Priority-aware scheduling method for trigger flow in avionics internal communication network

Abstract

The invention relates to the field of communication networks, and discloses a priority sensing scheduling method for trigger streams in an avionics internal communication network, which comprises the following steps: S1, constructing a first-order logic constraint model for conflict-free transmission of the trigger streams; and S2, sorting trigger streams based on priorities. And S3, an incremental scheduling strategy. According to the priority perception scheduling method for the trigger flow in the avionics internal communication network, priority scheduling of the high-priority TT flow is realized on the basis of meeting conflict-free transmission of the TT flow, and the problems of low solving efficiency and the like caused by multiple decision variables and high coupling between constraints in a large-scale AICN are solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of communication networks, in particular to a priority-aware scheduling method for triggering flows in an avionics intra-communication network. BACKGROUND

[0002] Avionics system covers the electronic systems of aircraft communication, navigation, management, control, etc. and the network system supporting information exchange and resource sharing among systems. With the progress of microelectronic technology, avionics system has developed from the initial discrete architecture, through the joint and integrated modular architecture, to the recent distributed integrated modular (DIMA) architecture. Under the DIMA architecture, the cooperation among various airborne subsystems is closer, thus higher performance requirements are put forward for the avionics intra-communication network (AICN). At present, time-triggered Ethernet (TTE) has become an ideal solution to meet these needs due to its determinism, low latency and high reliability. In the TTE network, three types of traffic are carried: time-triggered (TT), rate-constrained (RC) and best-effort (BE). Among them, TT flow has the characteristics of fixed period generation, strict delay requirement and zero jitter, and is widely used in flight control, navigation and control, etc. However, different TT flows carry tasks with different safety criticality levels and real-time constraints, for example, the priority of flight control command is usually much higher than that of navigation data update or regular status report. The existing scheduling techniques represented by satisfiability modulo theories solver (SMT), mixed integer programming (MLP) and heuristic algorithm, mostly adopt centralized static scheduling strategy, that is, a complete schedule table is generated once under the premise of knowing all TT flow constraints. This kind of method has high computational complexity and poor flexibility when facing large-scale TT traffic, and it is difficult to meet the needs of complex heterogeneous and large-scale communication scheduling in actual avionics system. In addition, this kind of method usually takes the conflict-free transmission of TT flow as the main target, and the scheduling basis mainly includes traffic period, transmission time and worst transmission delay, etc. parameters, without fully considering the priority difference of TT flow in actual application. Especially in the avionics intra-communication network, it is impossible to ensure that critical task traffic is scheduled first, resulting in difficulty in guaranteeing the timeliness of transmission of critical control information and the real-time performance of the system.

[0003] Therefore, on the basis of ensuring that all TT flows meet the worst delay constraint and conflict-free transmission, how to design a priority-aware scheduling method so that high-priority TT flows can obtain scheduling resources first has become a key challenge in the design of AICN scheduling. SUMMARY

[0004] In view of the above problems existing in the prior art, the purpose of the present application is to provide a priority-aware scheduling method for trigger flows in an avionics internal communication network.

[0005] To achieve the above purpose, the present application adopts the following technical solutions:

[0006] A priority-aware scheduling method for trigger flows in an avionics internal communication network, comprising:

[0007] S1. Constructing a first-order logic constraint model for conflict-free transmission of trigger flows, the first-order logic constraint model comprising: multi-cycle conflict-free constraints, path-dependent constraints, and end-to-end delay constraints;

[0008] S2. Priority-based trigger flow ordering, comprising:

[0009] According to the priority from high to low, the priority of each trigger flow is sorted;

[0010] S3. Incremental scheduling strategy, comprising:

[0011] All trigger flows to be scheduled are divided into multiple batches according to priority, and scheduling is performed in batches; in each round of scheduling, only trigger flow sets of the same priority are processed; then the current round of scheduling results are used as constraint conditions for subsequent rounds to advance the scheduling process.

