Inter-satellite routing method based on computing resources and link states

Abstract

The invention relates to an inter-satellite routing method based on computing resources and link states. The inter-satellite routing method comprises the following steps: S1, a current node periodically broadcasts a HELLO message to a one-hop node of the current node; s2, the current node determines a neighbor node set according to the HELLO message; the HELLO message comprises the computing resource degree and the position information of the current node and the computing resource degree and the link stability of the neighbor node of the current node; s3, the current node performs MPR node selection based on an improved routing protocol of the computing resource degree and the link stability; s4, the current node periodically sends and forwards TC messages in the whole network range through the selected MPR node to obtain topological information, and a topological database is formed; the TC message comprises computing resource degree and link stability information of neighbor nodes of the current node; and S5, the current node calculates and maintains a routing table according to the MPR node set and the topology database and a Djisktra algorithm.

Description

Technical Field

[0001] The present invention relates to the technical field of communication networks, and particularly to an inter-satellite routing method based on computing resources and link states. Background Art

[0002] With the booming development of new-generation communication and network technologies such as 5G and Wi-Fi6, as well as the improvement of launch technologies, the maturity of terrestrial Internet and Internet of Things, and the development needs of 6G technology and future space missions, large-scale low-Earth orbit Internet constellations have once again become a research hotspot in related industries.

[0003] Currently, with the construction of low-Earth orbit satellite constellations, the number of satellites in space and the types of network formations supported by satellites will gradually become richer. Due to constraints such as short intersection times of satellites in different orbital planes, reduced link survival times, and extended cross-layer network formation processing times, network topology changes occur continuously and frequently, weakening the periodicity of satellite operations. As user requirements and task complexity continue to increase, the capabilities and resources of individual satellite nodes are gradually reaching saturation and can no longer meet the growing task requirements. Therefore, the integration of computing and networking to study efficient and robust routing algorithms to address issues such as high-dynamic, weak-periodic topology changes, network congestion, and load balancing caused by uneven traffic flow distribution has become a new trend.

[0004] Currently, traditional virtual topology-based routing protocols use servers on the ground segment to centrally calculate and generate routing tables for each time period. The routing tables for all times are stored on the satellite and uploaded for updates regularly. The calculation process overly relies on centralized computing by ground stations. In high-dynamic, large-scale low-Earth orbit satellite scenarios, interruptions in space-ground communication will limit the flexibility and autonomy of routing; solidifying the routing topology within a certain period of time by dividing time slices may cause problems such as link switching and routing reconstruction due to high-speed node movement. At the same time, when services are running on nodes with network topology switching or reconstruction and are offloading tasks through expired routing tables, accurate task offloading and collaborative processing cannot be performed, resulting in task processing failures and resource waste; it is difficult to use link information to evaluate the remaining communication duration of neighbor nodes, leading to low efficiency in collaborative processing of compute-intensive tasks.

[0005] Therefore, the current development of space collaborative computing technology faces problems such as low computing task scheduling efficiency and insufficient computing resource utilization, and there is an urgent need for an inter-satellite routing method based on computing resources and link states. Summary of the Invention

[0006] To solve the above technical problems existing in the prior art, the object of the present invention is to provide an inter-satellite routing method based on computing resources and link status, which can find an offloading path with stable links and rich computing resources for on-satellite computing tasks, thereby reducing the task completion delay and improving the service quality.

[0007] To achieve the above object of the invention, the present invention provides an inter-satellite routing method based on computing resources and link status, including the following steps:

[0008] Step S1, the current node periodically broadcasts HELLO messages to its one-hop neighbors;

[0009] Step S2, the current node determines the neighbor node set according to the HELLO messages; the HELLO messages include the computing resource degree, link stability degree and location information of the current node, as well as the computing resource degree and link stability degree of its neighbor nodes;

[0010] Step S3, the current node selects MPR nodes based on an improved routing protocol of computing resource degree and link stability degree;

[0011] Step S4, the current node obtains topology information by periodically sending and forwarding TC messages through the selected MPR nodes in the whole network and forms a topology database; the TC messages include the computing resource degree and link stability degree information of the neighbor nodes of the current node;

[0012] Step S5, the current node calculates and maintains a routing table according to the MPR node set and the topology database according to the Djisktra algorithm.

