Satellite network hybrid control method based on SRv6

Abstract

The invention relates to the technical field of satellite communication, in particular to an SRv6-based satellite network hybrid control method, which comprises the following steps of: planning a routing strategy of satellite nodes by taking minimization of link cost as a target in an orbit period, and acquiring a global routing strategy of all satellite nodes; in a time slice of an orbital period, t nodes are selected from all satellites to serve as center nodes, and other nodes are divided to the corresponding center nodes to form t domains; according to the divided domains, routing optimization is carried out with the goal of minimizing the transmission cost between satellite nodes in the domains; according to the method, the requirement for high flexibility is met, meanwhile, data transmission delay and route updating burden caused by network topology change and link instability are remarkably reduced, and the overall performance of the low-orbit satellite network is improved.

Description

Technical Field

[0001] The present invention relates to the field of satellite communication technology, and in particular to a satellite network hybrid control method based on SRv6. Background Art

[0002] As the global demand for communications continues to grow, low-orbit satellite (LEO) networks have gradually become one of the key technologies supporting global Internet access due to their wide coverage, low latency and high bandwidth. Compared with high-orbit satellites (such as GEO satellites), LEO satellites have lower transmission latency and can provide effective communication support for remote areas and areas at sea where traditional networks are difficult to cover. However, the special operating environment of LEO satellite networks has led to a series of technical challenges in practical applications, especially when the network topology changes frequently and the link is unstable, and the traditional ground routing mechanism cannot be directly applied to low-orbit satellites.

[0003] First, LEO satellites are in low orbits and orbit the earth at a high speed, which causes the link connection status between satellites to change over time. The dynamic nature of the network topology makes it very difficult to adjust the routing in real time. In the prior art, the centralized control architecture is usually responsible for the global path calculation and management of the entire network by the ground control center. However, this centralized architecture has obvious bottlenecks. One is the risk of single point failure: the ground control center is the control node of the entire network. Once a failure occurs, the routing function of the entire satellite network will be paralyzed. The second is the high latency problem: due to the long-distance communication between the satellite and the ground station, the signal transmission delay is significant, especially when the routing table needs to be updated frequently, the response speed of the centralized architecture is difficult to meet the needs of low-latency application scenarios. On the other hand, the distributed control architecture disperses the control rights to each satellite node, and each node performs routing calculations and decisions autonomously, thereby reducing the dependence on the ground control center. The distributed architecture improves the robustness and flexibility of the system, but it also brings some new problems. Due to the lack of a global perspective, distributed routing is difficult to achieve global optimization when performing routing calculations, and is prone to local congestion and uneven utilization of network resources. Especially in the case of large-scale networks, distributed control is difficult to achieve global optimization of the entire network. Summary of the invention

[0004] In view of the above problems, the present invention proposes a satellite network hybrid control method based on SRv6, in which a ground control center sends a control instruction to a satellite node, and the satellite node forwards a data packet according to the control instruction. The method is characterized in that the method comprises the following steps:

[0005] In one orbital period, the routing strategy of satellite nodes is planned with the goal of minimizing link costs, and the global routing strategy of all satellite nodes is obtained;

[0006] In a time slice of an orbital cycle, t nodes are selected from all satellites as central nodes, and other nodes are divided into corresponding central nodes to form t domains;

[0007] According to the divided domain, the routing optimization is performed with the goal of minimizing the transmission cost between satellite nodes in the domain. Furthermore, the process of selecting t nodes as central nodes from all controlled nodes includes:

[0008] Select n satellites with the longest average normal working time from all satellite nodes as the central node candidate set;

[0009] The weight of each satellite node in the central node candidate set is calculated according to the remaining average normal working time of the satellite and the uplink and downlink distance of the satellite;

[0010] Select t satellites as the central satellites. The selection process includes:

[0011] max(△ i +μ×mem i +ν×cpu i )

[0012] Among them, Δ i is the weight of the ith central satellite, μ is the storage resource weight of the satellite, mem i is the available storage resource of the ith central satellite, v is the number of satellite nodes, cpu i is the available computing resources of the i-th central satellite.

