Cooperative scheduling method for deterministic and non-deterministic service flows

Abstract

The invention relates to the technical field of deterministic real-time transmission of computer networks, in particular to a collaborative scheduling method for deterministic and non-deterministic service flows. The method comprises the following steps: S1, deducing an ideal gap for minimizing the influence of a deterministic service flow on a non-deterministic service flow through network calculation, S2, optimizing a scheduling model of the deterministic service flow according to the deduced ideal gap, and constructing a deterministic service flow scheduling optimization model based on the ideal gap, and S3, proposing a heuristic strategy to solve the optimization model. According to the method, the ideal gap of the influence of the deterministic service flow on the non-deterministic service flow is minimized, the scheduling model of the deterministic service flow is optimized according to the deduced ideal gap, the heuristic strategy is provided to solve the optimization model, the optimization model and the heuristic strategy thereof are provided, and under the condition that the burst characteristics of the non-deterministic service flow are not assumed, the scheduling efficiency of the non-deterministic service flow is improved. By optimizing the transmission interval of the deterministic service flow, the service quality of the non-deterministic service flow is guaranteed as much as possible.

Description

Technical Field

[0001] The present invention relates to the technical field of deterministic real-time transmission of computer networks, and in particular to a collaborative scheduling method for deterministic and non-deterministic business flows. Background Art

[0002] Time Sensitive Networking (TSN) is a new type of industrial communication technology that can not only ensure the real-time and deterministic transmission of data in industrial control, but also be fully compatible with the existing Ethernet system. It is an important communication platform supporting the development of the Industrial Internet of Things. TSN accurately specifies the sending time of business flows in network nodes through time-triggered scheduling, thereby ensuring the end-to-end deterministic real-time transmission of business flows. TSN includes three types of business flows: TT flow, AVB flow and BE flow. TT flow is a deterministic business flow, and AVB flow and BE flow are collectively referred to as non-deterministic business flows. The service quality of deterministic business flows is met first and is exclusive, which seriously affects the service quality of non-deterministic business flows. Deterministic business flows use time-triggered transmission, and non-deterministic business flows use CBS (Credit-Based Shaper), ATS (Asynchronous Traffic Shaping) and priority scheduling transmission methods.

[0003] The Chinese patent publication number CN116346627A, "A time-triggered scheduling method, node, electronic device and storage medium", includes: obtaining the characteristics of each business flow in a preset business flow set, the path of the business flow and the characteristics of each preset node on the path of the business flow; constructing a directed graph according to the path of the business flow; determining the urgency of the business flow according to the directed graph, the characteristics of the business flow and the characteristics of the node; determining the sending time of the business flow according to the directed graph and the urgency of the business flow; and sending the sending time to the node.

[0004] In order to ensure the service quality when the three types of business flows are transmitted in a mixed manner, non-deterministic business flows are usually converted into deterministic business flows for scheduling. Deterministic business flows are scheduled first, and then non-deterministic business flows are incrementally scheduled on the basis of satisfying the schedulability of deterministic business flows. This processing method ignores the burst characteristics of non-deterministic business flows. When the traffic changes, the network needs to readjust the scheduling and cannot respond to the changes in burst traffic in time. In addition, the period and message length of the burst traffic obtained by statistical analysis may be too small or too large, causing queuing or waste of reserved bandwidth. Considering the unpredictable burst characteristics of non-deterministic business flows, the problem of deterministic business flow scheduling with the goal of minimizing the impact of deterministic business flows on non-deterministic business flows needs to be solved urgently. Network calculus is a theoretical method for analyzing the end-to-end worst-case delay of business flows in the network. In time-sensitive networks, network calculus is used to construct the arrival curve and service curve model of the business flows and network nodes under study. It is a mature theory for analyzing the end-to-end worst-case delay of the network.