[0013] According to a technical solution of the present invention, the HELLO message includes a Willingness field, a Link Code field, an Htime field, a Reserved field, a Link Message Size field and an idle field, and the idle field of the HELLO message includes the computing resource degree, link stability degree and location information of the satellite node, as well as the computing resource degree and link stability degree of the neighbor nodes of the satellite node;

[0014] The TC message includes a Reserved field, a Link Code field, a Link Message Size field, an Advertised Neighbor Address field and an idle field, and the idle field of the TC message includes the computing resource degree and link stability degree of the neighbor nodes of the satellite node.

[0015] According to a technical solution of the present invention, in the step S3, the computing resource degree of the satellite node includes the CPU computing resource degree, the GPU computing resource degree and the memory computing resource degree;

[0016] The calculation formula for the CPU computing resource degree is as follows:

[0017]

[0018] where idle cpu is the percentage of idle time of the satellite node CPU, ipc is the number of instructions completed per clock cycle of the tasks of the satellite node CPU, and load cpu is the average load of the satellite node CPU;

[0019] The calculation formula for the GPU computing resource degree is as follows:

[0020]

[0021] where idle gpu is the percentage of idle time of the satellite node GPU, flops is the floating-point computing power of the satellite node GPU, representing the number of floating-point operations per second, and load gpu is the average load of the satellite node GPU;

[0022] The calculation formula for the memory computing resource degree:

[0023]

[0024] where total is the total physical memory size of the satellite node memory, uised is the occupied physical memory size of the satellite node memory, cached and buffer are the memory sizes occupied by two types of caches of the satellite node memory, and pagefault is the page fault frequency of the tasks of the satellite node memory;

[0025] The calculation formula for the computing resource degree of the satellite node is as follows:

[0026] R = α 1 R cpu + α 2 α + α 3 R memory

[0027] where α 1 、α 2 and α 3 are the weights of the CPU computing resource degree, the GPU computing resource degree, and the memory computing resource degree respectively.

[0028] According to a technical solution of the present invention, in the step S3, the calculation method of the link stability is as follows:

[0029] Assume that at time t 1 the absolute position information x 1 、y1 Transmitted according to the first HELLO message, t 2 The absolute position information x at the moment 2 and y 2 , Transmitted according to the second HELLO message, and the time interval is the HELLO message transmission interval. Assuming the maximum communication distance between two satellites is |XY|, calculate the link survival time of satellite node X and its neighbor node Y within the HELLO message transmission interval, expressed as:

[0030]

[0031] where |xy| represents the current distance between satellite node X and satellite node Y, and α represents the included angle with , and v xy represents the relative velocity of satellite node X relative to its neighbor node Y;

[0032] Then the calculation formula for the link stability between satellite node X and its neighbor node Y is:

[0033]

[0034] where avg(t yz ) is the average link survival time between satellite node X and neighbor node Y, and its calculation formula is: n is the number of times of position information in the HELLO packet messages of neighbor node Y received by satellite node X; T max is the maximum existence time of the link between satellite node X and its neighbor node Y under the maximum communication distance.

[0035] According to one technical solution of the present invention, in the step S3, the selection of the MPR node specifically includes:

[0036] Step S31: Obtain all neighbor nodes of the current node and the one-hop neighbor node set N1 and two-hop neighbor node set N2 formed by its neighbor nodes to form the current node;

[0037] Step S32: Calculate the connection degree of the one-hop neighbor nodes in the one-hop neighbor node set N1;

[0038] Step S33: Add the one-hop neighbor nodes with a willingness value of 7 in the one-hop neighbor node set N1 to the MPR node set M of the current node;

[0039] Step S34: For the one-hop neighbor nodes with a willingness value less than 7 in the one-hop neighbor node set N1, the following algorithm is used for selection:

[0040] Step S341: If there are isolated points in the two-hop neighbor node set N2, add the one-hop neighbor nodes corresponding to the isolated points to the MPR node set M;

[0041] Step S342: If the two-hop neighbor node set N2 is non-empty, perform the following operations:

[0042] 1) Calculate the reachability of one-hop neighbor nodes in the one-hop node set N1.

[0043] 2) Select the one-hop neighbor node with the largest reachability and add it to the MPR node set M, then return to Step S342;

[0044] 3) If there are multiple one-hop neighbor nodes with the largest reachability, calculate the comprehensive evaluation index of these multiple one-hop neighbor nodes with the largest reachability, add the one-hop neighbor node with the largest comprehensive evaluation index value to the MPR node set M, and delete the two-hop neighbor nodes connected to it from the two-hop neighbor node set N2, then return to Step S342;

[0045] Step S343: If the two-hop neighbor node set N2 is empty, end the algorithm;

[0046] Among them, the calculation formula for the comprehensive evaluation index of one-hop neighbor nodes is:

[0047] RS = β 1 R + β 2 S

[0048] In the formula, β 1 and β 2 are respectively the weight of the computing resource degree of the one-hop neighbor node and the weight of the link stability degree between the one-hop neighbor node and the current node.