[0013] Furthermore, when other nodes are divided into corresponding central nodes to form t domains, the current satellite node is allocated to the domain where the central satellite node corresponding to the largest allocation coefficient is located according to the allocation score between the current satellite node and the central satellite node. When an orbital cycle is initialized, the allocation coefficient is calculated according to the distance between the current satellite node and the central satellite node, computing resources, storage resources and bandwidth requirements. If the system load is unbalanced in other time slices, the allocation coefficient is updated according to the load of the central node, otherwise the current allocation coefficient is maintained, where the higher the load of the central node, the lower the weight, and the lower the load, the higher the weight.

[0014] The present invention is applicable to efficient routing, resource allocation, load balancing and link fault recovery in low-orbit satellite (LEO) networks. The present invention combines logical centralized and distributed control strategies. By performing global routing calculations in the ground control center and combining the SRv6 protocol and routing algorithm on the satellite nodes, the satellite nodes can make real-time routing adjustments based on the local link status, thereby improving the response speed and stability of the network. On the other hand, the present invention selects relatively stable satellites as control nodes of the domain through a dynamic domain partitioning algorithm to ensure the high efficiency of data transmission within the domain. In summary, while meeting the high flexibility requirements, the present invention significantly reduces the data transmission delay and routing update burden caused by network topology changes and link instability, thereby improving the overall performance of the low-orbit satellite network. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 A schematic diagram of a single-layer satellite network consisting of a LEO satellite segment and a ground segment of the present invention;

[0016] Figure 2 A physical connection diagram of the satellite network hybrid control architecture of the present invention;

[0017] Figure 3 The logical architecture of the low-orbit satellite network hybrid control architecture of the present invention;

[0018] Figure 4 It is a global load balancing flow chart of the present invention;

[0019] Figure 5 This is a flow chart of the dynamic domain partitioning algorithm of the present invention. DETAILED DESCRIPTION

[0020] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0021] The present invention proposes a satellite network hybrid control method based on SRv6, wherein a ground control center sends a control instruction to a satellite node, and the satellite node forwards a data packet according to the control instruction. The method comprises the following steps:

[0022] In one orbital period, the routing strategy of satellite nodes is planned with the goal of minimizing link costs, and the global routing strategy of all satellite nodes is obtained;

[0023] In a time slice of an orbital cycle, t nodes are selected from all satellites as central nodes, and other nodes are divided into corresponding central nodes to form t domains;

[0024] According to the divided domains, routing optimization is performed with the goal of minimizing the transmission cost between satellite nodes in the domain.

[0025] like Figure 1 As shown in Figure 1, the topology of the LEO satellite network changes frequently due to the high-speed movement of the satellites. Satellite nodes communicate through inter-satellite links (ISLs), and ground stations also participate in some communication management. Therefore, the network topology is dynamic and constantly updated, making it difficult for traditional routing mechanisms to adapt. The SRv6 protocol can better adapt to these topology changes through segment routing technology.

[0026] In the SRv6-based routing optimization module of the present invention, the data packet carries a segment routing header (SRH). The segment routing header structure is:

[0027] SRH={S1,S2,…,S M}

[0028] The SRH segment routing header contains a series of path segments S1, S2, S3…S M , each segment defines the nodes that the data packet needs to pass through in the network. Each segment S in the segment list i For each satellite node, the data packet will be forwarded in the order of the segment list. Under the SRv6 routing mechanism, when each satellite node receives a data packet, it first reads the current segment in the segment routing header and determines the forwarding direction of the data packet based on the segment. The satellite node selects the appropriate next-hop port for data forwarding by simply reading and parsing the information in the SRH segment list. Compared with the traditional global routing search method, this SRH-based segment routing mechanism greatly reduces the complexity and time consumption of path search. This enables the system to respond quickly to changes in network topology and improves path search efficiency, especially in the case of satellite position changes that often occur in LEO networks.

