Cooperative allocation method for computing resources and link resources in satellite-ground network

Abstract

A collaborative allocation method for computing resources and link resources in a satellite-ground network comprises the steps that firstly, a timeline is divided into a plurality of time periods, a directed graph model is constructed in each time period, nodes are low-orbit satellites and ground stations, and sides represent communication links; step 2, defining the weight of the edge as the link bandwidth, and searching the shortest path from the source point to the destination; and calculating the weight of the path again and updating the shortest path by considering the calculation capability of the node. And allocating the bandwidth and computing resources on the path, and updating the remaining resources. The process is cyclically executed until resources are used up or no available path exists. And step 3, path searching and resource allocation are carried out on the satellite-ground network in each time period, and resource allocation on the whole timeline is ensured. According to the method, the multi-task transmission requirement, the satellite computing power and the link capacity limitation are comprehensively considered, and the data processing and transmission throughput of the satellite-ground network is remarkably improved.

Description

Technical Field

[0001] The invention relates to a method for collaboratively allocating computing resources and link resources in a satellite-to-ground network, and belongs to the technical field of satellite communications. Background Art

[0002] The continuous deployment of low-orbit satellite constellations has achieved full coverage of the earth, greatly improving the speed and coverage of data routing. In addition, by deploying computing resources and service function chains on low-orbit satellites, low-orbit satellites have a certain data computing capability, thereby supporting data to be directly processed and calculated on low-orbit satellites during transmission, which can effectively reduce the data transmission burden from satellites to the ground, improve the utilization rate of inter-satellite communication resources, and further optimize network resource allocation to provide users with more efficient and timely services.

[0003] However, existing methods generally transmit the original data to the destination before processing, and these methods have many limitations. First, centralized processing methods can easily cause the computing resources at the destination to become a bottleneck, especially when processing large amounts of data, the processing efficiency will drop significantly. Secondly, all data needs to be transmitted to the destination, which increases the pressure on network bandwidth, prolongs transmission delays, and affects the response speed of the overall system. In addition, this method is highly dependent on the computing resources at the destination and is difficult to adapt to complex and dynamic network environments. If the computing power of the destination is limited or equipment fails, the performance and reliability of the entire system will be seriously affected. The coordinated allocation of computing resources and link resources allows data to be calculated while it is transmitted on low-orbit satellites. In this way, the data can be transmitted to the destination after calculation, reducing the dependence on centralized computing resources at the destination.

[0004] The biggest challenge in achieving data transmission and computing in satellite-to-ground networks on low-orbit satellites is the limited computing power of low-orbit satellites and the limited energy provided by solar panels and batteries. In addition, multiple tasks require efficient data transmission and processing between multiple source points (such as different ground stations or sensors) and multiple sink points (such as specific ground receiving stations or other low-orbit satellites). Such multi-source and multi-sink tasks require the coordination of multiple traffic allocations at the same time. Therefore, considering the needs of multi-task transmission and the limitations of computing power and link capacity of low-orbit satellites, a method for coordinated allocation of computing resources and link resources in satellite-to-ground networks is studied and designed to make full use of the high-speed communication and computing capabilities between satellites to meet the growing data processing and real-time service needs in the future. Summary of the invention

[0005] The technical problem to be solved by the present invention is how to effectively maximize the throughput by considering both data transmission and data calculation in a satellite-to-ground network, and proposes a method for collaboratively allocating computing resources and link resources in a satellite-to-ground network that is oriented to network link bandwidth limitations and satellite node computing power limitations.

[0006] The present invention adopts the following technical solutions to solve the above technical problems:

[0007] The present invention proposes a collaborative allocation algorithm for computing resources and link resources in a satellite-to-ground network, comprising the following steps:

[0008] Step 1: Set a given time line range Divided into N time periods, for the nth time period The satellite-ground network of n∈[1,N] is modeled as a directed graph in is a set of network nodes, including satellites and ground stations, ε n Indicates the time period The collection of all links in the intra-satellite-ground network;

[0009] Step 2: Find the shortest path from the source to the destination for each task according to the link bandwidth, and define the link available bandwidth as a directed graph The length of the corresponding edge in the path is determined by the sum of the available bandwidths of all links on the path. Further considering the available computing resources of the node, the length of the shortest path is reassigned and updated according to the available computing resources of the node, the computing resources and link bandwidth on this shortest path are allocated, and the amount of data to be processed and transmitted on this path is determined. The available bandwidth and computing resources of the links and nodes are updated according to the allocation scheme. This step is repeated until all computing resources are allocated or there is no feasible path to transmit data.

