Multiple variable paths and variable bandwidth scheduling method based on high performance network

Abstract

The present application relates to a method for scheduling multiple variable paths and variable bandwidth based on a high-performance network. The method prioritizes the transmission of more urgent requests based on their priorities, and transmits more user requests within a transmission cycle. The method uses as many node-disjoint paths as possible to transmit data on the network, avoiding congestion caused by excessive load on a certain link, making more effective use of the entire network, improving overall bandwidth utilization, and increasing data transmission reliability. The method also takes into account the delay of path switching, making bandwidth scheduling more practical.

Description

Technical Field

[0001] The present application relates to the field of computer network technology, and in particular to a method for scheduling multiple variable paths and variable bandwidths based on a high-performance network. Background Art

[0002] In the era of big data, more and more data generated in science, industry, and commerce need to be analyzed and processed in a timely manner. High-performance networks (HPNs) with high-speed transmission capabilities have gradually become an effective solution for the scientific and industrial communities.

[0003] High-performance networks can use bandwidth reservation to provide dedicated transmission channels for data transmission, transferring data from its source to remote collaboration sites. Utilizing precious bandwidth resources is a crucial issue for data transmission, directly impacting network utilization and transmission efficiency. Many scientific applications, in particular, require flexible and schedulable network services. High-performance networks can use bandwidth schedulers to calculate and reserve paths, but bandwidth scheduling requires efficient scheduling algorithms to dispatch bandwidth reservation requests submitted by users. The quality of bandwidth scheduling algorithms, to a certain extent, determines network resource utilization and the success rate of request scheduling. Prior art researchers have explored bandwidth reservation strategies under various user models, designed algorithms under various constraints, and studied numerous issues related to data transmission using bandwidth reservation. Most of these issues focus on data transmission on one or two disjoint paths in dynamic high-performance networks. However, with the rapid growth of applications across various fields requiring large data transmission, there is a need to expand bandwidth reservation from one or two paths in high-performance networks to multiple paths to improve data transmission efficiency and reliability. Summary of the Invention

[0004] In order to overcome at least one deficiency in the prior art, the present application provides a method for scheduling multiple variable paths and variable bandwidths based on a high-performance network.

[0005] In a first aspect, a method for scheduling multiple variable paths and variable bandwidths based on a high-performance network is provided, comprising:

[0006] Step 1: sort multiple user requests to obtain a sorted request sequence;

[0007] Step 2: Select the first user request in the request sequence as the current request i. Set the earliest starting time slot of the current request i as the current time slot j. Pre-schedule the current request i starting from the current time slot j and determine the impact r of the path switching delay τ on data transmission in the current time slot j.

[0008] Determine the delayed data volume Q and the delay ratio P based on the relationship between the impact value r of the path switching delay τ in the current time slot j on data transmission and the total impact value prerecord of the path switching delay τ in the previous time slot j-1 on data transmission.

[0009] Step 3: Determine the path with the minimum bandwidth from the source node to the destination node of the current request i in the current time slot j as the current path; determine whether the current path meets the current request based on the amount of delayed data Q and the delay ratio P. If not, execute step 4; if so, execute step 5;

[0010] Step 4: Update the impact of the path switching delay on data transmission in the current time slot j, and update the data volume of the current request i in the current time slot j. Delete the current path in the high-performance network graph G and return to step 3 until no path in the current time slot j satisfies the current request. Record the impact of the path switching delay on data transmission in the current time slot j at this time as the total impact of the path switching delay τ on data transmission in the current time slot j, prerecord. Return to step 2 and set the next time slot j+1 as the current time slot. If no path in the last time slot of the current request i satisfies the current request i, the scheduling of the current request i fails. Execute step 6.

[0011] Step 5: Update the data volume of the current request i in the current time slot j. The current request i is successfully scheduled. Go to step 6.

[0012] Step 6: Restore the deleted path in the high-performance network graph G, return to step 2, and continue scheduling the next user request in the request sequence.

