A multicast routing and load balancing method applied to a low earth orbit satellite network

Abstract

The application discloses a multicast routing and load balancing method applied to a low-orbit satellite network and belongs to the field of low-earth orbit satellite network communication. The method acquires low-orbit satellite network topology and link state information, models a multicast tree construction problem into a joint path optimization game, and solves the problem through a cooperative game routing algorithm. Through the introduction of an incentive link cost sharing mechanism and the combination of a dynamically adjusted cost weight, the method optimizes multicast path selection, effectively shares link costs and realizes load balancing. A composite cost function integrating link load and delay is designed to calculate path total cost. The method provided by the application provides an innovative and efficient solution to the multicast routing and load balancing problem in the low-orbit satellite network, can enhance the transmission efficiency, service quality, network stability and overall throughput of multicast services in the low-orbit satellite network, and has significant advantages in improving network performance, guaranteeing service quality and enhancing network robustness.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of low earth orbit (LEO) satellite network communication, and more particularly relates to a multicast routing and load balancing method applied to a low earth orbit satellite network. BACKGROUND

[0002] Low earth orbit satellite networks have become a key technology path to support global communication, Internet of Things, remote sensing monitoring, and especially real-time video transmission and other applications, due to their inherent advantages of low delay, high bandwidth, and seamless global coverage. With the continuous expansion of the number of satellites in orbit and the explosive growth of emerging applications such as 4K / 8K ultra-high-definition video and virtual reality live streaming, the effective management of multicast traffic and the optimization of bandwidth utilization in satellite networks have become a higher requirement, and have become a core challenge to achieve efficient video distribution and other services.

[0003] Currently, there are many challenges in implementing efficient multicast services in LEO satellite networks. First, LEO satellite networks and ground networks have inherent differences in architecture and operating environment, and their topology structure frequently changes dynamically due to the high-speed movement of satellites, making it difficult for mature multicast technologies on the ground to be directly applicable to this network environment. For example, traditional IP multicast technologies, such as using a shortest path tree (SPT) to build a multicast routing, need to be frequently reconstructed under a dynamic topology, making it difficult to optimize bandwidth consumption. At the same time, the distributed nature of IP multicast limits the optimal perception of global network state and centralized resource scheduling, making it difficult to provide reliable QoS guarantees.

[0004] To address the above challenges, software-defined multicast (SDM) technology has been introduced into LEO satellite networks. SDM draws on the centralized control idea of software-defined networks (SDN), and collects global network information (such as topology, link load, delay, etc.) through a controller to dynamically optimize multicast tree construction, thereby improving bandwidth utilization and traffic scheduling accuracy. Existing research has explored the application of SDM to LEO satellite networks, such as using a Steiner tree algorithm to optimize the length of a multicast tree or to avoid obstacles in multicast routing.

[0005] Although SDM technology shows potential, current SDM research for LEO satellites still has deficiencies. First, many studies still mainly focus on the node positioning and construction optimization of the Steiner tree. However, this often leads to excessive utilization of part of the Inter-Satellite Link (ISL), thereby causing hotspot link congestion and poor network load balancing. Second, the Steiner tree has inherent limitations in multi-objective optimization, making it difficult to effectively balance multiple key QoS indicators such as bandwidth utilization, transmission delay, and reliability. Third, in the face of dynamic changes in satellite network topology, especially satellite node failures or link congestion, methods based on the Steiner tree often lack sufficient dynamic adaptability and rapid self-healing ability.

[0006] Therefore, how to achieve efficient multicast routing and resource management in the case of dynamic changes in LEO satellite network topology and frequent link failures, while considering multi-objective optimization such as bandwidth, delay, and load balancing, is a problem that needs to be solved. SUMMARY

[0007] In view of the above defects or improvement needs of the prior art, the present application provides a multicast routing and load balancing method applied to a low-orbit satellite network, thereby solving the problems of low routing efficiency, unbalanced load, insufficient QoS guarantee, and poor adaptability to network dynamics existing in existing multicast routing schemes.