[0012] Compared with the prior art, the priority-aware scheduling method for trigger flows in an avionics internal communication network according to some embodiments of the present disclosure can bring beneficial technical effects. For example, the priority-aware scheduling method for trigger flows in an avionics internal communication network according to some embodiments of the present disclosure realizes the priority scheduling of high-priority TT flows on the basis of meeting the conflict-free transmission of TT flows. For another example, the priority-aware scheduling method for trigger flows in an avionics internal communication network according to some embodiments of the present disclosure solves the problem of low solving efficiency caused by the large number of decision variables and high coupling between constraints in large-scale AICN. BRIEF DESCRIPTION OF DRAWINGS

[0013] Other features, objects and advantages of the present application will become more apparent from the following detailed description of non-limiting embodiments, made with reference to the accompanying drawings:

[0014] Figure 1A network topology architecture diagram of an aerial device is shown according to some embodiments of the present disclosure;

[0015] Figure 2 A sub-topology architecture diagram of an aerial device is shown according to some embodiments of the present disclosure;

[0016] Figure 3 A number of time slots occupied by TT flow completion transmission is shown according to some embodiments of the present disclosure;

[0017] Figure 4 A corresponding scheduling result Gantt chart is shown according to some embodiments of the present disclosure;

[0018] Figure 5 A comparison of the number of successful solutions of the centralized scheduling method and the incremental scheduling strategy according to some embodiments of the present disclosure under different numbers of trigger flows is shown;

[0019] Figure 6 A comparison of the initial solution generation time of the centralized scheduling method and the incremental scheduling strategy according to some embodiments of the present disclosure under different numbers of trigger flows is shown. DETAILED DESCRIPTION

[0020] The present application will be described in detail below with specific embodiments. The following embodiments will help those skilled in the art to further understand the present application, but do not limit the present application in any form. It should be noted that for those skilled in the art, without departing from the concept of the present application, a number of modifications and improvements can be made. These all belong to the protection scope of the present application.

[0021] Figure 1 A network topology architecture diagram of an aerial device is shown according to some embodiments of the present disclosure.

[0022] As Figure 1 shown, an aerial device includes 9 switch nodes (Switch, denoted as SW1, SW2…SW9), N end systems (End system, denoted as v i ), each of which generates a TT flow (denoted as f i ). The TT flow starts from the source terminal, is forwarded through a series of switches, and is finally transmitted to the destination terminal.

[0023] All nodes in the AICN communicate in full-duplex mode, i.e. any node can simultaneously transmit and receive at the same time. Each physical communication link connecting two nodes is defined as a “data flow link”. The set of data flow links L can be represented as:

[0024]

[0025] where V represents the set of all nodes.

[0026] In AICN, to ensure the deterministic transmission of all TT flows, a static scheduling method is used to pre-allocate the transmission time slot of each flow at each hop in the network, so as to ensure the conflict-free transmission of all TT flows in the network. This requires the system to globally plan the routing path, transmission delay, and transmission time of all TT flows in the scheduling stage. In AICN, the path of a TT flow starts from the source terminal and ends at the destination terminal, and the shortest path strategy is used for path planning. For any TT flow f i in AICN, it can be represented by the following six-tuple:

[0027]

[0028] where f i .period represents the generation period of f i , f i .birth represents the generation time of the first message of f i , f i .length represents the transmission delay of a TT flow, offset represents the transmission time of f i on the link [v k ,v l ], f i .route represents the transmission path of the TT flow, and f i .priority represents the priority of each TT flow.

[0029] In some embodiments of the present disclosure, the priority-aware scheduling method of the trigger flow in the avionics internal communication network can include: S1. constructing a first-order logic constraint model for conflict-free transmission of the trigger flow. The first-order logic constraint model includes: multi-cycle conflict-free constraint, path-dependent constraint, and end-to-end delay constraint.

[0030] In some embodiments of the present disclosure, the multi-cycle conflict-free constraint includes that for any two trigger flows f i ,f j , data cannot be transmitted to the same data flow link at the same time.

[0031] In some embodiments of the present disclosure, the multi-cycle conflict-free constraint can be represented as:

[0032]

[0033] where F = {f1, f2, …, f N} represents the set of TT flows in AICN, and g = gcd(f i .cycle, fj .cycle), i.e. the greatest common divisor of two TT flows, e1∈N, e2∈(f i .length, g - f i .length).

[0034] In some embodiments of the present disclosure, the path-dependent constraint includes that for any one trigger flow, in any two adjacent data flow links in the routing path, it must satisfy that the transmission offset time of the latter data flow link is strictly greater than the transmission offset time on the former data flow link plus the transmission delay of the data packet of the trigger flow on the data flow link.