[0013] In one embodiment, in step 2, determining the impact value r of the path switching delay τ on data transmission in the current time slot j includes:

[0014] Step 21: Determine the path with the minimum bandwidth from the source node to the target node of the current request i in the current time slot j as the current path l k If b k ·T j <δ i , where b k For path l k Bandwidth, T j is the duration of the current time slot j, δ i is the data volume of the current request i, then the current path l k If the current request is not satisfied, execute step S22; otherwise, if the current path satisfies the current request, execute step S23;

[0015] Step S22, calculate the current path l k The size of data that can be reserved D k =bk ·T j , update the impact value r of the path switching delay τ on data transmission in the current time slot j k , r k =r k-1 +D k , where r k-1 For the previous path l k-1 The updated impact of the path switching delay τ on data transmission in the current time slot j is k is the first path selected, i.e. k = 1, r k-1 is the initialization value; update the data volume δ of user request i in the current time slot j ik , δ ik =δ i(k-1) -D k , where δ i(k-1) For the previous path l k-1 Update and obtain the data volume of user request i in the current time slot j;

[0016] Delete the current path l in the high-performance network graph G k , continue to search for the next path with the smallest bandwidth as the current path, and return to step S21; until there is no path that meets the current request in the current time slot j, record the impact value r of the path switching delay on data transmission in the current time slot j k , as the impact value r of the path switching delay τ on data transmission in the current time slot j;

[0017] Step 23, calculate the current path l k The size of data that can be reserved D k =b k ·T j , update the impact value r of the path switching delay τ on data transmission in the current time slot j k , r k =r k-1 +D k , where r k-1 The impact of the path switching delay τ on data transmission in the current time slot j obtained by updating the previous path. If the current path l k is the first path selected, i.e. k = 1, r k-1 is the initialization value; record the impact value r of the path switching delay on data transmission in the current time slot j k , as the impact value r of the path switching delay τ on data transmission in the current time slot j.

[0018] In one embodiment, in step 2, determining the amount of delayed data Q and the delay ratio P based on the relationship between the impact value r of the path switching delay τ in the current time slot j on data transmission and the total impact value prerecord of the path switching delay τ in the previous time slot j-1 on data transmission includes:

[0019] If the current time slot j is the first time slot, then P = 1, Q = 0;

[0020] If the current time slot j is not the first time slot, if r>prerecord, then P=1, Q={[T j-1 -τ] / T j-1}·prerecord; if r≤prerecord, then Q=1, P={[T j -τ] / T j}; where T j-1 is the duration of time slot j-1, T j is the duration of time slot j, and τ is the path switching delay.

[0021] In one embodiment, determining whether the current path meets the current request based on the amount of delayed data Q and the delay ratio P includes:

[0022] If b m ·P·T j -Q<δ i , the current path does not satisfy the current request, otherwise, the current path satisfies the current request; where b m is the bandwidth of the current path, T j is the duration of the current time slot j, δ i The amount of data for the current request i.

[0023] In a second aspect, a computer-readable storage medium is provided, which stores a computer program. When the computer program is executed by a processor, it implements the above-mentioned multiple variable path and variable bandwidth scheduling method based on a high-performance network.

[0024] In a third aspect, a computer program product is provided, including a computer program / instruction, which implements the above-mentioned high-performance network-based multiple variable path and variable bandwidth scheduling method when the computer program / instruction is executed by a processor.

[0025] Compared with the prior art, this application has the following beneficial effects:

[0026] 1. This application transmits requests based on priority, giving priority to more urgent requests and delivering more user requests in one transmission cycle.

[0027] 2. This application uses as many non-intersecting paths as possible to transmit data in the network, avoiding congestion caused by excessive load on a certain link, making more full use of the entire network, improving the overall bandwidth utilization, and increasing the reliability of data transmission.