[0010] Step 3: Execute step 2 for the satellite-ground network information in each time period.

[0011] In the step 1:

[0012] Given a timeline range Divide into N time periods. When dividing the time periods, ensure that the topology of the satellite-ground network remains unchanged in each time period. n∈[1, N] satellite-to-ground network modeling, and obtain the directed graph Map each ground station and low-orbit satellite as a node in a directed graph, using Represents a set of network nodes. Each link in the satellite-to-ground network within n∈[1,N] is mapped to a graph The edge in nIndicates the time period Considering that the link status of the satellite-to-ground network may be different in different time periods, the directed graph constructed The edge sets in may also be different. The capacity of the edge is the available bandwidth of the corresponding link, and the capacity of the node is the available computing resources of the corresponding ground station or low-orbit satellite. Assume that there are K tasks that need to be calculated and transmitted in this satellite-ground network, k∈[1,K] represents the kth task, where s k For the task The source point, d k For the task The destination of The raw data of the task is from the source node s k During the transmission process, the low-orbit satellite passing by performs the required processing and calculation, and finally transmits the calculated data to the destination d k .

[0013] In the step 2:

[0014] Step 2.1) Take the capacity of each edge as the length of the edge, and assign the current task Find the source point s k The shortest path to each other node and calculate the distance between two nodes. The distance between two nodes is the sum of the lengths of all edges passed by the shortest path between them.

[0015] Step 2.2) For the s obtained in step 2.1) k The distance to other nodes is taken into account for the node capacity. If the capacity of the node passed by the path is 0, then s k Update the distance to this node to infinity, otherwise update it to the original distance plus the node capacity. k After the distance to each node is updated, select the current and s k The node v* with the shortest distance, traverses the neighboring nodes u of v*, if u is to s k The distance between v* and s is less than k The distance from v* to u is added to the length of the edge, then u is updated to s k The distance and shortest path of v*, otherwise unchanged; after visiting all neighbor nodes of v*, select the next one with s k The node with the shortest distance v*, until v* = d k Or all nodes have been visited.

[0016] Step 2.3) For the shortest path obtained in step 2.2), compare the minimum capacity of all edges on the path with the sum of the capacities of all nodes. If the minimum capacity of all edges on the path is less than the sum of the capacities of all nodes on the path, it means that the bottleneck of this path is the edge with the smallest capacity, and the maximum flow that can be processed and transmitted on the path is the capacity of this edge. The amount of data that each node on the path needs to calculate will be allocated according to the ratio of the capacity of this node to the sum of the capacities of all nodes on the path; otherwise, the flow is the sum of the capacities of all nodes on the path, and the amount of data that each node on the path needs to calculate is its own capacity.

[0017] Step 2.4) Update the capacity of the edges and nodes along the path according to the allocation plan of computing resources and link bandwidth of the shortest path in step 2.3): the edge capacity is updated to the original capacity minus the traffic flowing on this edge, and the node capacity is updated to the original capacity of the node minus the amount of data calculated on this node. Continue to execute steps 2.1)-2.3) for the next task. If all tasks have allocated resources, continue to execute steps 2.1)-2.3) for each task starting from the first task.

[0018] In the step three:

[0019] Let n=n+1, and execute step 2 until n=N. Obtain the allocation scheme of computing resources and bandwidth resources in the satellite-to-ground network and the data transmission path in each time period.

[0020] Compared with the prior art, the present invention adopts the above technical solution and has the following advantages:

[0021] 1. The present invention aims at the inefficiency problem caused by the separation of computing resources and link resources in the data transmission process in the existing satellite-to-ground network, and proposes a method for the coordinated allocation of computing resources and link resources in the satellite-to-ground network. Most of the existing satellite-to-ground network data transmission regards the low-orbit satellite network as a simple relay point. The original data is transmitted to the destination through the low-orbit satellite network and then processed. This method not only causes data redundancy in the transmission process, but also leads to single-point dependence on downlink bandwidth and destination computing resources, increasing the overall burden and bottleneck risk of the system. In order to solve these problems, the present invention proposes to innovatively combine the data calculation and transmission processes, allocate and utilize computing resources during the data transmission process, so that the original data can be distributedly calculated during the transmission process, and the destination directly receives the processed data without the need for subsequent calculations. This method can significantly reduce the pressure on the downlink, reduce the dependence on the destination computing resources, and improve the computing efficiency and transmission performance of the overall network.

