Method for avionics network virtual link configuration based on dynamic programming and top-down

Abstract

A kind of avionics network virtual link configuration method based on dynamic programming and top-down, detect IP fragmentation caused by MTU, network bandwidth limit and the jitter limit caused by multiple virtual link propagation, the bandwidth calculation problem of virtual link is modeled into the optimization problem of minimization;According to the limit condition, the initial parameters of virtual link that make the total bandwidth of network minimum are obtained, the message allocation method based on dynamic programming is used when the size of message is small, and the traversal type top-down message allocation method is used when the size of message is large, to obtain the corresponding virtual link and configuration parameters when each message flow is transmitted.The message allocation method with the target of improving bandwidth margin using dynamic programming idea and top-down structure respectively can significantly improve bandwidth margin to adapt to larger scale actual data, and has fast running speed.

Description

Technical Field

[0001] This invention relates to a technology in the field of avionics networks, specifically a virtual link configuration method based on dynamic programming and a top-down approach for Avionics Full Duplex Switched Ethernet (AFDX). Background Technology

[0002] AFDX is a standard protocol defined for data exchange in avionics systems. Figure 2 As shown, an AFDX network mainly consists of terminal systems, switches, and virtual links. A virtual link is a unidirectional logical path from a single transmitting system to a fixed set of receiving systems. Each virtual link occupies a fixed bandwidth. Setting appropriate parameters can reduce the bandwidth consumption of virtual links, which requires determining the message allocation on each virtual link. Existing message allocation methods that split and merge virtual links based on different strategies to improve bandwidth margins do not meet the real-world data scale of an AFDX network where a single terminal system can send more than 128 information streams. Summary of the Invention

[0003] To address the aforementioned shortcomings of existing technologies, this invention proposes a virtual link configuration method for avionics networks based on dynamic programming and a top-down approach. By utilizing dynamic programming and a top-down message allocation method aimed at increasing bandwidth margin, this method can significantly improve bandwidth margin to accommodate larger-scale real-world data and operates at high speed.

[0004] This invention is achieved through the following technical solution:

[0005] This invention relates to a method for configuring virtual links in avionics networks based on dynamic programming and a top-down approach. It detects IP fragmentation caused by MTU, network interface bandwidth limitations, and jitter limitations caused by the propagation of multiple virtual links. The bandwidth calculation problem of virtual links is modeled as a minimization optimization problem. Based on the constraints, the initial parameters of the virtual links that minimize the total network bandwidth are obtained. When the message size is small, a dynamic programming-based grouping algorithm (DPGA) is used; when the message size is large, a traversal top-down algorithm (TTDA) is used to obtain the corresponding virtual link and configuration parameters for each message stream transmission.

[0006] The aforementioned minimization optimization problem refers to an optimization objective that minimizes... Where: set n represents the total number of virtual links; constraints: Where: ni Let B be the number of message flows contained in the i-th virtual link, and let B be the network interface bandwidth, period. i，j payload i，j These represent the interval time of the sending cycle of the j-th message stream in the i-th virtual link and the message size of each cycle, respectively.

[0007] The initial parameters of the virtual link include: Maximum Transfer Unit (MTU) and Minimum Allocation Gap (BAG), where: MTU is the maximum frame size, i.e., IP fragmentation technology causes message flows larger than the MTU to be divided into multiple frames for transmission in the virtual link; BAG is the minimum interval between two consecutive frames, and these initial parameters of the virtual link are obtained by considering all feasible MTUs and BAGs for transmitting each message flow in a virtual link; according to constraints. Due to the IP fragmentation limitations caused by MTU, the MTU that minimizes the total network bandwidth can be obtained by using integer block division; that is, by using the integer block division method to calculate... Find the maximum BAG that satisfies the inequality after taking the union of the MTU sets of each message stream for all possible values. i .

[0008] The dynamic programming-based message allocation method refers to recursively breaking down a large-scale virtual link allocation problem into several smaller, easier-to-solve subproblems, thereby considering more options for allocating messages to different virtual links, so as to minimize the required bandwidth while satisfying jitter and network interface bandwidth constraints.