[0029] like Figure 2 The figure shows the physical connection of the hybrid control architecture proposed for a single-layer satellite network. The ground control center remotely controls the satellite to calculate the routing table, and sends it to the LEO satellite through the gateway station. The LEO satellite deploys a routing algorithm to calculate the routing table in the local domain. The hybrid control architecture of the present invention divides the single-layer satellite network into three logical planes, such as Figure 3As shown in the figure, they are the application plane, control plane and data plane. The application plane can be configured with flexible services such as topology reconstruction, fault recovery, and traffic balancing. Specific services are deployed on the application plane to provide an efficient mobility and switching management mechanism for the continuity of the connection served by the LEO satellite. The control plane consists of the ground satellite control center, the gateway station and the control part of each satellite. The ground satellite control center is responsible for monitoring the status of the entire LEO satellite through the gateway station and effectively collecting and analyzing the traffic in the network. The gateway station is responsible for receiving and compiling control signaling from the ground satellite control center and sending the compiled data stream to the LEO satellite, which is recognized by the southbound interface protocol.

[0030] The routing algorithm is deployed on the LEO satellite to perform local calculations in the local domain. Since it is extended based on the SDN architecture, the control plane is divided into two parts: centralized control and distributed routing control. The centralized control part interacts with the application plane through the northbound interface and interacts with the data plane through the southbound interface. The centralized part of the control plane is responsible for some global management, such as global routing calculation, topology discovery and maintenance, etc., while the distributed routing control part is responsible for fast local adaptive adjustment of real-time changes based on global management.

[0031] In the present invention, LEO satellite network load balancing is divided into global load balancing and local load balancing. The present invention proposes a hybrid traffic balancing solution that combines global traffic balancing with local traffic balancing strategies. A large number of predictable traffic change baselines determined by deterministic user geographical distribution and stable daily average traffic change intensity are the main part of network traffic demand. In contrast, a small amount of traffic fluctuations are unpredictable and are the main part of local traffic changes. Based on the above characteristics, the hybrid traffic balancing solution is divided into two steps of design. The first step is to complete the optimal traffic path for the entire network, and design a global load balancing solution based on predictable satellite motion trajectories for the entire network at the ground control center. In the second step, a single node uses the central satellite and transit satellite in the local domain to adjust the routing within the local domain in real time for unpredictable small-particle traffic fluctuations, and then reports the results to the ground control center.

[0032] The present invention designs a path preprocessing scheme based on time slices. The scheme obtains the average flow of the routing path in the past period of time in real time, and adjusts the link state in real time according to the load situation to achieve global flow load balancing. Based on the principle that the orbital period of low-orbit satellites is predictable, an orbital period T is divided into multiple equal time intervals, and the link topology remains unchanged within a single time interval. Such a time interval can be called a time slice. In a time slice, the shortest path of the inter-satellite routing is calculated based on the distance of all satellites.

[0033] like Figure 4The figure shows the flow chart of global load balancing. For a satellite link, the load ratio C port It can be calculated using the following formula:

[0034]

[0035] Among them, T ma Indicates the time slice time, M size Indicates the size of each packet, S port The data rate representing the satellite's motion.

[0036] The link cost can be calculated using the following formula:

[0037]

[0038] Among them, L′ budget represents the link cost of the global routing strategy; L budget α represents the cost of the link formed by the satellite node. Those skilled in the art may calculate the parameter by link communication cost, transmission cost, etc. or by any existing method in the art, and the present invention is not limited thereto. limit represents the link detection threshold, β cost Represents the link cost factor. Based on the above formula, data packets with shorter transmission distances will preferentially choose paths with shorter physical distances for transmission. When services with longer transmission distances bypass high-load areas, the optimal path under low-load pressure is selected on this basis.