[0028] 3. This application takes into account the delay of path switching, making bandwidth scheduling more practical. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] The present application may be better understood by referring to the following description in conjunction with the accompanying drawings, which together with the following detailed description are incorporated into and form a part of this specification. In the drawings:

[0030] Figure 1 shows a network topology diagram;

[0031] Figure 2 The following figure shows the comparison of the scheduling success rate (SSR) of the method of the present application and the existing method under different network sizes;

[0032] Figure 3 The following figure shows the comparison of the scheduling success rate SSR of the present application method and the existing method under different request numbers;

[0033] Figure 4 A comparison chart of the scheduling success rate SSR of the present application method and the existing method under different network sizes and different request numbers is shown. DETAILED DESCRIPTION

[0034] Exemplary embodiments of the present application are described below with reference to the accompanying drawings. For the sake of clarity and conciseness, not all features of actual embodiments are described in this specification. However, it should be understood that in the process of developing any such actual embodiment, many implementation-specific decisions may be made to achieve the developer's specific goals, and these decisions may vary from one implementation to another.

[0035] It is also necessary to explain here that, in order to avoid obscuring the present application due to unnecessary details, the accompanying drawings only show the device structure closely related to the solution according to the present application, while other details that are not closely related to the present application are omitted.

[0036] It should be understood that the present application is not limited to the described embodiments due to the following description with reference to the accompanying drawings. In this document, where feasible, the embodiments may be combined with each other, features between different embodiments may be replaced or borrowed, and one or more features may be omitted in one embodiment.

[0037] The present invention provides a method for scheduling multiple variable paths and variable bandwidths based on a high-performance network, including:

[0038] Step 1: sort multiple user requests to obtain a sorted request sequence.

[0039] Here, user request i can be expressed as in, is the source node of user request i; is the destination node of user request i; δ i is the amount of data requested by user i, δ i >0; is the start time of user request i, is the deadline for user request i, p i The priority of user request i;

[0040] To sort multiple user requests, you can first sort them according to user priority from high to low. For user requests with the same user priority, sort them in ascending order according to the requested data size, that is, requests with smaller data size are placed in front. For requests with the same data size, sort them in ascending order according to duration, where duration is That is, requests with shorter duration are queued first.

[0041] Here, multiple time windows in the high-performance network are sorted according to the starting time points and the ending time points to obtain a time slot sequence.

[0042] Step 2: Select the first user request in the request sequence as the current request i. Set the earliest starting time slot of the current request i as the current time slot j. Pre-schedule the current request i starting from the current time slot j and determine the impact r of the path switching delay τ on data transmission in the current time slot j.

[0043] The delayed data volume Q and the delay ratio P are determined based on the relationship between the impact value r of the path switching delay τ in the current time slot j on data transmission and the total impact value prerecord of the path switching delay τ in the previous time slot j-1 on data transmission.

[0044] Step 3: Determine the path with the minimum bandwidth from the source node to the destination node of the current request i in the current time slot j as the current path; determine whether the current path meets the current request based on the delay data volume Q and the delay ratio P. If not, execute step 4; if so, execute step 5.

[0045] Here, determining whether the current path meets the current request based on the delayed data volume Q and the delay ratio P may include:

[0046] If b m ·P·T j -Q<δi , then the current path does not meet the current request, and the current path is recorded as R m , otherwise, the current path satisfies the current request; where b m is the bandwidth of the current path, T j is the duration of the current time slot j, δ i The amount of data for the current request i.

[0047] Step 4: Update the impact of the path switching delay on data transmission in the current time slot j, and update the data volume of the current request i in the current time slot j. Delete the current path in the high-performance network graph G and return to step 3 until no path in the current time slot j satisfies the current request. Record the impact of the path switching delay on data transmission in the current time slot j at this time as the total impact of the path switching delay τ in the current time slot j on data transmission prerecord. Return to step 2 and use the next time slot j+1 as the current time slot. When no path in the last time slot of the current request i satisfies the current request i, the scheduling of the current request i fails. Execute step 6.