[0049] According to a technical solution of the present invention, in the said Step S5, the routing calculation specifically includes:

[0050] Step S51: Initialize the topology data table;

[0051] Step S52: Initialize the routing table;

[0052] Step S53: Judge whether it is necessary to perform routing calculation according to the received latest HELLO message and TC message;

[0053] Step S54: If so, update the topology database according to the HELLO message and TC message; otherwise, end the algorithm;

[0054] Step S55: Calculate the optimal path according to the updated topology database and neighbor table;

[0055] Step S56: Update the routing table according to the calculated optimal path.

[0056] According to a technical solution of the present invention, in step S55, the calculation of the optimal path is performed using the Djisktra algorithm, which specifically includes:

[0057] Step S551, initialization: create a node list and a current node, add all neighbor nodes in the topology database to the node list; add the current node to the node list; initialize the metrics of all satellite nodes in the node list to infinity; set the metric of the current node to 0; initialize the set of visited nodes as an empty set;

[0058] Step S552, select the satellite node with the smallest metric in the node list as the current processing node, and remove the current processing node from the node list and add it to the set of visited nodes;

[0059] Step S553, traverse each neighbor node of the current processing node in the topology database. If the neighbor node has not been added to the set of visited nodes, then execute:

[0060] Calculate the new metric from the current processing node to the neighbor node, where the calculation method of the new metric is:

[0061] New metric = the metric of the current processing node + the comprehensive metric of the neighbor node and the current processing node in the topology database;

[0062] The calculation formula of the comprehensive metric of the topology database is:

[0063] Newmetric ij = w 1 ·R j + w 2 ·S ij + w 3 ·Oldmetric

[0064] where i represents the current processing node, j represents the neighbor node of the current processing node i, R j represents the computing resource degree of the neighbor node j, S ij represents the link stability between the current processing node i and its neighbor node j, and Oldmetric is the distance metric value calculated by the traditional method;

[0065] If the new metric is less than the metric of the neighbor node, update the metric of the neighbor node;

[0066] Step S554, return to step S552 until the node list is not empty.

[0067] Compared with the prior art, the present invention has the following beneficial effects:

[0068] In view of the problems faced by the current development of space collaborative computing technology, such as low computing task scheduling efficiency and insufficient utilization of computing resources, the present invention proposes an inter-satellite routing method based on computing resources and link status. In the face of the characteristics of network dynamics, time-variation, and decentralized and heterogeneous computing power in the constellation environment, the protocol workflow, format modification of Hello packets and TC packets, MPR selection algorithm, and optimal routing calculation algorithm are designed. By introducing node computing resource information and link information as optimization routing protocol metrics, computing power resource awareness at the network layer helps to construct the shortest computing delay path, effectively utilize node resources on the edge or path for task scheduling, and achieve the construction of stable routing in the low-earth orbit satellite constellation and the efficient utilization of heterogeneous satellite computing resources. By introducing the definitions of two parameters, computing resource degree and link stability degree, the measurement of computing power and network characteristics is realized.

[0069] The present invention adopts an improved routing protocol based on computing resource degree and link stability degree, uses node computing power information and link information to construct stable routing and support efficient collaborative task processing, and solves the problems such as routing failure, task offloading failure, excessive computing delay, and large traffic of redundant routing topologies caused by the difficulty in perceiving the computing resources of adjacent nodes and the difficulty in constructing stable routing in satellite networking. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0071] Figure 1 Schematically showing the flowchart of the inter-satellite routing method based on computing resources and link status provided in the embodiment of the present invention;

[0072] Figure 2 Schematically showing the workflow diagram of the traditional OLSR protocol;

[0073] Figure 3 Schematically showing the format schematic diagram of the HELLO data packet in the inter-satellite routing method based on computing resources and link status provided in the embodiment of the present invention;

[0074] Figure 4 Schematically showing the format schematic diagram of the TC data packet in the inter-satellite routing method based on computing resources and link status provided in the embodiment of the present invention;

[0075] Figure 5 Schematically showing the schematic diagram of relative satellite motion in the embodiment of the present invention;

[0076] Figure 6 Schematic representation of a flowchart of a routing calculation process in an inter-satellite routing method based on computing resources and link status provided in an embodiment of the present invention. Detailed implementation manners

[0077] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0078] The present invention will be described in detail below in conjunction with the drawings and specific implementation manners. The implementation manners cannot be elaborated one by one here, but the implementation manners of the present invention are not limited to the following implementation manners.