[0022] 2. The present invention provides a resource allocation framework suitable for concurrent processing of multiple tasks in a satellite-to-ground network. At present, satellite-to-ground network data transmission is mostly based on a task model with a single starting point and a single end point, that is, the resource allocation problem of processing one source point to one destination each time. However, in a complex real-world application environment, multiple tasks may involve multiple different sources and destinations. Therefore, the present invention innovatively adopts the multi-commodity network flow theory, and introduces additional node computing power and resource constraints on the basis of the multi-commodity network flow theory, abstracts the satellite-to-ground network into a graph model with additional constraints, and on this basis realizes effective resource allocation for data flows with multiple starting points and end points at the same time, solving the resource allocation problem in multi-task concurrent scenarios.

[0023] 3. The present invention takes into account the limitations of link bandwidth and computing power of the satellite-to-ground network, converts the problem into a multi-commodity network maximum flow optimization problem that satisfies the link bandwidth constraints and the computing resource constraints within the network, studies and designs a multi-commodity network maximum flow algorithm that satisfies the limited computing resources and link bandwidth, and determines the path selection, link bandwidth allocation, and computing resource allocation for data processing and transmission tasks oriented to maximum throughput. The method proposed in the present invention can find a feasible collaborative allocation scheme for computing resources and link resources in the satellite-to-ground network to effectively maximize the throughput of data processing and transmission when the link bandwidth constraints and computing resource constraints of the satellite-to-ground network exist simultaneously. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 It is a flow chart of the whole invention. DETAILED DESCRIPTION

[0025] The technical solution of the present invention is further described in detail below in conjunction with the accompanying drawings:

[0026] It will be understood by those skilled in the art that, unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as those generally understood by those skilled in the art in the art to which the present invention belongs. It should also be understood that terms such as those defined in common dictionaries should be understood to have meanings consistent with the meanings in the context of the prior art, and will not be interpreted with idealized or overly formal meanings unless defined as herein.

[0027] The present invention aims at how to effectively maximize the throughput of data transmission and data calculation in a satellite-to-ground network, and proposes a satellite-to-ground network multi-task computing resource and link resource collaborative allocation method for network link bandwidth limitations and satellite node computing power limitations. Graph theory can be used to find fast and effective computing and transmission strategies, thereby improving the throughput of the entire network.

[0028] The present invention proposes a collaborative allocation algorithm for computing resources and link resources for a satellite-to-ground network, which specifies the maximum amount of data that a link can carry and the amount of computing that each low-orbit satellite can bear.

[0029] Under the above conditions, multiple tasks are processed directly on low-orbit satellites and transmitted to the destination. The steps of computing and routing tasks are as follows: Figure 1 As shown. Specifically including the following steps:

[0030] Step 1: Get a directed graph by modeling the satellite-ground network. The specific steps are as follows:

[0031] Step 1.1) Set the given timeline range Divided into N time periods Each time period n∈[1, N] are two specific time points t n-1 and t n the time interval between

[0032] Step 1.2) In each time period, the topology of the satellite-to-ground network remains unchanged. If the topology of the satellite-to-ground network changes, the time period is further divided until the requirements are met. The satellite-ground network modeling within n∈[1,N] is obtained as a directed graph Map each ground station and low-orbit satellite as a node in a directed graph, using Represents a set of network nodes. Each link in the satellite-to-ground network within n∈[1,N] is mapped to an edge in the graph, using ε n Indicates time period The edge set in the satellite-to-ground network. Considering that the link status of the satellite-to-ground network in different time periods may be different, the edge set in the directed graph constructed in each time period may also be different. B(e) represents the capacity of edge e, that is, the maximum available bandwidth of the corresponding link, and C(v) represents the capacity of node v, that is, the maximum amount of data that can be calculated on the corresponding low-orbit satellite.

[0033] Step 1.3) Use represents a set of K tasks, where each task κ∈[1,K] is a triple, s k , They represent the source and destination of the task respectively, and ξ represents the processor processing power required by the task, in cycles / bit. The original data of the task is transmitted from the source point s k During the transmission process, the low-orbit satellite passing by performs the required processing and calculation, and finally transmits the calculated data to the destination d k .