[0009] The aforementioned traversal-based top-down message allocation method refers to: initializing all messages in a single virtual link, and then allocating messages according to the LP value of the message flow, i.e. The virtual links are split or merged sequentially, and the method with the lowest bandwidth consumption is selected by iterative judgment.

[0010] This invention relates to a system for implementing the above method, comprising: a modeling unit, an integer partitioning unit, and a message allocation unit, wherein: the modeling unit models the bandwidth calculation problem as a minimization optimization problem based on constraints and optimization objectives; the integer partitioning unit performs integer partitioning processing based on the IP fragmentation limitations caused by MTU obtained by the modeling unit, to obtain the MTU and BAG of each virtual link that may minimize the total network bandwidth; the message allocation unit selects dynamic programming message allocation processing or traversal top-down message allocation processing based on the results of the integer partitioning unit, the network interface bandwidth limitations obtained by the modeling unit, and the jitter limitation information caused by the propagation of multiple virtual links obtained by the modeling unit, to obtain the corresponding virtual link and configuration parameters for each message stream transmission.

[0011] Technical effect

[0012] This invention employs an integer block method to obtain the MTU and BAG that minimize the total network bandwidth for each virtual link, and uses either the DPGA method (dynamic programming) or the TTDA method (a top-down, traversal-based message allocation method) to allocate messages based on the message size. Compared to existing technologies, this invention can increase bandwidth margin to accommodate larger-scale real-world data and operates at a faster speed. Attached Figure Description

[0013] Figure 1 This is a flowchart of the invention method;

[0014] Figure 2 This is a schematic diagram of the AFDX network structure;

[0015] Figure 3 A diagram illustrating message flow BAG and jitter;

[0016] Figure 4 This is a schematic diagram of step b in the TTDA method;

[0017] Figure 5 This is a diagram showing the comparison of bandwidth usage in actual experiments;

[0018] In the figure: a) compares the DPGA and TTDA methods with existing methods in experiments on terminal systems with a small number of message streams; b) compares the TTDA method with existing methods in experiments on terminal systems with a large number of message streams.

[0019] Figure 6 This is a diagram illustrating the time consumption of the actual experiment. Detailed Implementation

[0020] like Figure 1 As shown, this embodiment relates to a method for configuring virtual links in avionics networks based on dynamic programming and a top-down approach, including:

[0021] The first step is to model the bandwidth calculation problem as a minimization optimization problem:

[0022] Optimization goal: Minimize Where: set n represents the total number of virtual links;

[0023] Restrictions: Where: n i Let B be the number of message flows contained in the i-th virtual link, and let B be the network interface bandwidth, period. i，j payload i，j These represent the interval time of the sending cycle of the j-th message stream in the i-th virtual link and the message size of each cycle, respectively.

[0024] The second step is to optimize based on the MTU and BAG that minimize the total network bandwidth, according to the constraints: The integer block method is used to obtain the MTU and BAG that minimize the total network bandwidth for each virtual link, specifically including:

[0025] 2.1) When there is only one message stream in the i-th virtual link, the interval time of the message stream's transmission period is period. i，j and the message size payload for each cycle i，j For period, payload;

[0026] 2.2) Iterate continuously to find all possible values ​​of MTU. The iteration method is to initialize the variable m, and each time... Iteration stops when m > payload - 1;

[0027] 2.3) Let S be the set of all the m values ​​mentioned above. Then S contains all possible values ​​of MTU.

[0028] 2.4) For each element MTU in set S i Find the maximum BAG that satisfies the inequality under the constraints. i At this point, all possible (MTU) i BAG i This refers to the MTU and BAG that may minimize the total network bandwidth.

[0029] For the case of multiple message streams, repeat steps 2.2 and 2.3 for each message stream in N, take the union of the resulting N sets S, and then perform step 2.4 to obtain n. i The MTU and BAG that may minimize the total network bandwidth in the case of a single message flow.