[0079] As Figure 1 shown, an inter-satellite routing method based on computing resources and link status provided by the present invention includes the following steps:

[0080] Step S1: The current node periodically broadcasts HELLO messages to its one-hop neighbors;

[0081] Step S2: The current node determines a neighbor node set according to the HELLO messages; the HELLO messages include the computing resource degree and location information of the current node and the computing resource degree and link stability degree of its neighbor nodes;

[0082] The basic working process of the existing OLSR is as Figure 2 shown. The working process includes a neighbor node sensing process and a node-to-node routing establishment and maintenance process. First, the current node will periodically sense the information of surrounding nodes through the flooding of HELLO messages, and update the neighbor table according to the update of the HELLO messages, and perform the MPR selection algorithm based on this. Among them, the object of MPR selection is a subset of the nodes selected by the node from its neighbors, so that the broadcast messages forwarded by these selected subset nodes will be received by all nodes two hops away. After the node executes the MPR selection algorithm, it periodically updates the topology table according to the TC messages to ensure that each node obtains sufficient link status information, and finally calculates the routing table step by step through the node topology table information. In actual application scenarios, low-earth orbit satellites often can only select a few other satellites with relatively close distances around them to cooperate in processing computing services, and generally do not select to cooperate with satellites with relatively far link distances through multiple forwards, so as to avoid strong inter-satellite channel interference and network jitter and data loss caused by the position movement of satellites with relatively far distances. It is necessary to consider the ability to sense the computing resources of neighbor nodes and construct a stable routing.

[0083] In order to improve the efficiency of inter-satellite collaborative computing, the present invention proposes an inter-satellite routing protocol OLSR-RS based on computing resources and link status perception. On the basis of the OLSR protocol, by modifying the packet format, MPR selection algorithm and routing calculation algorithm, it can find an offloading path with stable links and rich computing resources for on-board computing tasks, thereby reducing the task completion delay and improving service quality.

[0084] In the traditional OLSR protocol, the Willingness field of the HELLO packet is used to indicate the willingness of a node to participate in the MPR node selection. The value range of the Willingness field (denoted as) is an integer from 0 to 7, where w 0 =0 means never participating in MPR, w 0 =7 means always selected as MPR. In the original protocol, most nodes are in the default state, that is, w 0 = 3. The ANSN field of the TC packet represents the advertised neighbor sequence number, which is incremented each time a node detects a change in its advertised neighbor set. The Advertised Neighbor Address field contains the primary address of the neighbor node. All primary addresses of the advertised neighbors of the originating node are placed in the TC message.

[0085] like Figure 3 As shown, in the present invention, the HELLO message includes a Willingness field, a Link Code field, a Htime field, a Reserved field, a Link Message Size field and an idle field. The idle field of the HELLO message includes the computing resource degree and location information of the current node and the computing resource degree and link stability of the neighboring node of the current node;

[0086] like Figure 4 As shown, the TC message includes a Reserved field, a Link Code field, a Link Message Size field, an Advertised Neighbor Address field, and an idle field. The idle field of the TC message includes the computing resource degree and link stability of the neighbor node of the current node.

[0087] The present invention adds the computing resource degree R and location information X of the current node, as well as the computing resource degree and link stability information of each neighbor node in the free field of the Hello packet, wherein the link stability information is calculated based on the historical location information in the neighbor Hello packet. The TC packet adds the computing resource degree and link stability information to the broadcasted neighbor information, so as to utilize the computing resource degree and link stability of the satellite node in the satellite network to perform MPR selection and routing calculation.

[0088] The computing power of satellite nodes is very important because it determines the role and value of nodes in the network. Quantifying the computing power of nodes can help other nodes and users better understand the contributions and capabilities of the nodes, and at the same time, it can also provide better reward and incentive mechanisms for the nodes. In the present invention, the computing power of satellite nodes in a satellite network is measured by the computing resource degree, and the computing resource degree of a satellite node includes the CPU computing resource degree, the GPU computing resource degree, and the memory computing resource degree;

[0089] The calculation formula for the CPU computing resource degree is:

[0090]

[0091] where idle cpu is the percentage of idle time of the satellite node's CPU, ipc is the number of instructions completed per clock cycle of the satellite node's CPU task, and load cpu is the average load of the satellite node's CPU;

[0092] The calculation formula for the GPU computing resource degree is:

