection points between the end system and the switch, and between switches themselves, are all considered as one output port. The switches are commercial switches supporting Time-Sensitive Networking Protocol (TSRP). The end system includes terminals such as computers, sensors, and displays in the airborne network. A computer is used to install the code and program for delay analysis of TSN network AVB traffic in incremental scenarios. This computer is configured with Windows 7 or later, at least 16GB of RAM, and at least 50GB of hard drive space; it comes pre-installed with the integrated development environment Matlab R2015a, Visual Studio Code 1.108.1, the iproute2 toolset, and the net-tools network toolkit. The configured display has a refresh rate of 60Hz or higher.

[0090] The case study includes 20 AVB streams from each of three priority classes: Class 1, Class 2, and Class 7. First, the worst-case latency of the AVB streams is calculated using this method, and the results are as follows. Figure 3 As shown. During the test, a scenario in which the number of TT streams gradually increases was set as an incremental scenario. In this scenario, the latency analysis method of this method and the traditional network calculus method were used respectively (the latency analysis method of the traditional network calculus method can be found in the literature "Performance Evaluation of 5G Wireless Hybrid Network Architecture in Civil Aircraft", Feng Youlin, He Feng et al., Acta Aeronautica Sinica, 2023, 44 (12) Sections 2.1-2.5). The cumulative calculation time of the two methods was counted and a comparison graph was drawn as shown. Figure 4 As shown in the figure, it can be seen that as the number of TT streams increases, the calculation time of traditional delay analysis methods is significantly longer than that of the method of this invention. The delay calculation time of AVB streams using the method of this invention is reduced by more than 80%, thus proving that the method of this invention can reduce the delay calculation time.

[0091] To address the technical challenges of traditional network computation and analysis in incremental airborne TSN networks, which rely on complete network information and scheduling tables, have high data acquisition and computational complexity, and struggle to support rapid recalculation of AVB flows after TT traffic changes, this invention proposes a worst-case latency analysis method for AVB flows as a technical means. Without analyzing the full global scheduling details, a fast TT arrival curve is constructed using the time slot upper bound parameter of TT traffic to characterize the occupancy of available services by TT traffic. Based on the leaky bucket parameters after ATS shaping, AVB flows of the same priority are aggregated to obtain class-level arrival curves. Simultaneously, a lower bound for the service curve is constructed by integrating factors such as TAS gating, CBS credit upper and lower bounds, and maximum frame length blocking for different priorities. Finally, the worst-case latency upper bound for each port is obtained from the maximum horizontal distance between the arrival curve and the service curve, and accumulated along the path to obtain the end-to-end worst-case latency upper bound. The resulting technical effects include: significantly reducing the strong dependence on the entire network's global traffic and the complete gating schedule table and improving the analysis efficiency in incremental scenarios; providing a conservative upper bound on latency when the parameters and construction rules are consistent; port-level results can be decomposed to locate bottlenecks, and calculation results can be reused or partially updated when the maximum number of time slots and the upper bound of the maximum time slot length of TT traffic remain unchanged, thereby reducing repeated full-scale solutions.

Claims

1. A delay analysis method for AVB traffic in TSN networks for incremental scenarios, characterized in that... The steps are as follows: S1, Read the association information of the TSN network topology; Read and organize TSN network and traffic information to obtain node information, topology connections, and port speeds; Read TT traffic information, including at least the maximum number of TT traffic time slots and the maximum time slot length; Read AVB traffic information, as well as the path, priority, and frame length of any AVB stream, and count the AVB traffic flowing through each node in the TSN network; S2, Construct a fast arrival curve model for TT traffic; A fast arrival curve for TT traffic occupancy at the node output port is established based on the maximum number of TT traffic slots and the maximum slot length. The fast arrival curve model refers to the model where, at time t, any network node... The arrival curve of the TT flow at the output port is represented as follows: When the remainder of time t divided by period g falls within the interval At that time, that is , When the remainder of time t divided by period g falls within the interval... At that time, that is , ; The shortest time required to reach a time slot with k TT streams within a period g is: ; Within a period g, it reaches [the desired state]. The shortest time required for a TT stream time slot is ; Within a period g, it reaches [the desired state]. The shortest time required for a TT stream time slot is ; S3, a network computation model for AVB streams is constructed based on the fast arrival curve of TT traffic; Step 301: Construct an arrival curve model for the AVB stream; At time t, any network node The priority at the output port is The arrival curves for AVB stream aggregation are obtained by summing the arrival curves of individual streams. The arrival curve model for AVB streams is as follows: ,and ; For integer queues; For nodes The priority at the output port is Shared queue; For integer queue Any one of the traffic flows; Step 302: Construct a service curve model for AVB streams based on the fast arrival curve of TT streams; At time t, any network node At the output port, class The service curve model for AVB traffic is ,and ,in, ; S4, calculate the end-to-end delay of the AVB stream; Step 401: Calculate the worst-case delay time of the AVB stream's output port; At time t, any network node The priority at the output port is Calculate the arrival curve of the AVB flow rate. With service curve The maximum horizontal distance between them is The Also a priority The worst-case latency of AVB traffic at the output port; Step 402: Calculate the worst-case end-to-end latency of the AVB stream; Traffic The end-to-end delay is denoted as ,and .

2. The delay analysis method for TSN network AVB traffic in incremental scenarios according to claim 1, characterized in that: In S2, the maximum number of time slots for continuously arriving TT flows involved in the fast arrival curve model is: ,and ; The shortest time required to reach a time slot with k TT streams within a period g is: ,and .

3. The delay analysis method for TSN network AVB traffic in incremental scenarios according to claim 1, characterized in that: The priority involved in the AVB stream service curve model in S302 The lower bound of the credit value of AVB traffic is denoted as ,and Priority The upper bound of the credit value of AVB traffic is denoted as ,and . Priority The lower bound of the credit value of AVB traffic; For CBS's credit growth rate; For CBS credit consumption rate; Priority lower than The maximum frame length of the traffic is denoted as ; Priority is The maximum frame length of the traffic is denoted as ; The maximum frame length of the BE stream (excluding TT and AVB streams) in a Time-Sensitive Network is denoted as . .

Images