[0005] In summary, developing a collaborative scheduling method for deterministic and non-deterministic business flows is still a key issue that needs to be urgently solved in the field of deterministic real-time transmission technology in computer networks. Summary of the invention

[0006] In view of the problems existing in the prior art, the purpose of the present invention is to provide a method for collaborative scheduling of deterministic and non-deterministic business flows.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] A method for co-scheduling deterministic and non-deterministic business flows, comprising the following steps:

[0009] S1. Derive the ideal gap that minimizes the impact of deterministic business flows on non-deterministic business flows through network calculation;

[0010] S2. According to the derived ideal gap, the scheduling model of the deterministic business flow is optimized, and a deterministic business flow scheduling optimization model based on the ideal gap is constructed;

[0011] S3. Propose a heuristic strategy to solve the optimization model.

[0012] The present invention is further configured as follows: in step S1, it is known through network calculation that the delay is related to the gap, and the optimal solution is obtained when the gaps are equal, that is, the TT flow gap with the smallest end-to-end worst delay of the AVB flow is a uniform gap.

[0013] The present invention is further configured as follows: in step S1, in actual situations, since the TT flow appears periodically, the uniform gap is only an ideal situation. If the AVB end-to-end worst delay is to be minimized, it can be obtained by adjusting the gap between the TT windows to be as close to the optimal solution as possible.

[0014] The present invention is further configured as follows: in step S2, the following steps are included:

[0015] S21. Objective function: Minimize the difference between the gap of the scheduling result and the ideal gap, so that the scheduling result is as close to the ideal gap as possible:

[0016]

[0017] Among them, m is the actual gap of scheduling, and z is the ideal gap, which is the average of the sum of gaps and the number of windows;

[0018]

[0019] S22, gap range constraint: Assume that m represents the gap between windows, M is the gap set, m i represents the gap size before the i-th window, and the gap range constraint is expressed as:

[0020]

[0021]

[0022] S23, time slot selection constraint: This constraint can be formalized as:

[0023]

[0024]

[0025]

[0026] S24, when b≥1, the conditional constraint must be met, ifb≥1, that is, one cycle apart:

[0027]

[0028] The present invention is further configured as follows: in step S23, the time slot selection constraint refers to limiting the service flow sending time to the time slot, each time slot for sending the service flow is separated by a gap, and for a flow sent multiple times within a macro cycle, the interval between two sendings is a fixed value of the cycle.

[0029] The present invention is further configured as follows: in step S3, the heuristic strategy is proposed to solve the optimization model, including the following steps:

[0030] S31, Initialize ideal gap Before implementing the conflict resolution strategy, the ideal clearance value on each edge needs to be calculated:

[0031]

[0032] S32, initializing the time slot exploration starting point;

[0033] S33, exploring conflict-free time slots;

[0034] S34. Narrow the ideal gap and continue exploring.

[0035] The present invention is further configured as follows: in step S32, the following steps are included:

[0036] S321, starting from the service flow with the highest urgency, for each service flow f i , let f i On the side k ,v j The initial sending time on ] is This is the starting point of scheduling, and the length is the business flow f i cycle;

[0037] S322, first determine the sending time of the closest previously scheduled flow, and set the sending time to Determine whether it is smaller than the ideal gap

[0038] S323: If yes, adjust the service flow f i The initial sending time is as follows, which can ensure that the gap with the nearest flow is an ideal gap:

[0039]

[0040] The present invention is further configured as follows: in step S33, the following steps are included:

[0041] S331. Determine the urgency level based on the conflict resolution mechanism. i Is the high business flow, that is, the business flow with a subscript less than i, consistent with f i The sending time conflicts;

[0042] S332, if there is no conflict, then the business flow f i If the scheduling is successful, i=i+1, the next flow is scheduled, or if i==n, all service flows have been scheduled;

[0043] S333. If there is a conflict, the sending time of the high-urgency service flow remains unchanged. i Sending time add That is, move the time slot of the ideal gap:

[0044]

[0045] S334, then continue to determine whether there is a conflict with the scheduled high-urgency flow, f i Sending time The search scope is:

[0046]

[0047] The present invention is further configured as follows: in step S34, the following steps are included:

[0048] S341. If there is still no available time slot after searching once according to step S33, the initial sending time when the first conflict occurs is adjusted to:

[0049]

[0050] S342, continue to explore the time slots in step S33 until the service flow is successfully scheduled, or after If no conflict-free time slot is found after the secondary gap is reduced, the service flow is marked as an unschedulable service flow;

[0051] S343, then i=i+1 to schedule the next flow, or i==n, that is, all service flows have been scheduled.