[0039] For unpredictable traffic fluctuations in each domain, the central satellite in each domain monitors the link status (including bandwidth utilization, delay, packet loss rate and other parameters) in real time, and dynamically adjusts the data transmission path when the link status reaches the preset threshold. Key indicators of link status such as bandwidth utilization U(t) and delay D(t) can be expressed by the following formula:

[0040]

[0041] D(t)=D trans +D queue +D prop

[0042] Among them, B used (t) represents the bandwidth used by the current link, B total is the total bandwidth of the link, d trans , D queue , D prop They represent transmission delay, queuing delay and propagation delay respectively. When the bandwidth utilization U(t) exceeds the threshold U threshOr when the delay D(t) increases to more than a predetermined value, the central satellite will trigger a local adjustment mechanism. According to the current link status and traffic distribution, the central satellite will select multiple paths for traffic diversion. The segment routing header SRH of the data packet will be updated according to the newly selected path, so that the data packet can be transmitted through the path with lighter load. The specific traffic diversion decision can be based on the dynamic adjustment of the link weight W, where the weight W ij (t) represents the transmission cost from link i to j, and the calculation formula is:

[0043] W ij (t) = α × D (t) + β × U (t)

[0044] Among them, W ij (t) represents the transmission cost between satellite node i and satellite node j in the current domain; α and β are weight coefficients, indicating the influence of delay and bandwidth utilization on path selection. When the weight of a link increases, the data packet will give priority to other paths with smaller weights, thereby achieving local load balancing and reducing the burden of global control. Through the distributed real-time adjustment mechanism, the network can respond quickly to sudden traffic fluctuations, thereby improving the robustness and fault tolerance of the network.

[0045] The present invention uses dynamic domain partitioning, and selects some relatively stable satellites with rich computing and storage resources as the satellite control plane of the satellite network during the partitioning, while the ground control center is still the main control plan. The low-orbit satellites are domain-divided, and their central satellites are selected in each domain, and the status information in the local domain is collected and the path is calculated. The domain ground control center exchanges data and updates the forwarding routing table after any change in the link status.

[0046] Taking into account the instability of the satellite network topology, it is necessary to select a suitable central satellite and divide the satellite domains reasonably. In the present invention, the satellite topology is defined as an undirected connected graph G (V, E, C) and the number of orbits is defined. In addition, since the ISL areas in the polar regions of the earth will be closed, satellites over these areas will not be able to transmit data. Therefore, satellites passing over the two poles of the earth are unstable. When dividing the regions, care should be taken to avoid having all unstable satellites in one area. Since the failure of the central satellite may cause great problems, it is necessary to select some younger satellites that do not pass through the polar regions as central satellites, which helps to reduce the possibility of failure. If the normal working time of the satellite from the last failure is called the age of the satellite (age), and the average interval between two consecutive failures of the satellite is called the average age of the satellite (avg_age), then the remaining average normal working time of the satellite is:

[0047] work_time i =avg_age-agei

[0048] Where avg_avg is the average time interval between satellite failures, ege i The working time since the satellite last failed.

[0049] The candidate set for selecting the central satellite should contain n satellites with the longest average normal working time. First, assign a weight Δ to each satellite in the candidate set. i , which is calculated as follows:

[0050] Δ i =α×work_time i +β×UDL i

[0051] Among them, α is the operating time weight of the central satellite, β is the UDL (uplink and downlink) distance weight of the central satellite, and UDL i is the satellite's UDL (uplink and downlink) distance / length.

[0052] After sorting in descending order of weight, select the first t pairs of satellites as the central satellite i of t domains, ensuring:

[0053] max(Δ i +μ×mem i +v×cup i )

[0054] Among them, μ is the storage resource weight of the satellite, mem i is the available storage resource of the satellite, v is the number of nodes, cpu i The available computing resources for the satellite.

[0055] After the central satellite is selected, the next focus is on the regional division algorithm of ordinary nodes. In the present invention, the node allocation process is first optimized based on satellite resources (computing resources, storage resources, bandwidth) and the physical distance between the node and the central satellite. By reasonably dividing the domain and allocating nodes, efficient utilization of network resources and load balancing can be ensured. After the initial division, ordinary nodes dynamically adjust their control domains according to the resource capabilities of the central satellite and changes in the network topology to ensure the flexibility and robustness of the network. The overall process is as follows: Figure 5 shown.