[0093]

[0094] where idle gpu is the percentage of idle time of the satellite node's GPU, flops is the floating-point operation ability of the satellite node's GPU, representing the number of floating-point operations per second, and load gpu is the average load of the satellite node's GPU;

[0095] The calculation formula for the memory computing resource degree:

[0096]

[0097] where total is the total physical memory size of the satellite node's memory, used is the occupied physical memory size of the satellite node's memory, cached and buffer are the memory sizes occupied by two types of caches of the satellite node's memory, and pagefault is the page fault frequency of the satellite node's memory tasks;

[0098] For the on-board heterogeneous computing module, in order to comprehensively evaluate the overall capabilities of multiple resources, a weighted summation method is adopted to combine the dynamic evaluation values of different resources to form a comprehensive computing resource degree. Then, the calculation formula for the computing resource degree of a satellite node is:

[0099] R = α 1 R cpu + α 2 R gpu + α 3 R memory

[0100] where α 1 、α 2 and α 3 are the weights of the CPU computing resource degree, GPU computing resource degree, and memory computing resource degree respectively, reflecting their relative importance in the overall resource evaluation. The above indicators may not be complete, and there may be different evaluation criteria for different application scenarios. Therefore, when quantifying the node computing ability, it is necessary to select appropriate indicators according to the specific application scenario and requirements, and comprehensively consider the weights of various indicators.

[0101] Since the satellite is not an object moving in a two-dimensional plane and cannot be simply described by two-dimensional coordinates, the calculation method of the link stability in the present invention is as follows:

[0102] Assume at time t 1 , the absolute position information x 1 、y 1 of the satellite node X and its neighbor node Y are transmitted according to the first HELLO message, and the absolute position information x 2 、y 2 at time t 2 are transmitted according to the second HELLO message. The time interval is the HELLO message transmission interval. Assume the maximum communication distance between the two satellites is |XY|. The relative motion schematic diagram of the satellite node X and its neighbor node Y is as shown in Figure 5 . Then, the link survival time of the satellite node X and its neighbor node Y within the HELLO message transmission interval is calculated and expressed as:

[0103]

[0104] where |xy| represents the current distance between the satellite node X and the satellite node Y, and α represents the included angle between and , and v xy represents the relative velocity of the satellite node X relative to its neighbor node Y; through the front and back position information of the satellite, the value of α can be obtained using the cosine theorem, and through the distance and the HELLO packet message interval, the value of the relative velocity v xy can be obtained.

[0105] Then, the calculation formula for the link stability between the satellite node X and its neighbor node Y is:

[0106]

[0107] where avg(t yz ) is the average link survival time between the satellite node X and the neighbor node Y, and its calculation formula is: n is the number of times of the location information in the Hello packet messages of all one-hop neighbor nodes in the one-hop neighbor node set of neighbor node Y received by satellite node X; T max is the maximum existence time of the link between satellite node X and its neighbor node Y under the maximum communication distance.

[0108] Step S3: The current node performs MPR node selection based on an improved routing protocol considering computational resource degree and link stability;

[0109] The MPR selection algorithm in the present invention comprehensively considers the computational resources and link dynamics of nodes. In step S3, the MPR node selection specifically includes:

[0110] Step S31: Obtain all neighbor nodes of the current node and the neighbor nodes form the one-hop neighbor node set N1 and the two-hop neighbor node set N2 of the current node;

[0111] Step S32: Calculate the connection degree of the one-hop neighbor nodes in the one-hop neighbor node set N1;

[0112] Step S33: Add the one-hop neighbor nodes with a willingness value of 7 in the one-hop neighbor node set N1 to the MPR node set M of the current node;

[0113] Step S34: For the one-hop neighbor nodes with a willingness value less than 7 in the one-hop neighbor node set N1, the following algorithm is used for selection:

[0114] Step S341: If there are isolated points in the two-hop neighbor node set N2, add the one-hop neighbor nodes corresponding to the isolated points to the MPR node set M;

[0115] Step S342: If the two-hop neighbor node set N2 is non-empty, make the following processing:

[0116] 1) Calculate the reachability of the one-hop neighbor nodes in the one-hop node set N1.