[0030] The time complexity of the second step is:

[0031] Third, after the above modeling, using the MTU and BAG of each virtual link obtained in the second step to minimize the total network bandwidth, the problem of rationally allocating messages in virtual links to improve bandwidth margin is transformed into an NP-hard problem, based on which a polynomial-time method can be designed. This embodiment includes two message allocation methods: the DPGA method is used when the message size is small, such as when the number of message streams sent by a single terminal system is less than or equal to 20, and the TTDA method is used when the message size is large.

[0032] The aforementioned message allocation method based on dynamic programming specifically includes:

[0033] i) Initialization: Set the jitter and bandwidth required for 0 messages to 0.

[0034] ii) Use recursion to solve the distribution scheme of n messages, enumerating the 2^n message streams S. n -2 non-empty proper subsets, and for each non-empty proper subset T, solve for the bandwidth and jitter under the corresponding optimal allocation scheme;

[0035] iii) Using the method in step 2, merge all remaining messages in set T that have not been considered into the same virtual link, that is, find the MTU and BAG in the message flow of set ST that minimize the total network bandwidth;

[0036] iv) Enumerate all possible (MTU) in step iii i BAG i ), calculate the required bandwidth and add it to the bandwidth usage of set T. If the constraints of the network interface bandwidth limit and jitter limit are still satisfied, then take all possible MTUs. i BAG i The one with the lowest bandwidth usage among them will be the final answer.

[0037] The time complexity of the dynamic programming-based message allocation method is O(3). n ).

[0038] The aforementioned traversal-based top-down message distribution method specifically includes:

[0039] a) Initialization: Place all message flows in a virtual link and check whether the network interface bandwidth limit and jitter limit are met. If both constraints are met, initialization is complete; otherwise, continuously select the message with the largest LP value and move it into the newly created virtual link until both constraints are met.

[0040] b) such as Figure 4As shown, a virtual link is selected, and its message streams are sorted in descending order of LP value. Messages with LP values ​​exceeding the average are moved to other existing virtual links or newly established virtual links, selecting the case with the lowest bandwidth usage. When all LP values ​​are the same, half of the message streams are selected. When there is only one message, it is equivalent to merging virtual links; the diagram shows three virtual links containing six, three, and two message streams respectively. When processing the first virtual link, if the LP values ​​of the first three messages exceed the average LP value, they are removed and placed into the second, third, and a new virtual link respectively. The case with the lowest bandwidth usage among these three cases and the original case is selected to continue this method.

[0041] The method for comparing bandwidth usage described in step b is implemented as follows: Using all feasible MTU and BAG pairs obtained in step two for each virtual link, based on a greedy algorithm, the MTU and BAG with the smallest bandwidth usage are selected for each virtual link sequentially. The MTU and BAG of unprocessed virtual links are then selected as the parameters with the smallest jitter. The jitter limit is checked once for each virtual link processed. If the jitter limit is not met, the previous virtual link settings are backtracked, sacrificing a small amount of bandwidth to meet the jitter limit.

[0042] c) Repeat step b for the virtual link until the bandwidth usage is at its lowest. In this case, the selected message is kept in the original virtual link. Then start the next virtual link and select the next virtual link to perform step b.

[0043] d) When all virtual links have been processed and the message allocation scheme is no longer updated, the message allocation scheme at this point is the final configuration scheme, and the MTU and BAG of each virtual link are the final parameters of the virtual link.

[0044] The time complexity of the described traversal top-down message distribution method is:

[0045] Through specific practical experiments, in a real-world AFDX network scenario, each terminal system has different components such as host applications and host functions, and each component has different ports to send or receive message streams. The dataset contains 108 terminal systems, each capable of sending 5 to 128 messages. The period of a message stream is 8 to 10,000 milliseconds. The first comparative method, SAPTDA, continuously splits the virtual link containing many message streams into two virtual links to reduce bandwidth consumption. The second comparative method (All In One, AIO) places all messages in a single virtual link. Each method has requirements on the number of message streams sent by the terminal system. The terminal systems were divided into two groups: the first group of 60 terminal systems sent no more than 22 message streams each, used to evaluate DPGA; the second group included all terminal systems that SAPTDA could handle.