[0035] In some embodiments of the present disclosure, the path-dependent constraint is that for any one TT flow f i in the AICN, its transmission times of the adjacent two data links [v k , v l ] and [v l , v m ] should satisfy the correct time logic constraint. The path-dependent constraint can be expressed as:

[0036]

[0037] In some embodiments of the present disclosure, the end-to-end delay constraint includes that for any one trigger flow, the end-to-end delay must satisfy that the transmission offset time on the terminal data flow link minus the transmission offset time on the starting data flow link plus the transmission delay of the data packet of the trigger flow on the data flow link must be strictly less than the maximum end-to-end delay allowed by the application requirement.

[0038] In some embodiments of the present disclosure, for any one TT flow f i in the AICN, its end-to-end delay cannot exceed the worst end-to-end delay allowed by the application requirement (denoted as maxDelay i ), and the constraint is expressed as:

[0039]

[0040] Wherein, first(f i .route) represents the starting link connected with the source terminal, and last(f i .route) represents the terminal link connected with the destination terminal. The present disclosure constrains that any one TT flow must complete transmission before the next TT flow is generated, i.e. maxDelay i = f i .period.

[0041] After the TT flow constraint modeling is completed, to enhance the guarantee ability of the scheduling strategy for critical traffic, the disclosure introduces a priority-based TT flow ordering mechanism. Therefore, in some embodiments of the disclosure, the priority-aware scheduling method of the triggered flow in the avionics internal communication network can further include: S2. Priority-based triggered flow ordering. The priority-based triggered flow ordering includes ordering the priority of each triggered flow from high to low according to the priority.

[0042] In some embodiments of the disclosure, first, according to the priority f i .priority of each TT flow, the priorities are ordered from high to low (in some embodiments of the disclosure, the highest priority is recorded as 1).

[0043] In some embodiments of the disclosure, for triggered flows with the same priority, an urgency indicator Urgency i is introduced as the basis for ordering.

[0044] In some embodiments of the disclosure, the urgency indicator is based on the worst-case delay of the triggered flow and the optimal transmission time.

[0045] In some embodiments of the disclosure, for triggered flows with the same priority, based on the worst-case end-to-end delay of any one time-triggered flow, subtracting the product of the total number of hops of its transmission path and its transmission delay, the urgency indicator is obtained:

[0046] Urgency i =maxDelay i -length(f i .route)*f i .length

[0047] Wherein, length(f i .route) represents the total number of hops of the transmission path of the i-th TT flow.

[0048] In some embodiments of the disclosure, the urgency indicator reflects the remaining margin between the ideal case where the TT flow does not occur waiting and its worst-case end-to-end delay. The smaller the value, the smaller the scheduling margin, the more urgent the flow, which should be processed first. For example, when two TT flows have the same priority, the system will preferentially schedule the flow with longer path and smaller worst-case end-to-end delay to reduce the potential conflict probability in the transmission process.

[0049] In some embodiments of the present disclosure, a first-order logical constraint model of TT flow conflict-free transmission is established by comprehensively considering constraints such as node conflict and end-to-end worst delay. In the scheduling process, first, according to the priority order of the TT flow, the flow with higher criticality is preferentially guaranteed; for the flows with the same priority, further combined with the remaining hop number, the single-hop transmission delay, and the difference between the worst delay and the current scheduling time, the scheduling priority is sorted, so as to improve the scheduling success rate of high urgency traffic.

[0050] After the constraint modeling and priority sorting of the TT flow are completed, in order to effectively cope with the problems such as high centralized scheduling calculation complexity, long solving time, and easy to appear non-convergence in large-scale AICN, the present disclosure proposes an incremental scheduling strategy. Therefore, in some embodiments of the present disclosure, the priority-aware scheduling method of the triggered flow in the avionics internal communication network can further include: S3. The incremental scheduling strategy.

[0051] In some embodiments of the present disclosure, the incremental scheduling strategy includes: dividing all the triggered flows to be scheduled into multiple batches according to the priority, and scheduling in batches; in each round of scheduling process, only the triggered flow set of the same priority is processed; then the current round scheduling result is taken as the constraint condition of the subsequent round, and the scheduling process is gradually promoted.