[0052] Beneficial Effects

[0053] Compared with the known public technology, the technical solution provided by the present invention has the following advantages:

[0054] Beneficial effects:

[0055] (1) In the present invention, an ideal gap for minimizing the impact of deterministic business flows on non-deterministic business flows is derived through network calculation, and the conditions for minimizing the impact of deterministic business flows on non-deterministic business flows are theoretically given. The ideal gap for minimizing the impact of deterministic business flows on non-deterministic business flows is compared with the previous collaborative scheduling method. The present invention models the impact of deterministic business flows on non-deterministic business flows through network calculation, with the goal of minimizing the end-to-end worst delay of non-deterministic business flows under the influence of deterministic business flows to ensure the transmission quality of non-deterministic business flows as much as possible, and gives the conclusion and derivation process of the best ideal gap, solving the problem of providing non-deterministic business flow service quality by optimizing the scheduling of deterministic business flows without assuming the burst characteristics of non-deterministic business flows.

[0056] (2) In the present invention, the scheduling model of the deterministic business flow is optimized according to the derived ideal gap, and an objective function, constraint conditions, and an ideal gap calculation method for minimizing the difference between the actual gap and the ideal gap are proposed. A heuristic strategy is also given to solve the optimization model. The optimization model and its heuristic strategy, without assuming the burst characteristics of the non-deterministic business flow, optimize the transmission interval of the deterministic business flow to ensure the service quality of the non-deterministic business flow as much as possible, providing theoretical and methodological guidance for exploring better deterministic business flow scheduling with better service quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 A method flow chart of a method for collaborative scheduling of deterministic and non-deterministic business flows;

[0058] Figure 2 A schematic diagram of an arrival curve and a service curve of a collaborative scheduling method for deterministic and non-deterministic business flows;

[0059] Figure 3 A schematic diagram of a collaborative scheduling method for deterministic and non-deterministic business flows in which three business flows are sent within one macrocycle. DETAILED DESCRIPTION

[0060] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0061] The present invention will be further described below in conjunction with the embodiments.

[0062] Please refer to Figure 1-Figure 3 As shown, a method for co-scheduling deterministic and non-deterministic business flows comprises the following steps:

[0063] Step 1: Derived the ideal gap that minimizes the impact of deterministic traffic flow on non-deterministic traffic flow through network calculation; Modeled the gap of TT traffic flow, analyzed the worst delay using arrival curve and service curve according to network calculation theory, and then minimized the delay;

[0064] The relevant symbols involved in this section are shown in Table 1 below:

[0065] Table 1 Symbol comparison table

[0066]

[0067] 11) The TT arrival curve and AVB service curve are:

[0068]

[0069]

[0070] 12) Substituting the TT arrival curve into the AVB service curve yields:

[0071]

[0072] 13) Because idSl M -sdSl M =C, then:

[0073]

[0074] 14) Through network calculation theory derivation, it can be known that the forwarding delay Where σ is the service flow length, θ and r are service curve parameters,

[0075] and so is a fixed value, and the forwarding delay depends on the parameter θ, that is:

[0076]

[0077] 15) Therefore, the problem of minimizing the delay can be transformed into min{θ}, that is:

[0078]

[0079] 16) Simplify the constant term in the above formula to obtain:

[0080]

[0081] 17)Canceling the constants yields:

[0082]

[0083] 18) Finally, after complete simplification, we get:

[0084]

[0085] 19) Let the size of the gap introduced before window j be It can also be assumed that the gap is after the window. The present invention takes setting the gap before the window as an example. The following is an equation for the relationship between the window start time, the window length and the gap:

[0086]

[0087]

[0088] Among them, when j = i, That is, the reference zero point is introduced into the variable gap m get:

[0089]

[0090] 110) According to the relationship equation between the window start time, window length and gap, we continue to decompose it and finally get:

[0091]

[0092]

[0093] 111) Since the length of the TT window is a constant, it can be simplified to:

[0094]

[0095] 112) Strategy 1 (time slot strategy): Divide the entire time axis into several time slots, which are the minimum time units for scheduling and allocation. The size of the time slot is sufficient to send a complete service flow message. When the service flow message is an extra-long message, it can be divided into several sub-messages for sending according to the size of the time slot. In this case, as long as the terminal node supports the function of dividing sub-messages, it can be used. Since the time slot is the minimum unit for scheduling and allocation, the minimum forwarding time of the switching node is also an integer multiple of the time slot. During scheduling, all time allocations are allocated according to the size of the time slot. Introducing time slot strategy 1, so Then the above formula is transformed into:

[0096]

[0097] 113) Lemma 1: max(min 0≤i≤n-1 {x0,x1,x2,…,x n-1}) and x0+x1+x2+…+x n-1 =C (C is a constant), then x0 = x1 = x2 = ... = x n-1 Get the optimal solution when

[0098] 114) Since the total gap is a constant, we know that:

[0099] is a constant;

[0100] 115) Then the optimal solution obtained by Lemma 1 is:

[0101]

[0102] In summary, through network calculation derivation, it can be known that the delay is related to the gap, and the optimal solution is obtained when the gaps are equal, that is, the TT flow gap with the smallest end-to-end worst-case delay of the AVB flow is a uniform gap; in actual situations, due to the periodic appearance of the TT flow, the uniform gap is only an ideal situation. If the AVB end-to-end worst-case delay is to be minimized, it can be obtained by adjusting the gap between the TT windows as close to the optimal solution as possible.

[0103] Step 2: According to the derived ideal gap, the scheduling model of the deterministic business flow is optimized, and a deterministic business flow scheduling optimization model based on the ideal gap is constructed, including the following steps:

[0104] S21. Objective function: Minimize the difference between the gap of the scheduling result and the ideal gap, so that the scheduling result is as close to the ideal gap as possible:

[0105]

[0106] Among them, m is the actual gap of scheduling, and z is the ideal gap, which is the average of the sum of gaps and the number of windows;

[0107]

[0108] S22. (Constraint 1) Gap range constraint: Assume that m represents the gap between windows, M is the gap set, and m i represents the gap size before the i-th window, and the gap range constraint is expressed as:

[0109]

[0110]

[0111] S23. (Constraint 2) Time slot selection constraint: The time slot selection constraint means limiting the service flow sending time to the time slot. Each time slot for sending a service flow is separated by a gap. For a flow that is sent multiple times within a macrocycle, the interval between two sendings is a fixed value of the cycle. This constraint can be formalized as:

[0112]

[0113]

[0114]

[0115] S24, when b≥1, the conditional constraint must be met, ifb≥1, that is, one cycle apart:

[0116]

[0117] Step 3: Propose a heuristic strategy to solve the optimization model, including the following steps:

[0118] S31, Initialize ideal gap Before implementing the conflict resolution strategy, the ideal clearance value on each edge needs to be calculated:

[0119]

[0120] S32, initializing the time slot exploration starting point, including the following steps:

[0121] S321, starting from the service flow with the highest urgency, for each service flow f i , let f i On the side k ,v j The initial sending time on ] is This is the starting point of scheduling, and the length is the business flow f i cycle;

[0122] S322, first determine the sending time of the closest previously scheduled flow, and set the sending time to Determine whether it is smaller than the ideal gap