[0117] 2) Select the one-hop neighbor node with the largest reachability and add it to the MPR node set M, and return to step S342;

[0118] 3) If there are multiple one-hop neighbor nodes with the largest reachability, calculate the comprehensive evaluation index of the multiple one-hop neighbor nodes with the largest reachability, add the one-hop neighbor node with the largest comprehensive evaluation index value to the MPR node set M, and delete the two-hop neighbor nodes connected to it from the two-hop neighbor node set N2, and return to step S342;

[0119] Step S343: If the two-hop neighbor node set N2 is empty, the algorithm ends;

[0120] Among them, the calculation formula for the comprehensive evaluation index of one-hop neighbor nodes is as follows:

[0121] RS = β 1 R + β 2 S

[0122] In the formula, β 1 and β 2 are respectively the weights of the computing resource degree of one-hop neighbor nodes and the link stability degree between one-hop neighbor nodes and the current node, reflecting the relative importance of computing resources and link dynamics in the comprehensive evaluation index.

[0123] The pseudo-code for MPR selection is as follows:

[0124] Input: The set N1 of one-hop neighbor nodes of the node, the set N2 of all two-hop neighbor nodes

[0125] Output: The set M of MPRs of the node

[0126] Calculate the connection degree D(j) of all nodes in N1

[0127] Add the nodes in N1 with the willingness value W(j) of 7 to the set M

[0128] while W(j) < 7 do

[0129] Add the N1 nodes that connect the isolated nodes in N2 to M

[0130] Delete the nodes in N2 that can be connected by M

[0131] while N2 is not empty do

[0132] Add the node with the maximum connection degree D among the N1 nodes that connect each N2 node to M

[0133] if multiple one-hop nodes have the same D

[0134] Calculate the comprehensive evaluation index RS value

[0135] Add the one-hop neighbor node with the maximum RS value to M

[0136] Delete the two-hop neighbor nodes connected by this one-hop node from N2

[0137] end if

[0138] end while

[0139] end while

[0140] Step S4: The current node obtains topology information by periodically sending and forwarding TC messages through the selected MPR nodes across the network and forms a topology database; the TC message includes the computing resource degree and link stability degree information of the neighbor nodes of the current node.

[0141] Step S5: The current node calculates and maintains a routing table according to the MPR node set and the topology database based on the Djisktra algorithm.

[0142] The OLSR routing table is calculated based on the information in the neighbor table and the topology database. The MPR nodes are responsible for periodically broadcasting topology control (TC) packets. Each node collects the network-wide topology information by receiving TC packets and forms a topology database. As Figure 6 shown, in Step S5, the routing calculation specifically includes:

[0143] Step S51: Initialize the topology data table;

[0144] Step S52: Initialize the routing table;

[0145] Step S53: Determine whether routing calculation needs to be performed according to the received latest HELLO message and TC message;

[0146] Step S54: If so, update the topology database according to the HELLO message and the TC message; otherwise, end the algorithm;

[0147] Step S55: Calculate the optimal path according to the updated topology database and the neighbor table;

[0148] Step S56: Update the routing table according to the calculated optimal path.

[0149] In the present invention, the OLSR-RS routing protocol modifies two parts in the traditional OLSR routing calculation process, namely the topology database and the optimal path calculation. When recording the neighbor node information of each node in the topology database, the computing resource degree and link stability degree of the node (obtained through the TC packet) are added. When calculating the metric value of the path in the optimal path calculation algorithm, the shortest path and the computing resource degree and link stability degree of the nodes on the path are considered to form a comprehensive metric value. The Djisktra algorithm is used to calculate the optimal path, which specifically includes:

[0150] Step S551: Initialization: Create a node list and the current node, add all neighbor nodes in the topology database to the node list; add the current node to the node list; initialize the metric of all satellite nodes in the node list to infinity; set the metric of the current node to 0; initialize the set of visited nodes as an empty set;

[0151] Step S552: Select the satellite node with the smallest metric in the node list as the current processing node, remove the current processing node from the node list, and add it to the set of visited nodes;

[0152] Step S553: Traverse each neighbor node of the current processing node in the topology database. If the neighbor node has not been added to the set of visited nodes, then execute:

[0153] Calculate the new metric from the current processing node to the neighbor node, where the calculation method of the new metric is:

[0154] New metric = metric of the current processing node + comprehensive metric of the topology database between the neighbor node and the current processing node;

[0155] The calculation formula of the comprehensive metric of the topology database is:

[0156] Newmetric ij = w 1 ·R j + w 2 ·S ij + w 3 ·Oldmetric

[0157] where i represents the current processing node, j represents the neighbor node of the current processing node i, R j represents the computing resource degree of the neighbor node j, S ij represents the link stability between the current processing node i and the neighbor node j, and Oldmetric is the distance metric value calculated by the traditional method;

[0158] If the new metric is less than the metric of this neighbor node, update the metric of this neighbor node;

[0159] Step S554: Return to Step S552 until the node list is not empty.