[0008] To achieve the above-mentioned purpose, according to the first aspect of the present application, a multicast routing and load balancing method applied to a low-orbit satellite network is provided, comprising:

[0009] S1, determining the network topology structure and link state information of the satellite constellation under the current time slice, mapping each satellite to a virtual node one by one, and constructing an undirected graph ; wherein, is a set of virtual nodes, is a set of communication links between virtual nodes;

[0010] S2, according to the user request of the ground station, determining the source virtual node and the target virtual node set , regarding each target virtual node as a game participant to build a joint path optimization game model for cooperative game;

[0011] wherein, the goal of each game participant is to minimize the total cost of the path from to itself by selecting and adjusting the path from to The total cost of a path is the sum of the costs of all links in that path; the total cost of a link between any two adjacent virtual nodes is the weighted sum of the load rate and the link delay of that link, and this total cost is shared by all target virtual nodes that transmit data through that link; j=1,2,…,n; n is the number of target virtual nodes;

[0012] S3, obtained by solving the joint path optimization game model. arrive The optimal path is determined and a multicast path routing table is generated, thereby obtaining a multicast tree that satisfies the user's request.

[0013] The solution process includes: initialization arrive The path, total path cost, participating nodes and link load of each link; traversal Search from arrive If the total cost of the new path is lower than the total cost of the current path, then update... arrive The path, total path cost, participating nodes and link load of each link are not updated unless the path is found; the search continues from... arrive The game continues until no new path that reduces the total cost of the path can be found. If no new path can be found that reduces the total cost of the path, the current path is the optimal path, and the game ends.

[0014] According to a second aspect of the present invention, an electronic device is provided, comprising: a computer-readable storage medium and a processor;

[0015] The computer-readable storage medium is used to store executable instructions;

[0016] The processor is configured to read executable instructions stored in the computer-readable storage medium and execute the method as described in the first aspect.

[0017] According to a third aspect of the invention, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions for causing a processor to perform the method as described in the first aspect.

[0018] According to a fourth aspect of the invention, a computer program product is provided, comprising a computer program or instructions that, when executed by a processor, implement the method described in the first aspect.

[0019] In summary, compared with the prior art, the above-described technical solutions conceived by this invention can achieve the following beneficial effects:

[0020] 1. The multicast routing and load balancing method for low-orbit satellite networks provided by the application models the multicast tree construction problem as a joint path optimization game (JPOG). In the JPOG, the calculation of link cost not only considers the transmission delay and load of the link, but also considers the resource consumption and network topology of the link. This multi-dimensional link cost model enables the algorithm to make more accurate optimization decisions in complex network environments. In addition, the JPOG has a load balancing mechanism that can effectively balance the resource demands between different multicast groups by reasonably allocating resources, avoiding performance bottlenecks caused by overloading on some links. In this way, the algorithm optimizes the allocation of network resources while ensuring the performance of each multicast group is effectively guaranteed. In actual network application scenarios, due to the dynamic changes in the number of multicast groups, network traffic inevitably presents a characteristic of stage-wise uneven distribution. To address this issue, the JPOG can dynamically adjust resource allocation strategies in real time, continuously optimizing link load distribution to ensure fairness between different multicast groups and maximize network resource utilization. In this game model, each multicast target virtual node is considered a game participant whose goal is to minimize its total path cost by selecting the optimal path.

[0021] Furthermore, to promote cooperation and achieve global optimization, a link cost sharing mechanism is adopted: when multiple target virtual nodes' paths share a link, the cost of that link is shared by all target virtual nodes that pass through it. This mechanism encourages target virtual nodes to share link resources, thereby reducing their own costs and reducing overall network resource consumption.

[0022] In addition, the method uses a composite cost function that combines link load and delay to evaluate the quality of a link; link cost is positively related to the current load rate and transmission delay of the link, and low-cost links mean low load and low delay, which helps to improve network transmission efficiency and ensure quality of service (QoS). The link load rate reflects the degree of link congestion, while the link delay takes into account the transmission delay of data packets and the queuing delay on satellite nodes. By adjusting the weight coefficients of load and delay in the cost function, optimization can be performed according to different application requirements.

[0023] 2、As a further preferred scheme, the method provided by the application adopts a two-stage A* path generation algorithm based on Manhattan distance as a stage switching condition in path search, which dynamically switches the search strategy according to the Manhattan distance between the current node and the target virtual node: in the initial stage of path search (when the distance to the target virtual node is far), a breadth-first strategy is adopted to explore multiple directions of the adjacent nodes in order to quickly find a potential low-cost path; when approaching the target virtual node, the search is focused on the direction towards the target virtual node in order to reduce redundant path expansion and improve path generation efficiency.