[0048] Here, the current path is recorded as R m , represents the path selected for the mth time, if the current path R m is the path selected for the first time, the impact of the path switching delay on data transmission in the current time slot j is the initialization value, and the data volume of the current request i is the initialization value. If the current path R m This is not the first selected path, and the updated impact value is represented as prerecord m , prerecord m =prerecord m-1 +δ im , prerecord m-1 The updated impact value of the path selected for the m-1th time; the updated data volume of the current request i is expressed as δ im , δ im =δ i(m-1) -b m ·P·T j -Q, where δ i(m-1) The data volume of the current request i obtained by updating the path selected for the m-1th time, b m is the current path R m Bandwidth, T j is the duration of the current time slot j.

[0049] Step 5: Update the data volume of the current request i in the current time slot j. The current request i is scheduled successfully. Execute step 6.

[0050] Step 6: Restore the deleted path in the high-performance network graph G, return to step 2, and continue scheduling the next user request in the request sequence.

[0051] In one embodiment, in step 2, determining the impact value r of the path switching delay τ on data transmission in the current time slot j may include:

[0052] Step 21: Determine the path with the minimum bandwidth from the source node to the target node of the current request i in the current time slot j as the current path l k If b k ·T j <δ i , where b k The current path l k Bandwidth, T j is the duration of the current time slot j, δ i is the data volume of the current request i, then the current path l k If the current request is not satisfied, step S22 is executed; otherwise, the current path satisfies the current request, step S23 is executed; here, the Dijkstra algorithm may be used to determine the path with the minimum bandwidth.

[0053] Step S22, calculate the current path l k The size of data that can be reserved D k =b k ·T j , update the impact value r of the path switching delay τ on data transmission in the current time slot j k , r k =r k-1 +D k , where r k-1 The impact of the path switching delay τ on data transmission in the current time slot j obtained by updating the previous path. If the current path l k is the first path selected, i.e. k = 1, r k-1 is the initialization value; update the data volume δ of user request i in the current time slot j ik , δ ik =δ i(k-1) -D k , where δ i(k-1) For the previous path l k-1 Update and obtain the data volume of user request i in the current time slot j;

[0054] Delete the current path l in the high-performance network graph G k , continue to search for the next path with the smallest bandwidth as the current path, and return to step S21; until there is no path that meets the current request in the current time slot j, record the impact value r of the path switching delay on data transmission in the current time slot j k, as the impact value r of the path switching delay τ on data transmission in the current time slot j;

[0055] Step 23, calculate the current path l k The size of data that can be reserved D k =b k ·T j , update the impact value r of the path switching delay τ on data transmission in the current time slot j k , r k =r k-1 +D k , where r k-1 The impact of the path switching delay τ on data transmission in the current time slot j obtained by updating the previous path. If the current path l k is the first path selected, i.e. k = 1, r k-1 is the initialization value; record the impact value rk of the path switching delay under the current time slot j on data transmission as the impact value r of the path switching delay τ under the current time slot j on data transmission.

[0056] In one embodiment, in step 2, determining the amount of delayed data Q and the delay ratio P based on the relationship between the impact value r of the path switching delay τ in the current time slot j on data transmission and the total impact value prerecord of the path switching delay τ in the previous time slot j-1 on data transmission includes:

[0057] If the current time slot j is the first time slot, then P = 1, Q = 0;

[0058] If the current time slot j is not the first time slot, if r>prerecord, P=1, Q={[T j-1 -τ] / T j-1}·prerecord;

[0059] If r≤prerecord, Q=1, P={[T j -τ] / T j}

[0060] Among them, T j-1 is the duration of time slot j-1, T j is the duration of time slot j, and τ is the path switching delay.

[0061] In order to better explain the method of this application, Figure 1 The network topology diagram is shown in Figure 1 The user request in Table 1 is transmitted on the network topology diagram shown, and the specific processing is as follows:

[0062] Table 1

[0063] <![CDATA[v s ]]> <![CDATA[v d ]]> δ(Gb) <]]> P i＝1 a d 10 [0，3] 1 i＝2 a d 15 [2，4] 2 i＝3 b d 7 [2，3] 2

[0064] Sorting the three user requests in Table 1, user request 1 has a higher priority than user request 3 and user request 2, and user request 3 transmits less data than user request 2, so the sorted request sequence is user request 1, user request 3, and user request 2.