[0160] The pseudo-code of the optimal path calculation algorithm is as follows:

[0161] Input: Topology database, current node

[0162] Output: Shortest paths to all target nodes

[0163] Add all neighbor nodes in the topology database to the node list;

[0164] Add the current node to the node list;

[0165] Distance metric = {node: infinity for node in node list};

[0166] Distance metric{current node} = 0;

[0167] Set of visited nodes = empty set;

[0168] while the node list is not empty:

[0169] Current node = the node with the minimum distance metric in the selected node list;

[0170] Remove the current node from the node list;

[0171] Add the current node to the set of visited nodes;

[0172] for each neighbor node in the topology database[current node]:

[0173] if the neighbor node is not in the set of visited nodes:

[0174] New metric = metric[current node] + topology database comprehensive metric[current node][neighbor node];

[0175] if the new metric < metric[neighbor node]:

[0176] metric[neighbor node] = new metric;

[0177] end if

[0178] end if

[0179] end while

[0180] It should also be noted that in this article, the term "including", "comprising" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or terminal device including a series of elements not only includes those elements, but also includes other elements not explicitly listed, or elements inherent to such a process, method, article or terminal device. Without further limitation, an element defined by the statement "including one..." does not exclude the existence of another identical element in the process, method, article or terminal device including the said element.

[0181] Finally, it should be noted that the above is the preferred embodiment of the present invention. It should be pointed out that although the preferred embodiments of the present invention have been described, for those skilled in the art of this technology, once the basic creative concept of the present invention is known, without departing from the principle of the present invention, several improvements and refinements can still be made, and these improvements and refinements should also be regarded as the protection scope of the present invention. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the embodiments of the present invention.

Claims

1. An inter-satellite routing method based on computing resources and link status, characterized in that: The following steps are involved: Step S1, the current node regularly broadcasts a HELLO message to its first-hop node; Step S2: The current node determines a set of neighbor nodes according to a HELLO message; the HELLO message includes the computing resource degree and location information of the current node and the computing resource degree and link stability of its neighbor nodes; Step S3: The current node selects an MPR node based on an improved routing protocol of computing resource degree and link stability; Step S4: The current node periodically sends and forwards TC messages in the entire network through the selected MPR node to obtain topology information and form a topology database; the TC message includes the computing resource degree and link stability information of the neighboring nodes of the current node; Step S5: The current node calculates and maintains the routing table according to the MPR node set and the topology database and the modified Djisktra algorithm.

2. The inter-satellite routing method based on computing resources and link status according to claim 1, characterized in that: The HELLO message includes a Willingness field, a Link Code field, a Htime field, a Reserved field, a LinkMessage Size field, and an idle field. The idle field of the HELLO message includes the computing resource degree and location information of the current node and the computing resource degree and link stability of the neighboring node of the current node; The TC message includes a Reserved field, a Link Code field, a Link Message Size field, an Advertised Neighbor Address field, and an idle field. The idle field of the TC message includes the computing resource degree and link stability of the neighbor node of the current node.

3. The inter-satellite routing method based on computing resources and link status according to claim 1, characterized in that: In step S3, the computing resource degree of the satellite node includes the CPU computing resource degree, the GPU computing resource degree and the memory computing resource degree; The calculation formula for CPU computing resources is: Among them, idle cpu is the percentage of idle time of the satellite node CPU, ipc is the average number of instructions completed per clock cycle of the satellite node CPU task, load cpu is the average load of the satellite node CPU; The calculation formula for GPU computing resources is: Among them, idle gpu is the percentage of idle time of the satellite node GPU, flops is the floating point computing capability of the satellite node GPU, indicating the number of floating point operations per second, load gpu is the average load of the satellite node GPU; The calculation formula for memory computing resources is: Among them, total is the total physical memory size of the satellite node memory, used is the physical memory size occupied by the satellite node memory, cached and buffer are the memory sizes occupied by the two caches of the satellite node memory, and pagefault is the page fault frequency of the satellite node memory task; The calculation formula of the computing resource degree of the satellite node is: R=α1R cpu +α2R gpu +α3R memory Among them, α1, α2 and α3 are the weights of CPU computing resources, GPU computing resources and memory computing resources respectively.