[0034] Step 2: Find the shortest path from the source to the destination for each task according to the link bandwidth, and define the link available bandwidth as a directed graph The length of the corresponding edge in the path is determined by the sum of the available bandwidths of all links on the path. Further considering the available computing resources of the node, reassign the length of the shortest path and update the shortest path based on the available computing resources of the node, allocate the computing resources and link bandwidth on this shortest path, and decide the amount of data to be processed and transmitted on this path. Update the available bandwidth and computing resources of the links and nodes according to the allocation plan. Repeat this step until all computing resources are allocated or there is no feasible path to transmit data. The specific steps are as follows:

[0035] Step 2.1) Take the capacity of each edge as the length of the edge, and assign the current task Find the source point s k The shortest path to each other node and calculate the distance between two nodes. The distance between two nodes is the sum of the lengths of all edges that the shortest path between them passes through. The specific steps are as follows:

[0036] Step 2.1.1) Define the length of each edge e as the edge capacity B(e).

[0037] Step 2.1.2) Set the current task to k∈[1, K], use Dijkstra algorithm to find the source point s for the current task k The shortest path to all other nodes v is path(s k , v), and calculate the distance between nodes dis(s k , v). The distance between any two nodes is the sum of the lengths of all edges that the shortest path between them passes through.

[0038] Step 2.2) For the s obtained in step 2.1) k The distance to other nodes is taken into account for the node capacity. If the capacity of the node passed by the path is 0, then s k The distance to this node is updated to infinity, otherwise it is updated to the original distance plus the node capacity. k After the distance to each node is updated, select the current and s k The node v* with the shortest distance, traverses the neighboring nodes u of v*, if u is to s k The distance between v* and s is less than k The distance from v* to u is added to the length of the edge, then u is updated to s k The distance and shortest path of v*, otherwise unchanged; after visiting all neighbor nodes of v*, select the next one with s kThe node with the shortest distance v*, until v* = d k Or all nodes have been visited. The specific steps are as follows:

[0039] Step 2.2.1) For the value obtained in step 2.1) from s k The shortest path to other nodes v k , v) and distance dis(s k , v). Then reassign the distances of these nodes: Let the set S = {s k}.

[0040] Step 2.2.2) Select the k The node with the shortest distance And v* is not in S, let S = S + v*. For v*'s neighbor node u, update u's path and distance:

[0041]

[0042] After visiting all the neighbor nodes of v*, repeat this step until Or visit d κ .

[0043] Step 2.3) For the shortest path obtained in step 2.2), compare the minimum capacity of all edges on the path with the sum of the capacities of all nodes. If the minimum capacity of all edges on the path is less than the sum of the capacities of all nodes on the path, it means that the bottleneck of this path is the edge with the smallest capacity, and the maximum flow that can be processed and transmitted on the path is the capacity of this edge. The amount of data that each node on the path needs to calculate will be allocated according to the ratio of the capacity of this node to the sum of the capacities of all nodes on the path; otherwise, the flow is the sum of the capacities of all nodes on the path, and the amount of data that each node on the path needs to calculate is its own capacity. The specific steps are as follows:

[0044] Step 2.3.1) For step 2.2) obtained from s k to d k Pathpath(s k , d k ), let e* be path(s k , d k ) is the edge with the smallest capacity, that is,

[0045] Step 2.3.2) Let f path (s k , d k ) is path(s k , d k ) on:

[0046]

[0047] Let p(v) be the amount of data processed on node v, and f(e) be the amount of data flowing on edge e. Allocate p(v) and f(e) according to the following method:

[0048]

[0049]

[0050] Step 2.4) Based on the allocation plan of computing resources and link bandwidth of the shortest path in step 2.3), update the capacity of the edges and nodes along the path: the edge capacity is updated to the original capacity minus the traffic flowing on this edge, and the node capacity is updated to the original capacity of the node minus the amount of data calculated on this node. Continue to perform steps 2.1)-2.3) for the next task. If all tasks have allocated resources, continue to perform steps 2.1)-2.3) for each task starting from the first task. The specific steps are as follows:

[0051] Step 2.4.1) Based on the node calculation resource allocation p(v) and link bandwidth allocation f(e) in step 2.2), update each node capacity C(v) = C(v) - p(v), Update the capacity of each edge to B(e) = B(e) - f(e),

[0052] In step 2.4.2), let k=k+1, and perform steps 2.1.2)-2.4.1) again for the next task. If k=K, then set k=1.