[0065] User requests are processed in sequence according to the order of the request sequence. Table 2 shows Figure 1 The bandwidth of each time slot in the network topology.

[0066] User Request 1

[0067] In Table 2, we search for time slots with constant bandwidth as 0, 1, 2, 3, and 4. We start with time slot 0 and reserve the current network links and the bandwidth of each link. Specifically, we search for a path from time slot 0 and use the Dijkstra algorithm to find the path v a -v b -v c -v d , the path bandwidth b1=9Gb / s, at this time b1·P·(T1-T0)-Q<δ1, the current path is not enough to meet the data transmission of user request 1, r=9,δ1=10-9=1Gb, update v a -v b -v c -v d The bandwidth is used and this path is deleted. At this time, there is no extra path to continue transmitting data in the 0 time slot.

[0068] Restore the network link and bandwidth. At this time, the current time slot j is the first time slot, so Q = 0, p = 1;

[0069] Find the path from time slot 0 and find the path v a -v b -v c -v d , path bandwidth b1 = 9 Gb / s, at this time b1·P·(T1-T0)-Q<δ1, the current path is insufficient to satisfy user request 1;

[0070] Update δ1 = 10-9 = 1Gb and delete v from the graph a -v b -v c -v d , there is no extra path to continue transmitting data in time slot 0, at this time δ1>0, record prerecord=9+0=9 and continue to the next time slot.

[0071] Continue with the pre-bandwidth from time slot 1 and use Dijkstra's algorithm to find the path v a -v c -vd , the path bandwidth b1 = 6 Gb / s. At this time, b2·(T2 - T1)>δ1, and the current path can satisfy data transmission. r = 6, δ1 = 1 - 9 = -8;

[0072] Perform real reservation, restore the network link and the bandwidth of each link. At this time, since r < prerecord, calculate P = 0.9, and find the path v from slot 1 using Dijkstra's algorithm a -v c -v d , the path bandwidth b1 = 6 Gb / s. At this time, since b1·P·(T2 - T1)-Q = 6×0.9 = 5.4>δ1, the current path can transfer and complete the data, the successful scheduling count +1, and restore the deleted path.

[0073] User request 3

[0074] Start reservation from the starting slot 2, save the current network link and the bandwidth of each link; perform pre-reserved bandwidth to find the path v a -v b -v d , the path bandwidth b1 = 9 Gb / s. At this time, b1·(T3 - T2)<δ2 is not enough to satisfy data transmission, b1·(T3 - T2)<δ2, update the path v a -v b -v d of the bandwidth and delete this path, continue to find the path v a -v c -v d , the path bandwidth b2 = 13 Gb / s. At this time, b2·(T3 - T2)>δ1 can transmit the data completely, r = 22.

[0075] Restore the network link and the link bandwidth. At this time, the current slot j is the first slot, then Q = 0, p = 1;

[0076] Perform real reservation, find the path v at slot 2 a -v b -v d , at this time b1·P·(T3 - T2)-Q < δ1, the current path is not enough to satisfy user request 3;

[0077] Update δ2 = 15 - 9 = 6 Gb, delete v in the figure a -v b -v d and update the link bandwidth, continue to find the path v a -v<00001​, the path bandwidth b2 = 13 Gb / s, at this time b1·P·(T3-T2)-Q=13>δ2, the data can be transmitted completely, the number of successful scheduling +1, and the deleted path is restored.

[0078] User Request 2

[0079] Start reserving from the starting time slot 2, save the current network link and the bandwidth of each link; pre-reserve bandwidth to find the path v a -v c -v d , path bandwidth b1 = 7 Gb / s, at this time b1·(T3-T2)>δ3 can transmit all the data, r = 7;

[0080] Restore the network link and bandwidth. At this time, the current time slot j is the first time slot, so Q = 0, p = 1;

[0081] Find path v in time slot 2 a -v c -v d , at this time b1·P·(T3-T2)-Q=7=δ3, the current path can meet the data transmission request;

[0082] Update δ3 = 7 - 7 = 0 and delete v from the graph a -v c -v d The link bandwidth is updated, the number of successful scheduling is increased by 1, and the deleted path is restored.