[0046] like Figure 5 As shown, the figure compares the bandwidth usage of different methods in actual experiments, with the horizontal axis sorted from largest to smallest bandwidth usage. Figure a compares the DPGA / TTDA method with the AIO / SABTDA method, and figure b compares the TTDA method with the AIO / SABTDA method. It can be seen that both message allocation methods proposed in this embodiment can improve bandwidth margin. Specifically, for the first group of terminal systems, the bandwidth usage rates of AIO, SABTDA, TTDA, and DPGA are 3.45%, 3.20%, 2.76%, and 2.68%, respectively. The bandwidth usage of DPGA and TTDA is reduced by 22.3% and 20.0% compared to AIO, respectively. For the second group of terminal systems, the bandwidth usage rates of AIO, SABTDA, and TTDA are 18.15%, 11.69%, and 10.86%, respectively. The bandwidth usage of TTDA is reduced by 40.2% compared to AIO. Furthermore, TTDA can handle terminal systems with more message streams that SABTDA cannot handle, which also demonstrates the significant value of message allocation methods.

[0047] like Figure 6 The diagram illustrates the time consumption of the TTDA method under different BAG value ranges. The BAG values ​​for TTDA1 are set to powers of 2, while the BAG values ​​for TTDA2 are set to integer multiples of 0.5. The diagram shows that the maximum time consumption for TTDA1 is less than 1.8 seconds, while the maximum time consumption for TTDA2 is 1025.6 seconds. This indicates that the message distribution method runs quickly.

[0048] Compared with existing technologies, this invention can significantly increase bandwidth margin to accommodate larger-scale real-world data and has a faster operating speed.

[0049] The above-described specific implementations can be partially adjusted by those skilled in the art in different ways without departing from the principles and purpose of the present invention. The scope of protection of the present invention is defined by the claims and is not limited to the above-described specific implementations. All implementation schemes within the scope of the claims are bound by the present invention.

Claims

1. A method for configuring virtual links in avionics networks based on dynamic programming and a top-down approach, characterized in that, The study detects IP fragmentation caused by MTU, network interface bandwidth limitations, and jitter limitations caused by propagation through multiple virtual links. It models the bandwidth calculation problem of virtual links as a minimization optimization problem. Based on the constraints, it obtains the initial parameters of the virtual links that minimize the total network bandwidth. When the number of message streams sent by a single terminal system is less than or equal to 20, a dynamic programming-based message allocation method is used; when the message size is large, a traversal-based top-down message allocation method is used. This yields the corresponding virtual link and configuration parameters for each message stream transmission, specifically including: The first step is to model the bandwidth calculation problem as a minimization optimization problem: Optimization goal: Minimize Where: set n represents the total number of virtual links; Restrictions: ,in: Let B be the number of message flows contained in the i-th virtual link, and let B be the network interface bandwidth. They represent the first time. The first virtual link The interval between message stream sending cycles and the message size in each cycle; The second step is to use the Maximum Transmission Unit (MTU) and Minimum Interval Time (BAG) that minimize the total network bandwidth as optimization objectives, based on the following constraints: The integer block method is used to obtain the MTU and BAG of each virtual link that minimize the total network bandwidth, specifically including: 2.1) When the first If there is only one message stream in a virtual link, then the interval of the message stream's sending period is... and the message size for each cycle for ; 2.2) Iterate continuously to find all possible values ​​of MTU. The iterative method is to change the variable... Initialization, each time , until Stop iteration when the time is right; 2.3) Let S be the set of all the m values ​​mentioned above. Then S contains all possible values ​​of MTU. 2.4) For each i element in set S Find the maximum value that satisfies the inequality under the constraints. At this point, all possible These are the MTU and BAG that may minimize the total network bandwidth; For the case of multiple message streams, repeat steps 2.2 and 2.3 for each message stream in N, take the union of the resulting set S, and then perform step 2.4 to obtain... The MTU and BAG that may minimize the total network bandwidth in the case of multiple message flows; Third, after the above modeling, using the MTU and BAG of each virtual link obtained in the second step to minimize the total network bandwidth, the problem of rationally allocating messages in the virtual links to increase bandwidth margin is transformed into an NP-hard problem, and a polynomial-time method is designed accordingly. The aforementioned message allocation method based on dynamic programming specifically includes: i) Initialization: Set the jitter and bandwidth required for 0 messages to be 0. ii) Use recursion to solve the distribution scheme of n messages, enumerating the n message streams. Given a non-empty proper subset T, solve for the bandwidth and jitter under the corresponding optimal allocation scheme for each non-empty proper subset T; iii) Using the method in step ii, merge all remaining messages not considered in set T into the same virtual link, and find the MTU and BAG in the message flow of set S-T that minimize the total network bandwidth; iv) Enumerate all possible steps in step iii Calculate the required bandwidth and add it to the bandwidth usage of set T. If the constraints of the network interface bandwidth limit and jitter limit are still satisfied, then take all possible values. The one with the lowest bandwidth usage will be the final answer; The aforementioned traversal-based top-down message distribution method specifically includes: a) Initialization: Place all message streams in a virtual link and check whether the network interface bandwidth limit and jitter limit are met. If both constraint equations are met, the initialization is complete. Otherwise, continuously select the message with the largest LP value and move it into the newly created virtual link until both constraint equations are met. The LP value is ; b) Select a virtual link, sort the message streams in it in descending order of LP value, select messages with LP values ​​exceeding the average value and move them to other existing virtual links or newly established virtual links, selecting the case with the lowest bandwidth usage; when all LP values ​​are the same, select half of the message streams; when there is only one message, it is equivalent to merging virtual links. c) Repeat step b for the virtual link until the bandwidth usage is at its lowest. In this case, the selected message is kept in the original virtual link. Then start the next virtual link and select the next virtual link to perform step b. d) When all virtual links have been processed and the message allocation scheme is no longer updated, the message allocation scheme at this point is the final configuration scheme, and the MTU and BAG of each virtual link are the final parameters of the virtual link.