[0123] S323: If yes, adjust the service flow f i The initial sending time is as follows, which can ensure that the gap with the nearest flow is an ideal gap:

[0124]

[0125] S33, exploring a time slot without conflict, comprising the following steps:

[0126] S331. Determine the urgency level based on the conflict resolution mechanism. i Is the high business flow, that is, the business flow with a subscript less than i, consistent with f i The sending time conflicts;

[0127] S332, if there is no conflict, then the business flow f i If the scheduling is successful, i=i+1, the next flow is scheduled, or if i==n, all service flows have been scheduled;

[0128] S333. If there is a conflict, the sending time of the high-urgency service flow remains unchanged. i Sending time add That is, move the time slot of the ideal gap:

[0129]

[0130] S334, then continue to determine whether there is a conflict with the scheduled high-urgency flow, f i Sending time The search scope is:

[0131]

[0132] S34, narrowing the ideal gap and continuing the exploration, including the following steps:

[0133] S341. If there is still no available time slot after searching once according to step S33, the initial sending time when the first conflict occurs is adjusted to:

[0134]

[0135] S342, continue to explore the time slots in step S33 until the service flow is successfully scheduled, or after If no conflict-free time slot is found after the secondary gap is reduced, the service flow is marked as an unschedulable service flow;

[0136] S343, then i=i+1 to schedule the next flow, or i==n, that is, all service flows have been scheduled.

[0137] In the present invention, the collaborative scheduling method of a deterministic and non-deterministic service flow first derives an ideal gap that minimizes the impact of the deterministic service flow on the non-deterministic service flow through network calculation, that is, the ideal gap Theoretically, the conditions for minimizing the impact of deterministic business flows on non-deterministic business flows are given; the ideal gap for minimizing the impact of deterministic business flows on non-deterministic business flows in the present invention: Compared with the previous collaborative scheduling methods, the present invention models the impact of deterministic business flows on non-deterministic business flows through network calculations, with the goal of minimizing the end-to-end worst delay of non-deterministic business flows under the influence of deterministic business flows to ensure the transmission quality of non-deterministic business flows as much as possible, and gives the conclusion and derivation process of the best ideal gap, which solves the problem of optimizing the scheduling of deterministic business flows without assuming the burst characteristics of non-deterministic business flows. The problem of providing service quality for non-deterministic business flows; then based on this condition, according to the derived ideal gap, the scheduling model of deterministic business flows is optimized, and the objective function, constraint conditions, and ideal gap calculation method that minimize the gap between the actual gap and the ideal gap are proposed, and a heuristic strategy is given to solve the optimization model. This optimization model and its heuristic strategy, without assuming the burst characteristics of non-deterministic business flows, optimize the transmission interval of deterministic business flows to ensure the service quality of non-deterministic business flows as much as possible, providing theoretical and methodological guidance for exploring better deterministic business flow scheduling with better service quality;

[0138] There are two specific implementation methods in the present invention, using an ILP solver to solve directly and using a heuristic algorithm based on a heuristic strategy to solve. When using an ILP solver to solve directly, it is necessary to add constraints and an objective function, and use a CPLEX solver to solve. This method can obtain the optimal solution of the proposed optimization model, but because the solution takes a long time, it is suitable for offline solution scheduling. The heuristic algorithm based on the heuristic strategy can achieve fast solution scheduling according to the heuristic strategy.

[0139] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for collaborative scheduling of deterministic and non-deterministic business flows, characterized in that: The following steps are involved: S1. Derive the ideal gap that minimizes the impact of deterministic business flows on non-deterministic business flows through network calculation; S2. According to the derived ideal gap, the scheduling model of the deterministic business flow is optimized, and a deterministic business flow scheduling optimization model based on the ideal gap is constructed; S3. Propose a heuristic strategy to solve the optimization model.