[0053] Step 2.4.3) Repeat steps 2.1.1)-2.4.2) and stop when any of the following situations occurs: (1) C(v) = 0, (2) There is no path (s k , d k ), (3)dis(s k , d k )=∞,

[0054] Step 3: Let n=n+1, and execute step 2 until n=N. Obtain the allocation scheme of computing resources and bandwidth resources in the satellite-to-ground network and the data transmission path in each time period.

[0055] The above descriptions are only some embodiments of the present invention. It should be pointed out that, for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for collaboratively allocating computing resources and link resources in a satellite-to-ground network, comprising the following steps: Step 1: Divide a given time line range τ into N time periods. For the nth time period τ n , n∈[1,N], the satellite-ground network is modeled as a directed graph in is a set of network nodes, including satellites and ground stations, ε n Indicates that in the time period τ n The collection of all links in the intra-satellite-ground network; Step 2: Find the shortest path from the source to the destination for each task according to the link bandwidth, and define the link available bandwidth as a directed graph The length of the corresponding edge in the path is determined by the sum of the available bandwidths of all links on the path. Further considering the available computing resources of the node, the length of the shortest path is reassigned and updated according to the available computing resources of the node, the computing resources and link bandwidth on this shortest path are allocated, and the amount of data to be processed and transmitted on this path is determined. The available bandwidth and computing resources of the links and nodes are updated according to the allocation scheme. This step is repeated until all computing resources are allocated or there is no feasible path to transmit data. Step 3: Execute step 2 for the satellite-ground network information in each time period. In the step 1: Divide the given timeline range τ into N time periods. When dividing the time periods, ensure that the topology of the satellite-ground network remains unchanged in each time period. For the nth time period τ n , n∈[1,N] satellite-ground network modeling, and obtain the directed graph Map each ground station and low-orbit satellite as a node in a directed graph, using Represents the set of network nodes, in τ n , each link in the satellite-to-ground network within n∈[1,N] is mapped to a graph The edge in n Indicates that in the time period τ n Considering that the link status of the satellite-to-ground network may be different in different time periods, the directed graph constructed The edge sets in may also be different. The capacity of the edge is the available bandwidth of the corresponding link, and the capacity of the node is the available computing resources of the corresponding ground station or low-orbit satellite. Assume that there are K tasks that need to be calculated and transmitted in this satellite-ground network, κ∈[1,K] represents the κth task, where s κ For the task The source point, d k For the task The destination of The raw data of the task is from the source node s k During the transmission process, the low-orbit satellite passing by performs the required processing and calculation, and finally transmits the calculated data to the destination d k . In the step 2: Step 2.1) Take the capacity of each edge as the length of the edge, and assign the current task Find the source point s k The shortest path to each other node and calculate the distance between two nodes. The distance between two nodes is the sum of the lengths of all edges passed by the shortest path between them. Step 2.2) For the s obtained in step 2.1) k The distance to other nodes is taken into account for the node capacity. If the capacity of the node passed by the path is 0, then s k The distance to this node is updated to infinity, otherwise it is updated to the original distance plus the node capacity. k After the distance to each node is updated, select the current and s k The node v* with the shortest distance, traverses the neighboring nodes u of v*, if u is to s k The distance between v* and s is less than k The distance from v* to u is added to the length of the edge, then u is updated to s k The distance and shortest path of v*, otherwise unchanged; after visiting all neighbor nodes of v*, select the next one with s κ The node with the shortest distance v*, until v* = d k Or all nodes have been visited. Step 2.3) For the shortest path obtained in step 2.2), compare the minimum capacity of all edges on the path with the sum of the capacities of all nodes. If the minimum capacity of all edges on the path is less than the sum of the capacities of all nodes on the path, it means that the bottleneck of this path is the edge with the smallest capacity, and the maximum flow that can be processed and transmitted on the path is the capacity of this edge. The amount of data that each node on the path needs to calculate will be allocated according to the ratio of the capacity of this node to the sum of the capacities of all nodes on the path; otherwise, the flow is the sum of the capacities of all nodes on the path, and the amount of data that each node on the path needs to calculate is its own capacity. Step 2.4) Update the capacity of the edges and nodes along the path according to the allocation plan of computing resources and link bandwidth of the shortest path in step 2.3): the edge capacity is updated to the original capacity minus the traffic flowing on this edge, and the node capacity is updated to the original capacity of the node minus the amount of data calculated on this node. Continue to execute steps 2.1)-2.3) for the next task. If all tasks have allocated resources, continue to execute steps 2.1)-2.3) for each task starting from the first task. Step 2.5) loops through steps 2.1)-2.4) until all computing resources are allocated or there is no feasible path available to transmit data. In the step three: Let n=n+1, and execute step 2 until n=N. Obtain the allocation scheme of computing resources and bandwidth resources in the satellite-to-ground network and the data transmission path in each time period.