4. The inter-satellite routing method based on computing resources and link status according to claim 3, characterized in that: In step S3, the link stability is calculated as follows: Assume that at time t1, the absolute position information x1 and y1 of satellite node X and its neighbor node Y are transmitted according to the first HELLO message, and the absolute position information x2 and y2 at time t2 are transmitted according to the second HELLO message. The time interval is the HELLO message transmission interval. Assuming that the maximum communication distance between the two satellites is |XY|, the link survival time between satellite node X and its neighbor node Y in the HELLO message transmission interval is calculated, which is expressed as: Among them, |xy| represents the current distance between satellite nodes X and Y, and α represents and The angle, v xy Indicates the relative speed of satellite node X relative to its neighbor node Y; The calculation formula for the link stability between satellite node X and its neighbor node Y is: Among them, avg(t yz ) is the average link survival time between satellite node X and its neighbor node Y, and its calculation formula is: n is the number of times satellite node X receives the location information in the HELLO packet message from its neighbor node Y; T max is the maximum existence time of the link between satellite node X and its neighbor node Y under the maximum communication distance.

5. The inter-satellite routing method based on computing resources and link status according to claim 4, characterized in that: In step S3, the MPR node selection specifically includes: Step S31, obtaining all neighbor nodes of the current node and their neighbor nodes forming a one-hop neighbor node set N1 and a two-hop neighbor node set N2 of the current node; Step S32, calculating the connectivity of one-hop neighbor nodes in the one-hop neighbor node set N1; Step S33: Add the one-hop neighbor node with a willingness value of 7 in the one-hop neighbor node set N1 to the MPR node set M of the current node; Step S34: For the one-hop neighbor nodes whose willingness values ​​are less than 7 in the one-hop neighbor node set N1, the following algorithm is used for selection: Step S341: If there is an isolated point in the two-hop neighbor node set N2, then the one-hop neighbor node corresponding to the isolated point is added to the MPR node set M; Step S342: If the two-hop neighbor node set N2 is not empty, perform the following processing: 1) Calculate the reachable number of one-hop neighbor nodes in the one-hop node set N1; 2) Select the one-hop neighbor node with the largest reachable number to add to the MPR node set M, and return to step S342; 3) If there are multiple one-hop neighbor nodes with the largest reachable number, calculate the comprehensive evaluation index of the multiple one-hop neighbor nodes with the largest reachable number, add the one-hop neighbor node with the largest comprehensive evaluation index value to the MPR node set M, and delete the two-hop neighbor node connected to it from the two-hop neighbor node set N2, and return to step S342; Step S343: If the two-hop neighbor node set N2 is empty, the algorithm ends; Among them, the calculation formula of the comprehensive evaluation index of a one-hop neighbor node is: RS=β1R+β2S Where β1 and β2 are the weight of the computing resource degree of the one-hop neighbor node and the weight of the link stability between the one-hop neighbor node and the current node, respectively.

6. The inter-satellite routing method based on computing resources and link status according to claim 5, characterized in that: In step S5, the route calculation specifically includes: Step S51, initializing the topology data table; Step S52, initializing the routing table; Step S53: judging whether it is necessary to perform routing calculation according to the latest HELLO message and TC message received; Step S54: If yes, update the topology database according to the HELLO message and the TC message; otherwise, end the algorithm; Step S55, calculating the optimal path according to the updated topology database and neighbor table; Step S56: Update the routing table according to the calculated optimal path.

7. The inter-satellite routing method based on computing resources and link status according to claim 6, characterized in that: In step S55, the optimal path is calculated using a modified Djisktra algorithm, which specifically includes: Step S551, initialization: add all neighbor nodes in the topology database to the node list; add the current node to the node list; initialize the metrics of all nodes in the node list to infinity; set the metric of the current node to 0; initialize the visited node set to an empty set; Step S552: Select the node with the smallest metric in the node list as the current processing node, remove the current processing node from the node list and add it to the visited node set; Step S553: ​​traverse each neighbor node of the current processing node in the topology database. If the neighbor node is not added to the visited node set, execute: Calculate the new metric from the current processing node to the neighboring node, where the new metric is calculated as follows: New metric = metric of the current processing node + comprehensive metric of the topology database of neighboring nodes and the current processing node; The calculation formula of the comprehensive metric of the topology database is: Newmetric ij =w1·R j +w2·S ij +w3·Oldmetric Among them, i represents the current processing node, j represents the neighbor node of the current processing node i, and R j represents the computing resource degree of neighbor node j, S ij It represents the link stability between neighbor nodes j of the current processing node i. Oldmetric is the distance metric calculated in the traditional way. If the new metric is smaller than the metric of the neighboring node, update the metric of the neighboring node; Step S554, return to step S552, until the node list is not empty.

Images