[0024] 3、As a further preferred scheme, the method provided by the application adopts a lightweight member management mechanism and a fault recovery mechanism. The member management mechanism can handle the dynamic joining and exiting of multicast group members, independently execute the CGMR algorithm to access the existing session when a new member joins, and orderly release resources when a member exits to avoid affecting other members; through the member management mechanism, the dynamic joining and exiting of multicast group members are managed to ensure the stability of the session and the effective use of resources. When a link fault is detected, the fault recovery mechanism sets the cost of the fault link to infinity through the SDM controller, forces the path generation algorithm to avoid the fault link, and allows the affected nodes to perform local path re-planning, thereby quickly recovering communication, enhancing the stability and throughput of the network, realizing the fast self-healing of communication services and guaranteeing the overall stability of the network. BRIEF DESCRIPTION OF DRAWINGS

[0025] Figure 1 A flowchart of a multicast routing and load balancing method applied to a low-orbit satellite network provided by an embodiment of the application.

[0026] Figure 2 A mapping relationship diagram of a network topology and a virtual node provided by an embodiment of the application.

[0027] Figure 3 A flowchart of a CGMR algorithm provided by an embodiment of the application.

[0028] Figure 4 A schematic diagram of a two-stage A* path generation algorithm provided by an embodiment of the application.

[0029] Figure 5 A multicast routing schematic diagram. DETAILED DESCRIPTION

[0030] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0031] This invention provides a multicast routing and load balancing method for low-Earth orbit satellite networks, such as... Figure 1 As shown, it includes:

[0032] S1, determine the network topology and link status information of the satellite constellation under the current time slice, map each satellite to a virtual node, and construct an undirected graph. ;in, A set of virtual nodes A set of communication links between virtual nodes;

[0033] S2, based on the user request from the ground station, determines the source virtual node. and the target virtual node set Each target virtual node is regarded as a game participant, and a joint path optimization game model is constructed for cooperative game.

[0034] In this game, the goal of each participant is to autonomously choose and adjust their strategy from... The path to itself, in order to minimize its total path cost; from arrive The total cost of a path is the sum of the costs of all links in that path; the total cost of a link between any two adjacent virtual nodes is the weighted sum of the load rate and the link delay of that link, and this total cost is shared by all target virtual nodes that transmit data through that link; j=1,2,…,n; n is the number of target virtual nodes;

[0035] S3, obtained by solving the joint path optimization game model. arrive The optimal path is determined and a multicast path routing table is generated, thereby obtaining a multicast tree that satisfies the user's request.

[0036] The solution process includes: initialization arrive The path, total path cost, participating nodes and link load of each link; traversal Search from arrive If the total cost of the new path is lower than the total cost of the current path, then update... arrive the path, the total cost of the path, and the participating nodes and link load of each link, otherwise not update; continuously find new paths from to , until no new path can reduce the total cost of the path, then the current path is the optimal path, and the game ends.

[0037] The method provided by the embodiment of the application constructs a cooperative game-theoretic multicast routing (CGMR) algorithm, which is applied to multicast tree construction in a low earth orbit satellite network of a software defined multicast architecture. The typical architecture is divided into three levels of architecture of an application data layer, a control layer and an application service layer: the data layer is composed of a LEO satellite constellation and inter-satellite links (ISLs) and satellite-ground links, and undertakes an actual data packet forwarding task; the control layer is deployed in a ground control station, has global network topology, link state and traffic information, and the CGMR algorithm for calculating the cooperative game-theoretic multicast routing proposed in the application runs in the layer and issues control instructions to the data layer satellite; and the application service layer is an interface of a ground terminal user or a base station accessing the satellite network, and provides multicast service access.