2. The method for collaboratively allocating computing resources and link resources in a satellite-to-ground network according to claim 1, characterized in that In the step 1: Step 1.1) Divide the given timeline range τ into N time periods τ = {τ1, τ2, ..., τ N }, each time period τ n =[t n-1 , t n ], n∈[1,N] are two specific time points t n-1 and t n the time interval between Step 1.2) In each time period, the topology of the satellite-to-ground network remains unchanged. If the topology of the satellite-to-ground network changes, the time period is further divided until the requirements are met. n , n∈[1,N] within the satellite-ground network model, and obtain the directed graph Map each ground station and low-orbit satellite as a node in a directed graph, using Represents the set of network nodes, in τ n , each link in the satellite-to-ground network within n∈[1,N] is mapped to an edge in the graph, represented by ε n Represents the time period τ n The edge set in the satellite-to-ground network. Considering that the link status of the satellite-to-ground network in different time periods may be different, the edge set in the directed graph constructed in each time period may also be different. B(e) represents the capacity of edge e, that is, the maximum available bandwidth of the corresponding link, and C(v) represents the capacity of node v, that is, the maximum amount of data that can be calculated on the corresponding low-orbit satellite. Step 1.3) Use represents a set of K tasks, where each task k∈[1,K] is a triple, They represent the source and destination of the task respectively, and ξ represents the processor processing power required by the task, in cycles / bit. The original data of the task is transmitted from the source point s k During the transmission process, the low-orbit satellite passing by performs the required processing and calculation, and finally transmits the calculated data to the destination d k .

3. The method for collaboratively allocating computing resources and link resources in a satellite-to-ground network according to claim 1, characterized in that The specific steps described in step 2.1) include: Step 2.1.1) Define the length of each edge e as the edge capacity B(e). Step 2.1.2) Set the current task to Use Dijkstra algorithm to find the source point s for the current task k The shortest path to all other nodes v is path(s k , v), and calculate the distance between nodes dis(s k , v). The distance between any two nodes is the sum of the lengths of all edges that the shortest path between them passes through.

4. The method for collaboratively allocating computing resources and link resources in a satellite-to-ground network according to claim 1, characterized in that The specific steps described in step 2.2) include: Step 2.2.1) For the value obtained in step 2.1) from s k The shortest path to other nodes v (s k , v) and distance dis(s κ , v). Considering the node capacity, the distance dis(s k , v) Reassign: Initialize the set S = [s κ }. Step 2.2.2) Select a point s away from the source k The node with the shortest distance that is not in the set S Here, arg min represents the extreme point, that is, the value of the variable when the function behind it reaches the minimum value. Add v* to the set S, that is, S = S ∪ v*. For the neighbor node u of v*, use the following rules to update the source point s k The shortest path and distance to node u: After visiting all the neighbor nodes of v*, repeat this step until Or u*=d k .

5. The method for collaboratively allocating computing resources and link resources in a satellite-to-ground network according to claim 1, characterized in that The specific steps described in step 2.3) include: Step 2.3.1) For the source point s obtained in step 2.2) k To destination k The path path(s k , d k ), let e* be path(s k , d k ) is the edge with the smallest capacity, that is, Step 2.3.2) Let f path (s k , d k ) is path(s k , d k ) is the maximum flow that can be calculated and transmitted on the path, and this flow is called path(s κ , d k ), the bottleneck flow is defined as follows: Let p(v) be the amount of data computed on node v, and f(e) be the amount of data flowing on edge e. Allocate p(v) and f(e) according to the following method:

6. The method for collaboratively allocating computing resources and link resources in a satellite-to-ground network according to claim 1, characterized in that The specific steps described in step 2.4) include: Step 2.4.1) Update the capacity of each node based on the node resource allocation p(v) and link bandwidth allocation f(e) calculated in step 2.3) Update the capacity of each edge Step 2.4.2) Set k=k+1, and perform steps 2.1.2-2.4.1 again for the next task. If k=K, set k=1. Step 2.4.3) Repeat steps 2.1.1) - 2.4.2) and stop when any of the following situations occur: (1) (2) None of the (3)

Images