2. The avionics network virtual link configuration method based on dynamic programming and top-down approach as described in claim 1, characterized in that, The initial parameters of the virtual link include: Maximum Transmission Unit (MTU) and Minimum Interval (BAG), where: MTU is the maximum frame size; IP fragmentation technology allows message streams exceeding the MTU to be divided into multiple frames for transmission in the virtual link; BAG is the minimum interval between two consecutive frames, and these initial parameters are obtained by considering all feasible MTUs and BAGs for transmitting multiple message streams in a single virtual link; and by applying constraints. Due to IP fragmentation limitations caused by MTU, integer block partitioning is used to obtain the MTU and BAG that minimize the total network bandwidth. The integer block partitioning method is then used to calculate... All possible values ​​and the corresponding maximum satisfying inequality .

3. The avionics network virtual link configuration method based on dynamic programming and top-down approach as described in claim 1, characterized in that, In step b: using all feasible MTU and BAG pairs obtained in step two for each virtual link, based on the idea of ​​a greedy method, select the MTU and BAG with the smallest bandwidth for each virtual link in turn, and select the MTU and BAG of the virtual links that have not yet been processed as the parameters with the smallest jitter; check the jitter limit once for each virtual link processed. If the jitter limit is not met, backtrack the previous virtual link settings and sacrifice a small amount of bandwidth to meet the jitter limit.

4. A system for implementing the avionics network virtual link configuration method based on dynamic programming and top-down approach as described in any of claims 1-3, characterized in that, include: The system comprises a modeling unit, an integer partitioning unit, and a message allocation unit. The modeling unit models the bandwidth calculation problem as a minimization optimization problem based on constraints and optimization objectives. The integer partitioning unit performs integer partitioning based on the IP fragmentation limitations caused by the MTU derived by the modeling unit, obtaining the MTU and BAG that minimize the total network bandwidth for each virtual link. The message allocation unit, based on the results from the integer partitioning unit, the network interface bandwidth limitations derived by the modeling unit, and the jitter limitations caused by the propagation of multiple virtual links derived by the modeling unit, selects either dynamic programming message allocation processing or traversal-based top-down message allocation processing according to the message size, obtaining the corresponding virtual link and configuration parameters for each message stream transmission.

Images