[0052] In some embodiments of the present disclosure, the incremental scheduling strategy can include:

[0053] S31. According to the number of triggered flows corresponding to each priority, a priority-scheduling step mapping table Sp is constructed, and according to the scheduling step, the start and end position pointer variables Begin and End of the triggered flow set in each round of scheduling are set;

[0054] S32. In each round of scheduling, a batch of triggered flows under the current priority is selected to form a scheduling subset F temp ;

[0055] S33. Based on the first-order logical constraint model, combined with the scheduled flow set F sch (initially, ), a decision variable set Var temp (initially, ) is constructed, and a constraint condition is added to the decision variable set;

[0056] S34. After adding all the constraint conditions, the decision variable set Var temp is imported into the Gurobi solver for solving, if the solving is successful, the scheduling result is incorporated into the scheduled flow set, and the pointer start and end position pointer variables Begin and End are updated, and the next round of scheduling is continued, and the cycle is repeated until all the flows are scheduled.

[0057] In some embodiments of this disclosure, the set of decision variables is not an empty set (i.e. If the answer is ), then the solution is successful.

[0058] In some embodiments of this disclosure, the pseudocode for the incremental scheduling strategy is shown in Alg.1.

[0059]

[0060]

[0061] in,

[0062] In some embodiments of this disclosure, the incremental scheduling strategy effectively alleviates the scheduling bottleneck problem of centralized scheduling in large-scale network environments by adopting a "batch processing and gradual advancement" approach, reducing the overall solution complexity while ensuring priority awareness. Through the incremental scheduling strategy, the scale of each round of scheduling is controllable and the computational overhead is limited, achieving both priority scheduling of critical tasks and increasing the scale of schedulable TT flows.

[0063] Figure 2 A diagram of an aviation equipment sub-topology architecture according to some embodiments of the present disclosure is shown.

[0064] In some embodiments of this disclosure, Figure 2 Taking the sub-topology of the aviation equipment network topology as an example, 10 TT flows are randomly generated, with priority ranging from 1 to 4, as shown in Table 1.

[0065] Table 1. Detailed parameter information for 10 TT streams (sorted)

[0066]

[0067] Figure 3 The amount of time slot resources required for a TT stream to complete transmission according to some embodiments of this disclosure is shown. Figure 4 A Gantt chart showing the corresponding scheduling results according to some embodiments of this disclosure is shown. Figure 3 The horizontal axis represents the trigger flow index (1-10), corresponding to... Figure 4 (Flow1-Flow10 as identified in the text).

[0068] like Figure 3 and Figure 4 As shown, in AICN, the time slot resources occupied by decision variables do not overlap when any TT stream is transmitted on each hop link, and a strict time dependency relationship is satisfied between adjacent hops. Furthermore, for each data stream link, the higher-priority TT stream is always sent first. Figure 3The total number of time slots occupied by the 10 TT streams is 32, which is the same as... Figure 4 The number of time slot resource blocks matches, further verifying the correctness of the scheduling result. It is evident that the priority-aware scheduling method for triggered flows in the avionics internal communication network according to some embodiments of this disclosure can achieve priority scheduling of high-priority traffic while ensuring conflict-free transmission of TT flows in AICN.

[0069] In some embodiments of this disclosure, Figure 1 Taking the network topology of the aviation equipment shown (9 switch nodes and 18 terminal nodes) as an example, comparative experiments were conducted in TT flow scenarios of different scales.

[0070] Figure 5 The results show a comparison of the number of successful solutions obtained by the centralized scheduling method and the incremental scheduling strategy according to some embodiments of this disclosure for different numbers of TT streams.

[0071] Figure 6 The diagram shows a comparison of the initial solution generation time between a centralized scheduling method and an incremental scheduling strategy according to some embodiments of this disclosure, under different TT flow numbers.