2. The method for collaborative scheduling of deterministic and non-deterministic business flows according to claim 1, characterized in that: In step S1, it can be known through network calculation that the delay is related to the gap, and the optimal solution is obtained when the gaps are equal, that is, the TT flow gap with the smallest end-to-end worst delay of the AVB flow is a uniform gap.

3. The method for collaborative scheduling of deterministic and non-deterministic business flows according to claim 1, characterized in that: In step S1, in actual situations, since the TT flow appears periodically, the uniform gap is only an ideal situation. If the AVB end-to-end worst-case delay is to be minimized, it can be obtained by adjusting the gap between the TT windows to be as close to the optimal solution as possible.

4. The method for collaborative scheduling of deterministic and non-deterministic business flows according to claim 1, characterized in that: In step S2, the following steps are included: S21. Objective function: Minimize the difference between the gap of the scheduling result and the ideal gap, so that the scheduling result is as close to the ideal gap as possible: Among them, m is the actual gap of scheduling, and z is the ideal gap, which is the average of the sum of gaps and the number of windows; S22, gap range constraint: Assume that m represents the gap between windows, M is the gap set, m i represents the gap size before the i-th window, and the gap range constraint is expressed as: S23, time slot selection constraint: This constraint can be formalized as: S24, when b≥1, the conditional constraint must be met, ifb≥1, that is, one cycle apart:

5. The method for collaborative scheduling of deterministic and non-deterministic service flows according to claim 4, characterized in that: In step S23, the time slot selection constraint refers to limiting the service flow sending time to the time slot, each time slot for sending the service flow is separated by a gap, and for a flow sent multiple times within a macro cycle, the interval between two sendings is a fixed value of the cycle.

6. The method for collaborative scheduling of deterministic and non-deterministic business flows according to claim 5, characterized in that: In step S3, the heuristic strategy is proposed to solve the optimization model, including the following steps: S31, Initialize ideal gap Before implementing the conflict resolution strategy, the ideal clearance value on each edge needs to be calculated: S32, initializing the time slot exploration starting point; S33, exploring conflict-free time slots; S34. Narrow the ideal gap and continue exploring.

7. A method for collaborative scheduling of deterministic and non-deterministic business flows according to claim 6, characterized in that: In step S32, the following steps are included: S321, starting from the service flow with the highest urgency, for each service flow f i , let f i On the side k ,v j The initial sending time on ] is This is the starting point of scheduling, and the length is the business flow f i cycle; S322, first determine the sending time of the closest previously scheduled flow, and set the sending time to Determine whether it is smaller than the ideal gap S323: If yes, adjust the service flow f i The initial sending time is as follows, which can ensure that the gap with the nearest flow is an ideal gap:

8. The method for collaborative scheduling of deterministic and non-deterministic service flows according to claim 7, characterized in that: In step S33, the following steps are included: S331. Determine the urgency level based on the conflict resolution mechanism. i Is the high business flow, that is, the business flow with a subscript less than i, consistent with f i The sending time conflicts; S332, if there is no conflict, then the service flow f i If the scheduling is successful, i=i+1, the next flow is scheduled, or if i==n, all service flows have been scheduled; S333. If there is a conflict, the sending time of the high-urgency service flow remains unchanged. i Sending time add That is, move the time slot of the ideal gap: S334, then continue to determine whether there is a conflict with the scheduled high-urgency flow, f i Sending time The search scope is: [Initial sending time, initial sending time + Min (f i .period, maximum end-to-end delay)].

9. A method for collaborative scheduling of deterministic and non-deterministic service flows according to claim 8, characterized in that: In step S34, the following steps are included: S341. If there is still no available time slot after searching once according to step S33, the initial sending time when the first conflict occurs is adjusted to: S342, continue to explore the time slots in step S33 until the service flow is successfully scheduled, or after If no conflict-free time slot is found after the secondary gap is reduced, the service flow is marked as an unschedulable service flow; S343, then i=i+1 to schedule the next flow, or i==n, that is, all service flows have been scheduled.

Images