[0038] A Walker constellation configuration composed of satellites. The satellites are connected with each other through in-orbit plane inter-satellite links and cross-orbit plane inter-satellite links (for example, X-type links), and form a mesh-shaped topological structure. In view of the topological dynamic problem caused by high-speed movement of the satellites, the embodiment of the application introduces the concept of time slicing (Time Slice), and in each time slicing period, the network topology is regarded as relatively stable. Meanwhile, a virtual node mapping technology is adopted, and a virtual rectangular network ( 、 is a virtual network dimension parameter). As shown in Figure 2 , each physical satellite is mapped to a virtual node , and the dynamic binding relationship between the satellite and the virtual node is described through an indication function . In a single time slice, the network topology is modeled as an undirected graph , wherein is a virtual node set, and is a communication link set between virtual nodes. With the passage of time, the virtual topological structure maintains logical static characteristics, and when the physical satellites move, they will be rebound with the virtual nodes, and link state information such as topology configuration, satellite load and energy consumption is synchronously updated.

[0039] The method provided in this invention aims to efficiently generate multicast routes within each time slice, thereby minimizing the cost of inter-satellite links (ISL).

[0040] Define the link between two nodes The cost is the weighted sum of the link's load rate and link latency. Link latency is mainly composed of transmission delay. and queue delay It consists of two parts.

[0041] Preferably, an M / M / 1 queuing model is used to account for the queuing delay of links in the satellite network. To perform modeling, that is, .

[0042] link Total cost The calculation formula is:

[0043]

[0044]

[0045]

[0046] in, Indicates the average data packet size. This indicates the average load on the link. Indicates the capacity of the link. Indicates the bandwidth of the link. This represents the average arrival rate of data packets. Indicates the service rate of the link. Indicates the load rate of the link. These are the preset weighting coefficients.

[0047] In the traditional unicast path problem, each target virtual node typically chooses its path solely to minimize its own cost. However, in satellite multicast scenarios, the situation becomes more complex because multiple paths often need to share a common physical link. This resource-sharing characteristic means that the path decision of any node not only affects its own cost but also has a direct cascading effect on the costs of other nodes sharing the link. Therefore, to address this multi-party interactive optimization problem, the multicast tree construction problem can be modeled as a Joint Path Optimization Game (JPOG).

[0048] In a specific multicast service, the source virtual node is defined as The target set of virtual nodes is (where n is the number of target virtual nodes), where each target virtual node Each of the n participants, j = 1, 2, …, n, aims to minimize its own total path cost by making independent decisions and adjusting its path from the source node to itself.

[0049] Therefore, the path selection optimization problem for the target virtual node can be formulated as:

[0050]

[0051] wherein, denotes the complete path from the source node to the target virtual node , denotes a link constituting the path , is the cost of the link.

[0052] To facilitate the optimization of the global path, the method provided by the embodiment of the present application introduces a link cost sharing mechanism: when the path of the target virtual node contains a link, the target virtual node is added to the set of participating nodes of the link, and the cost allocated to the target virtual node on the link is the total cost of the link divided by the total number of target virtual nodes passing through the link.

[0053] Under this mechanism, the total cost of a physical link will be shared by all target virtual nodes sharing the link, instead of being independently borne by a single user. Specifically, a "participant" list is maintained for each link to track all target virtual nodes sharing the link. Therefore, the total cost of the complete path from the source node to the target virtual node is the sum of the allocated costs of all links on the path:

[0054]

[0055] wherein, denotes the "participant" list of the link , i.e., the set of target virtual nodes passing data through the link , denotes the total cost of the path .

[0056] This cost-sharing mechanism aims to incentivize each target virtual node to choose shared link resources. The more nodes that share, the lower the cost per node, thus reducing not only its own total path cost but also the cost for other nodes sharing the link, creating a "cooperative game" situation. When a target virtual node... To minimize its total path cost, its current path is... Change to a new path This path change process typically involves one or more new links. The addition (i.e.) and ), and one or more old links Removal (i.e.) and Accordingly, the target virtual node It will be added to all newly joined links The set of participating nodes In, and from all the old links that were removed The set of participating nodes Removed from the middle.

[0057] To mitigate the risk of frequent path changes (i.e., path oscillations), as a further optimization, after the path is updated, the target virtual node... The method for calculating the total path cost will be adjusted. Its new total path cost... Given by the following formula:

[0058]

[0059]

[0060] in, Indicates the updated path Links in The weighting ratio Indicates the updated path The total cost.