[0072] like Figure 5 As shown, with the increase in the number of TT flows, the success rate of the traditional centralized scheduling method decreases significantly, and it becomes impossible to complete the scheduling solution when the number of TT flows reaches 140. In contrast, the incremental scheduling strategy according to some embodiments of this disclosure maintains high solution stability under the same conditions, with only a slight decrease in the number of successful solutions. Figure 6 As shown, the incremental scheduling strategy according to some embodiments of this disclosure significantly outperforms the centralized method at all scales, and the difference in solution time between the two methods widens as the number of TT flows increases. It is evident that the priority-aware scheduling method for trigger flows in avionics internal communication networks according to some embodiments of this disclosure, by introducing an incremental scheduling strategy, solves the problems of low solution efficiency caused by numerous decision variables and high coupling between constraints in large-scale AICNs. The incremental scheduling strategy schedules TT flows in batches after prioritization. The scheduling result of each round is fed back as a constraint to subsequent scheduling stages, effectively reducing the complexity of a single solution and improving scheduling convergence speed and feasibility.

[0073] The specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art can make various modifications or variations within the scope of the claims, which do not affect the essence of the present invention.

Claims

1. A priority-aware scheduling method for trigger flows in an avionics internal communication network, characterized in that, include: S1. Construct a first-order logical constraint model for triggering conflict-free transmission of the stream, wherein the first-order logical constraint model includes: multi-cycle conflict-free constraint, path dependency constraint and end-to-end latency constraint. S2. Priority-based trigger stream sorting, including: Sort the priority of each trigger stream from high to low. S3. Incremental scheduling strategy, including: All trigger flows to be scheduled are divided into multiple batches according to priority, and scheduled sequentially by batch. In each round of scheduling, only the set of trigger flows with the same priority is processed. Then, the scheduling result of the current round is used as the constraint condition for the subsequent rounds to advance the scheduling process.

2. The priority-aware scheduling method for trigger flows in an avionics internal communication network according to claim 1, characterized in that, The multi-cycle conflict-free constraint includes the requirement that for any two triggering flows, data must not be transmitted to the same data flow link at the same time.

3. The priority-aware scheduling method for trigger flows in an avionics internal communication network according to claim 1, characterized in that, The path dependency constraint includes the following: for any triggering flow in any two adjacent data flow links in the routing path, the following must be satisfied: the transmission offset time of the latter data flow link is strictly greater than the transmission offset time of the former data flow link plus the transmission delay of the data packet of the triggering flow on the data flow link.

4. The priority-aware scheduling method for trigger flows in an avionics internal communication network according to claim 1, characterized in that, The end-to-end latency constraint includes the requirement that the end-to-end latency of any triggering flow must meet the following condition: the transmission offset time on the endpoint data flow link minus the transmission offset time on the starting data flow link plus the transmission latency of the data packet of the triggering flow on the data flow link must be strictly less than the maximum end-to-end latency allowed by the application requirements.

5. The priority-aware scheduling method for trigger flows in an avionics internal communication network according to claim 1, characterized in that, For triggering flows with the same priority, an urgency index is introduced as the sorting criterion. The urgency index is obtained based on the worst-case delay and the best transmission time of the triggering flow.

6. The priority-aware scheduling method for trigger flows in an avionics internal communication network according to claim 5, characterized in that, For triggering flows with the same priority, the urgency index is obtained by subtracting the product of the total number of hops in the transmission path and the transmission delay from the worst end-to-end delay of any time-triggered flow.

7. The priority-aware scheduling method for trigger flows in an avionics internal communication network according to claim 1, characterized in that, The incremental scheduling strategy includes: S31. Based on the number of trigger flows corresponding to each priority, construct a priority-scheduling step size mapping table, and set the start and end position pointer variables of the trigger flow set in each round according to the scheduling step size; S32. In each round of scheduling, select a batch of trigger flows under the current priority to form a scheduling subset; S33. Based on the first-order logic constraint model, and combined with the set of scheduled traffic, construct a set of decision variables, and add constraints to the set of decision variables; S34. Import the set of decision variables into the Gurobi solver for solving. If the solution is successful, merge the scheduling result into the set of scheduled traffic, update the start and end pointer variables, and continue the next round of scheduling. Repeat until all traffic is scheduled.

8. The priority-aware scheduling method for trigger flows in an avionics internal communication network according to claim 7, characterized in that, The set of scheduled traffic is initially empty, and the set of decision variables is initially empty.

9. The priority-aware scheduling method for trigger flows in an avionics internal communication network according to claim 7, characterized in that, If the set of decision variables is not empty, then the solution is successful.

Images