[0083] In the above, all three user requests are processed, and the success rate is SSR = total number of successfully scheduled requests = 33 = 100%.

[0084] The present application also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, it implements the above-mentioned multiple variable path and variable bandwidth scheduling method based on a high-performance network.

[0085] The present application also provides a computer program product, including a computer program / instruction, which implements the above-mentioned high-performance network-based multiple variable path and variable bandwidth scheduling method when the computer program / instruction is executed by a processor.

[0086] In order to further analyze the performance of the method proposed in this application, simulation experiments were conducted, and the method FBR-MDVPVB of this application was compared with the existing MinD-MDVPVB and MaxB-MDVPVB.

[0087] First, we create random topology networks with different numbers of nodes and links, as shown in Table 3. In the simulation, we set the total number of time slots to 60 and the bandwidth of each link to follow a normal distribution: where bmax The data rate is 80Gb / s, and x is a random variable between (0, 1]. 100-1500 user requests are randomly generated in the time slot of [0, 60s], and each user request is expressed as Randomly generate source nodes and destination node δ i is a random integer in the interval [1GB, 100GB], starting at time t s Take a random value in [0,30s], and end at time t e In [t s ,60s] and pick a random value.

[0088] Table 3

[0089]

[0090] Figure 2 The following figure shows a comparison of the scheduling success rate SSR (Schuduling success radio) of the present invention and the existing method under different network sizes. When the network size changes, 500 random user requests are set in the 12 different network sizes in Table 3 to compare the scheduling success rate SSR of the three methods. Figure 2 shown.

[0091] Figure 3 The following figure shows the scheduling success rate (SSR) comparison of the method of the present application and the existing method under different request numbers. The performance of the three algorithms is tested when the amount of data requested by the user changes. 100-1500 random user requests are scheduled in a large-scale network with 150 nodes and 600 links, and the scheduling success rate (SSR) of the three methods is compared. Figure 3 shown.

[0092] Figure 4 The following figure shows the scheduling success rate (SSR) comparison of the method of the present application and the existing method under different network sizes and different request numbers. The performance of the three algorithms is tested when the network size and the number of user requests change simultaneously. 100-1500 random user requests are scheduled in 6 different random large-scale networks, and the scheduling success rate (SSR) of the three methods is compared. Figure 4 shown.

[0093] according to Figure 2-Figure 4 It can be seen that the scheduling success rate of the method of the present application is better than that of the existing methods.

[0094] The above descriptions are merely examples of various embodiments of the present application, but the scope of protection of the present application is not limited thereto. Any modifications or substitutions that can be readily conceived by a person skilled in the art within the technical scope disclosed in the present application should be included within the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