[0061] Weighting ratio The design allows nodes with higher total path costs to receive relatively higher weight on shared links, while reducing the risk of frequent path selection changes (i.e., path oscillations).

[0062] For multicast services JPOG achieves Pure Strategy Nash Equilibrium (PSNE) through an iterative method. The overall framework of the Cooperative Game Multicast Routing (CGMR) algorithm provided in this embodiment is shown in Table 1.

[0063]

[0064] The specific flow of the algorithm is shown in Figure 3 The flow of CGMR includes:

[0065] Initialization: The ground control station initializes the satellite network as an undirected graph Each target virtual node is a game participant, and initializes the path list , initializes the participant node list of the link and the link load . Each target virtual node randomly selects a shortest path from the available paths as the initial path, calculates the total cost of the link and stores it in the cost list .

[0066] Path selection and list update: when the target virtual node selects a path, if the path contains a link , update the participant node list according to , and update the link load according to .

[0067] Optimal path search and replacement: traverse each target virtual node , find the optimal path by two-stage A* algorithm, and calculate the total cost of the link . If , replace the path and update the participant node list according to the rules :

[0068]

[0069] Move from the participant node list of the old path to the participant node list of the new path , and store the updated path in the path list p, which is dynamically updated with the change of the path. If the cost is the same, it is not updated.

[0070] Load and cost update: at the end of each round of path selection, update the relevant link load and recalculate the total cost of the path of each target virtual node . The CGMR process continues until each target virtual node cannot find a better path that can reduce its own cost, at which point the path is the optimal path and the game ends.

[0071] To improve the path search efficiency, preferably, in the path search, a two-stage A* algorithm is used to perform optimal path search. In the initial stage of path search, when the Manhattan distance between the current virtual node and the target virtual node is greater than or equal to a preset threshold, it is in the first stage, a breadth-first strategy is used to explore multiple directions of the neighboring nodes; when the Manhattan distance is less than the preset threshold, it is in the second stage, a search is performed in the direction focusing on the target virtual node.

[0072] Specifically, the two-stage A* algorithm improves the path generation efficiency by dynamically adjusting the neighbor node traversal strategy and using different search strategies in different stages. The path search process is divided into two stages: the first stage phase1 and the second stage phase2, and the stage switching condition is defined as follows:

[0073]

[0074] Wherein, is the Manhattan distance from the current neighbor node to the target virtual node, is an adjustment factor for controlling the entry into the later stage.

[0075] The two-stage A* algorithm determines whether to switch to the later stage according to the distance between the current node and the target virtual node, and then adjusts the search focus. As shown in Figure 4 According to this rule: when the current node searched is still far from the final target virtual node, it is in the phase1 stage, at this time all directions of the neighbor nodes of the current node are explored for extensive search. The purpose of this is to avoid locking a direction too early, so that the global and possible low-cost path can be found faster; when the current node searched is close to the target virtual node, it is switched to the phase2 stage, and the search strategy becomes more purposeful, focusing on the direction towards the target virtual node. This reduces the redundant calculation and path expansion in irrelevant directions, thereby significantly accelerating the accurate search near the end point.

[0076] After obtaining the optimized paths of each target virtual node by the CGMR algorithm, the ground control station is configured to perform path integration operation to generate a multicast path routing table. The system uses a distributed routing mechanism, as shown in Figure 5 The ground control station establishes a communication connection with the related satellites and sends corresponding next-hop forwarding information to each satellite node, without obtaining the global routing topology, thereby achieving the technical effects of reducing satellite signaling overhead and enhancing system scalability.

[0077] The method provided by the application collects satellite network topology structure, link load, and ground station user request and the like in real time; converts the satellite network topology structure into a standardized network model and a game scenario according to the above data; simulates path selection games among each target virtual node by using a cooperative game routing (CGMR) algorithm and using a cost sharing mechanism according to the constructed model; and finally calculates a globally optimal multicast routing scheme, which is converted into a specific next-hop forwarding table and is issued to each satellite node for execution.