[0056] First, calculate the physical distance d between each ordinary node i and the central satellite j ij Considering the dynamic nature of the satellite network, the distance depends on the orbital position of the satellite. The specific calculation formula is:

[0057] d ij =R×arccos(sin(θ i)×sin(θ j )+cos(θ i )×cos(θ j )×cos(φ i -φ j ))

[0058] Among them, θ i ,θ j are the latitudes of satellites i and j, respectively, φ i ,φ j is their longitude, R is the radius of the earth, and this formula estimates the physical distance between satellite nodes through the spherical distance calculation method. The closer the distance, the higher the node priority.

[0059] The resource requirements of each node i include the computing capacity C req,i , storage space S req,i and bandwidth requirement B req,i ,These requirements are usually determined by the application tasks or data transmission rates of the nodes. The specific calculation method is as follows:

[0060] C rerq,i =Task complexity × calculation time

[0061] S req,i = Buffer size + Packet size

[0062]

[0063] These requirements will affect the resource allocation of nodes. The higher the node requirements, the greater the possibility of being allocated to a central satellite with sufficient resources. ij It is the result of comprehensive consideration of satellite distance, computing resources, storage resources and bandwidth requirements. The formula is as follows:

[0064]

[0065] Among them, d ij is the distance between ordinary satellite node i and central satellite j. The smaller the distance, the higher the score. j , S j , B j are the computing, storage and bandwidth resources of the central satellite j. α, β, γ and δ are weighted coefficients used to balance the priorities of different resources. The size of the coefficients is set according to different network optimization goals.

[0066] In order to ensure the load balance of each central satellite, check the resource utilization of each central satellite j, and the calculation formula is:

[0067]

[0068] Among them, D j It is the set of nodes managed by the central satellite j. If the resource utilization of a central satellite exceeds the preset threshold, it is necessary to re-evaluate the node allocation and reallocate some ordinary satellite nodes to central satellites with more sufficient resources to avoid overload.

[0069] When the satellite network topology changes (such as satellite position, link status changes, etc.), that is, in the present invention, the balance standard value of the current central satellite node j exceeds the set range, the node allocation will be dynamically adjusted, and the allocation score of node i will be recalculated based on the real-time link status and resource conditions:

[0070]

[0071] Among them, load factor j It is a factor of the current load of the central satellite. If the load of the central satellite is high, its allocation score will decrease, and nodes will be allocated to central satellites with lower loads first.

[0072] The system will perform load balancing adjustments based on the load conditions of each domain in the network. The load imbalance criteria are:

[0073]

[0074] Among them, loadimbalance j is the balance standard value of the current central satellite node j, which measures the difference between the load of satellite j and the overall network load, reduces the load imbalance, and thus optimizes the node allocation; B req,i is the bandwidth resource of satellite node i, B j is the bandwidth resource of the central satellite node j, B req,k is the bandwidth resource of the central satellite node k; N is D j In addition to the central satellite node j, other satellite nodes in B total is the total bandwidth requirement of all nodes in the network, D j are all satellite nodes including the central satellite node j in the domain where the current central satellite node j is located.

[0075] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A satellite network hybrid control method based on SRv6, wherein a ground control center sends a control instruction to a satellite node, and the satellite node forwards a data packet according to the control instruction, characterized in that: The method comprises the following steps: In one orbital period, the routing strategy of satellite nodes is planned with the goal of minimizing link costs, and the global routing strategy of all satellite nodes is obtained; In a time slice of an orbital cycle, t nodes are selected from all satellites as central nodes, and other nodes are divided into corresponding central nodes to form t domains; According to the divided domains, routing optimization is performed with the goal of minimizing the transmission cost between satellite nodes in the domain.

2. According to claim 1, a satellite network hybrid control method based on SRv6 is characterized in that: When planning the routing strategy for satellite nodes, the link cost is expressed as: Among them, L′ budget represents the link cost of the global routing strategy; L budget represents the cost of the link formed by the satellite node; α limit represents the link detection threshold, β cost represents the link cost factor; C port Indicates the load ratio of a satellite link.