Claims

1. A method for scheduling multiple variable paths and variable bandwidth based on a high-performance network, characterized in that: include: Step 1: sort multiple user requests to obtain a sorted request sequence; User request i is represented as ,in, is the source node of user request i; The destination node for user request i; is the amount of data requested by user i, ; is the start time of user request i, ; is the deadline for user request i, ; The priority of user request i; Sort multiple user requests by user priority from high to low. For user requests with the same priority, sort them in ascending order according to the requested data size, that is, requests with smaller data size are placed in front. For requests with the same data size, sort them in ascending order according to duration, where duration is , that is, requests with shorter duration are placed in front; Step 2: Select the first user request in the order of the request sequence as the current request i, set the earliest starting time slot of the current request i as the current time slot j, pre-schedule the current request i starting from the current time slot j, and determine the path switching delay in the current time slot j. The impact value r on data transmission; According to the path switching delay in the current time slot j The impact value r on data transmission and the path switching delay in the previous time slot j-1 Total impact on data transmission The size relationship determines the amount of delayed data and delay ratio P; Step 3: Determine the path with the smallest bandwidth from the source node to the target node of the current request i in the current time slot j as the current path; and the delay ratio P to determine whether the current path meets the current request, if not, execute step 4; if so, execute step 5; Step 4: Update the impact value of the path switching delay on data transmission in the current time slot j, and update the data volume of the current request i in the current time slot j; delete the current path in the high-performance network graph G, return to step 3, until there is no path in the current time slot j that meets the current request, record the impact value of the path switching delay on data transmission in the current time slot j at this time, and use it as the path switching delay in the current time slot j. Total impact on data transmission , return to step 2, set the next time slot j+1 as the current time slot, and when there is no path that satisfies the current request i in the last time slot of the current request i, the scheduling of the current request i fails; execute step 6; The current path is recorded as , indicating the m The path selected this time, if the current path is the path selected for the first time, the impact of the path switching delay on data transmission in the current time slot j is the initialization value, and the data volume of the current request i is the initialization value. If the current path This is not the first selected path, and the updated impact value is expressed as , , For the m- The influence value obtained by updating the selected path once, is the updated data volume of the current request i, ,in, For the m -1 selected path update for the current request i, For the current path bandwidth, is the duration of the current time slot j; Step 5: Update the data volume of the current request i in the current time slot j, and the current request i is successfully scheduled; then execute step 6; Step 6: Restore the deleted path in the high-performance network graph G, return to step 2, and continue to schedule the next user request in the request sequence; Among them, in step 2, the path switching delay in the current time slot j is determined The impact value r on data transmission includes: Step 21: Determine the path with the smallest bandwidth from the source node to the target node of the current request i in the current time slot j as the current path ;like ,in, For path bandwidth, is the duration of the current time slot j, is the data volume of the current request i, then the current path If the current request is not satisfied, step S22 is executed; otherwise, if the current path satisfies the current request, step S23 is executed; Step S22: Calculate the current path The size of data that can be reserved , update the path switching delay in the current time slot j Impact on data transmission , ,in, For the previous path Updated path switching delay in current time slot j Impact value on data transmission, if the current path is the first path selected, i.e. k =1, is the initialization value; update the data volume of user request i in the current time slot j , ,in, For the previous path Update and obtain the data volume of user request i in the current time slot j; Delete the current path in the high-performance network graph G , continue to search for the next path with the smallest bandwidth as the current path, and return to step S21; until there is no path in the current time slot j that meets the current request, record the impact value of the path switching delay on data transmission in the current time slot j at this time , as the path switching delay in the current time slot j Impact on data transmission r ; Step 23: Calculate the current path The size of data that can be reserved , update the path switching delay in the current time slot j Impact on data transmission , ,in, The path switching delay in the current time slot j obtained by updating the previous path Impact value on data transmission, if the current path is the first path selected, i.e. k=1, is the initialization value; records the impact of the path switching delay on data transmission at the current time slot j , as the path switching delay in the current time slot j The impact value r on data transmission; Among them, in step 2, according to the path switching delay in the current time slot j Impact on data transmission r and the path switching delay of the previous time slot j-1 Total impact on data transmission The size relationship determines the amount of delayed data and delay ratio P, including: If the current time slot j is the first time slot, then P =1, Q =0; If the current time slot j is not the first time slot, if ,but P =1, ;like ,but Q =1 ，P ;in, Time slot duration, For the current time slot duration, is the path switching delay; In step 3, according to the amount of delayed data and the delay ratio P to determine whether the current path meets the current request, including: if ， The current path does not satisfy the current request; otherwise, the current path satisfies the current request; wherein, is the bandwidth of the current path, is the duration of the current time slot j, The amount of data for the current request i.

2. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, it implements the method for scheduling multiple variable paths and variable bandwidths based on a high-performance network according to claim 1.

3. A computer program product, characterized in that The method comprises a computer program / instruction, which, when executed by a processor, implements the method for scheduling multiple variable paths and variable bandwidth based on a high-performance network as claimed in claim 1.

Images