[0078] As a further preferred solution, when a new member node dynamically joins a stable multicast session, there is no need to rebuild the complete multicast tree structure. Given that rebuilding the multicast tree can cause dramatic fluctuations in link state, increase path reconstruction calculation overhead, and even cause temporary network performance degradation, the method provided by the embodiment of the application adopts the following new member management mechanism:

[0079] The new member node independently executes the CGMR algorithm based on the real-time acquired network link load and delay state parameters, selects the optimal access link through path total cost optimization calculation, and accesses the existing multicast session in a manner of minimizing the total path cost, thereby effectively avoiding excessive disturbance to the overall multicast tree structure.

[0080] When a member node needs to exit the multicast session due to device failure, unstable link, or network environment change, the control layer is configured to manage the link corresponding to the node to maintain the overall stability of the network. The specific implementation is as follows: when the member node exits, the control layer triggers a link disconnection process, which will recursively disconnect the related links on the path of the member node to release the corresponding network resources. It needs to be particularly limited that if it is detected that a link is still used by other member nodes, the disconnection operation of the link is inhibited to avoid interference with the normal communication of other multicast members. This mechanism aims to maintain the stability of the multicast tree while minimizing the negative impact caused by the node exit. During the link removal process, the control layer queries the update of the participating node list of each link to determine whether the corresponding link needs to be disconnected, ensuring the accuracy and efficiency of the management measures.

[0081] The link fault recovery mechanism is a key technical means to ensure the stable operation of the LEO satellite network. Influenced by factors such as solar activity, satellite maintenance, and network environment change, link faults are difficult to completely avoid in actual operation. When a link fails, the multicast service path relying on the link needs to be reconstructed in a timely manner to ensure the continuity of the communication service and the overall stability of the network. As a further preferred solution, the method provided by the embodiment of the application adopts the following link fault recovery mechanism:

[0082] When a link fails, the software-defined multicast (SDM) controller sets the cost or load parameter of the failed link to infinity to force the path generation algorithm to avoid the failed link.

[0083] Allow the affected node to carry out local path re-planning (the path of the unaffected node remains unchanged), and when path reconstruction, the search range is expanded to all adjacent nodes, rather than only using the two-stage A* search strategy.

[0084] Specifically, to minimize the impact of link failure on the overall network performance, the path reconstruction operation is limited to the multicast destination virtual nodes directly affected by the failure, without the need for large-scale adjustment of the overall network path layout. Under the unified management architecture of the SDM controller, the controller is configured to set the load parameter of the failed link to infinity, and through this parameter configuration, the path generation algorithm is forced to automatically avoid the failed link, effectively preventing communication interruption problems caused by link failure.

[0085] To improve the efficiency of path reconstruction, the path generation algorithm under the link failure scenario usually disables the shortest path search strategy. At this time, instead of using the two-stage A* search strategy, the search range is directly expanded to all adjacent nodes. Since the shortest path search algorithm relies on complete and stable network state information, if the search path contains a failed link, it may cause path calculation to fail to converge or even fail to generate, therefore, by expanding the search range, the failed link can be effectively avoided, ensuring the reliability of path generation and improving the success rate and efficiency of path reconstruction.

[0086] The method provided by the embodiment of the application models the multicast tree construction problem as a joint path optimization game (JPOG), and innovatively proposes a cooperative game multicast routing (CGMR) algorithm. The algorithm effectively encourages link sharing between target virtual nodes by introducing a link cost allocation mechanism, thereby optimizing multicast path selection and achieving effective utilization and load balancing of network resources. Moreover, the complex cost function designed by the application, which combines link load and delay, combined with the efficient two-stage A* path generation algorithm, significantly improves network transmission efficiency and guarantees quality of service (QoS). In addition, the method provided by the embodiment of the application adopts a lightweight member management mechanism and a failure recovery mechanism, enabling the system to quickly adapt to the dynamic changes in the topology of the low-orbit satellite network and link failures, greatly enhancing the stability, reliability and overall throughput of the network.

[0087] In summary, the method provided by the embodiment of the application provides an innovative and efficient solution to the multicast routing and load balancing problem in low-orbit satellite networks, which has significant advantages in improving network performance, guaranteeing service quality and enhancing network robustness.

[0088] The embodiment of the application provides a multicast routing and load balancing system applied to a low-orbit satellite network, which can be deployed on an SDM architecture, and the system comprises a computer readable storage medium and a processor.

[0089] The computer readable storage medium is configured to store executable instructions.

[0090] The processor is configured to read the executable instructions stored in the computer readable storage medium, and execute the method according to any one of the above embodiments.