3. The SRv6-based satellite network hybrid control method according to claim 2, characterized in that: The load ratio of a satellite link is C port It is expressed as: Among them, T ma Indicates the time slice time, M size Indicates the size of each packet, S port The data rate representing the satellite's motion.

4. The SRv6-based satellite network hybrid control method according to claim 1, characterized in that: The process of selecting t nodes from all controlled nodes as central nodes includes: Select n satellites with the longest average normal working time from all satellite nodes as the central node candidate set; The weight of each satellite node in the central node candidate set is calculated according to the remaining average normal working time of the satellite and the uplink and downlink distance of the satellite; Select t satellites as the central satellites. The selection process includes: max(D i +μ×mem i +v×cpu i ) Among them, Δ i is the weight of the ith central satellite, μ is the storage resource weight of the satellite, mem i is the available storage resource of the ith central satellite, v is the number of satellite nodes, cpu i is the available computing resources of the i-th central satellite.

5. The SRv6-based satellite network hybrid control method according to claim 4, characterized in that: The weight of the i-th central satellite Δ i It is expressed as: D i =α×work_time i +β×UDL i Among them, α is the operating time weight of the central satellite, work_time i is the remaining average normal working time of the i-th central satellite, β is the uplink and downlink distance weight of the central satellite, UDL i is the uplink and downlink distance of the ith central satellite.

6. The SRv6-based satellite network hybrid control method according to claim 5, characterized in that: The remaining average normal working time of the i-th central satellite is the average interval between its two consecutive failures minus the normal working time since its last failure.

7. The SRv6-based satellite network hybrid control method according to claim 1, characterized in that: When other nodes are divided into corresponding central nodes to form t domains, the current satellite node is allocated to the domain where the central satellite node with the largest allocation coefficient is located according to the allocation score between the current satellite node and the central satellite node. When an orbital cycle is initialized, the allocation coefficient is calculated according to the distance between the current satellite node and the central satellite node, computing resources, storage resources and bandwidth requirements. If the system load is unbalanced in other time slices, the allocation coefficient is updated according to the load of the central node, otherwise the current allocation coefficient is maintained. The higher the load of the central node, the lower the weight, and the lower the load, the higher the weight.

8. The SRv6-based satellite network hybrid control method according to claim 7, characterized in that: The allocation score between the current satellite node and the central satellite node is expressed as: Among them, score ij represents the allocation score between satellite node i and central satellite node j; d ij represents the distance between satellite node i and central satellite node j; C req,i is the computing resource of satellite node i, C j is the computing resource of the central satellite node j; S req,i is the storage resource of satellite node i, S j is the storage resource of the central satellite node j; B req,i is the bandwidth resource of satellite node i, B j is the bandwidth resource of the central satellite node j; α, β, γ, δ are the weights of distance, computing resources, storage resources, and bandwidth resources.

9. The SRv6-based satellite network hybrid control method according to claim 7 or 8, characterized in that: Calculate the balance standard value between the current central satellite node and the system. If the standard value exceeds the set range, the current central satellite node load is unbalanced. The calculation of the balance standard value includes: Among them, load imbalance j is the equilibrium standard value of the current central satellite node j; B req,i is the bandwidth resource of satellite node i, B j is the bandwidth resource of the central satellite node j, B req,k is the bandwidth resource of the central satellite node k; n is the number of all satellite nodes, B total is the total bandwidth of the system, D j are all satellite nodes in the domain where the current central satellite node j is located.

10. The SRv6-based satellite network hybrid control method according to claim 1, characterized in that: The transmission cost between satellite nodes in the domain is expressed as: W ij (t)=α×D(t)+β×U(t) D(t)=D trans +D queue +D prop Among them, W ij (t) represents the transmission cost between satellite node i and satellite node j in the current domain; α and β are weight coefficients, U(t) represents the bandwidth utilization of the current link, and D(t) represents the delay of the current link; B used (t) represents the bandwidth used by the current link, B total is the total bandwidth of the link; D trans Denotes the transmission delay, D queue represents the queue delay, D prop Represents the propagation delay.

Images