[0091] An embodiment of the present application provides a computer readable storage medium, which stores computer instructions, and the computer instructions are configured to make a processor execute the method according to any one of the above embodiments.

[0092] An embodiment of the present application provides a computer program product, which comprises computer programs or instructions, and the computer programs or instructions are executed by a processor to implement the method according to any one of the above embodiments.

[0093] Those skilled in the art can easily understand that the above description is only the preferred embodiment of the present application, and is not intended to limit the present application, and any modification, equivalent replacement and improvement made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A multicast routing and load balancing method applied to low-Earth orbit satellite networks, characterized in that, include: S1, determine the network topology and link status information of the satellite constellation under the current time slice, map each satellite to a virtual node, and construct an undirected graph. ;in, A set of virtual nodes A set of communication links between virtual nodes; S2, based on the user request from the ground station, determines the source virtual node. and the target virtual node set Each target virtual node is regarded as a game participant, and a joint path optimization game model is constructed for cooperative game. In this game, the goal of each participant is to autonomously choose and adjust their strategy from... The path to itself, in order to minimize its total path cost; from arrive The total cost of a path is the sum of the costs of all links in that path; the total cost of a link between any two adjacent virtual nodes is the weighted sum of the load rate and the link delay of that link, and this total cost is shared by all target virtual nodes that transmit data through that link; j=1,2,…,n; n is the number of target virtual nodes; S3, obtained by solving the joint path optimization game model. arrive The optimal path is determined and a multicast path routing table is generated, thereby obtaining a multicast tree that satisfies the user's request. The solution process includes: initialization arrive The path, total path cost, participating nodes and link load of each link; traversal Search from arrive If the total cost of the new path is lower than the total cost of the current path, then update... arrive The path, total path cost, participating nodes and link load of each link are not updated unless the path is found; the search continues from... arrive The game continues until no new path that reduces the total cost of the path can be found. If no new path can be found that reduces the total cost of the path, the current path is the optimal path, and the game ends.

2. The method as described in claim 1, characterized in that, from arrive path Total path cost The calculation formula is: in, To pass through the link The target set of virtual nodes for data transmission To pass through the link The number of target virtual nodes for data transmission. for One of the links in the chain.

3. The method as described in claim 2, characterized in that, from arrive path The path is Updated to back, arrive Total path cost The calculation formula is: in, , for Total path cost, for Total path cost, From arrive The path, for The target virtual node in To pass through the link The target set of virtual nodes for data transmission for One of the links in, For link The total cost.

4. The method according to any one of claims 1-3, characterized in that, In step S3, the path search process is divided into a first stage and a second stage; the first stage uses a breadth-first search strategy to search for paths, and the second stage explores the direction towards the target virtual node. Wherein, if the Manhattan distance from the current neighbor node to the target virtual node is Satisfying the relation If the path search is successful, the path search process is in the first stage; otherwise, it is in the second stage. As a regulating factor, , The dimension parameters are used to map each satellite to a virtual node to form a virtual network.

5. The method as described in claim 1, characterized in that, Also includes: When a new satellite virtual node joins, it connects to the existing multicast session by minimizing its total path cost, based on real-time link status information, without changing the optimal path from the source node to other virtual nodes. When a satellite virtual node exits, the links on its path are recursively disconnected; however, if a link on the path is still being used by another satellite virtual node, it is not disconnected.

6. The method as described in claim 1, characterized in that, Also includes: When a link fails, the total cost of the failed link is set to infinity, and the affected satellite virtual nodes expand the path search range to all adjacent nodes to replan the path.

7. The method as described in claim 1, characterized in that, The link delay between any two virtual nodes includes transmission delay and queuing delay.

8. A multicast routing and load balancing system for low-Earth orbit satellite networks, characterized in that, include: Computer-readable storage media and processors; The computer-readable storage medium is used to store executable instructions; The processor is configured to read executable instructions stored in the computer-readable storage medium and execute the method as described in any one of claims 1-7.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing a processor to perform the method as described in any one of claims 1-7.

10. A computer program product, comprising a computer program or instructions, characterized in that, When the computer program or instructions are executed by a processor, they implement the method as described in any one of claims 1-7.

Images