A satellite network topology discovery method, path construction method and device

Abstract

The application discloses a satellite network topology discovery method, a path construction method and device, the method comprising: dividing the satellite network into autonomous domains and establishing a SID matrix and a satellite node adjacency matrix by assigning each satellite an extended SRv6 SID containing a network prefix, an orbital plane number, a node number, a service type and a resource margin; the satellite node perceives its own resource state, periodically detects the link state by using a BFD protocol, synchronizes the values of each element in the matrix by using an incremental update mechanism, and realizes dynamic topology perception; meanwhile, a routing path construction method based on the SID matrix and the adjacency matrix is proposed, which supports the satellite node to locally perform path calculation. The method supports heterogeneous service identification and resource state perception, can effectively reduce signaling overhead, improve the accuracy of satellite network topology discovery and routing efficiency, and improve the cross-constellation computing power pooling and collaborative computing capability.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of satellite networks, and in particular to a satellite network topology discovery method and a path construction method and device. BACKGROUND

[0002] SRv6 (Segment Routing over IPv6) is a key technology for current network architecture innovation, which simplifies the complex protocol stack of traditional networks into a programmable IPv6 (Internet Protocol Version 6) data plane. By embedding an SRH (Segment Routing Header) extension header in an IPv6 packet, SRv6 realizes flexible arrangement of a 128-bit segment identifier (Segment Identifier, hereinafter referred to as SID), wherein a Locator identifies a node position, a Function defines a forwarding behavior, and Args carries service parameters, forming an intelligent routing system of "position-function-parameter" trinity.

[0003] However, in the traditional SRv6, the traditional SID only supports ground network topology identification and cannot effectively describe the dynamic altitude and position of a satellite constellation and on-board heterogeneous computing resources (such as FPGA, CPU, GPU, etc.). At the same time, due to the lack of a standardized space computing identification system, different satellite manufacturers use private coding schemes, which seriously restricts the development of cross-constellation computing pool and collaborative computing.

[0004] On the other hand, in the traditional satellite network topology discovery protocol, network nodes mainly master the network topology information within the autonomous domain (Autonomous System, AS) through the autonomous domain level management mechanism. Network nodes lack global network topology structure and are heavily dependent on controllers for path construction.

[0005] Through the search of the prior art, it is found that Shandong Lancid Science Research Institute Co., Ltd. proposes in the invention patent "Computing power routing system and method based on SRv6" (application number CN202210458725.X) that an algorithm power routing controller determines the computing power service SRv6 SID and the forwarding action policy of the computing power service gateway based on the computing power service information and the computing power routing information. Beijing Century Interconnection Broadband Data Center Co., Ltd. proposes in the invention patent "Computing power information notification method, computing power network element node, equipment and storage medium" (application number CN202311014426.8) that the computing power state information is encapsulated by using an extended border gateway protocol BGP-link state LS, and the computing power state information includes a segment identifier SID of the computing power network element node. However, the above methods do not modify the traditional SRv6 SID structure and cannot identify heterogeneous computing power resources. SUMMARY

[0006] In view of the above problems in the prior art, the present application provides a satellite network topology discovery method and device, and a path construction method, which solves the problem that satellite nodes cannot perceive global on-board heterogeneous computing power resources and network topology in the prior art.

[0007] The object of the present application is achieved by the following technical solutions.

[0008] A satellite network topology discovery method comprises the following steps:

[0009] S1: a controller assigns an extended SRv6 SID to each satellite according to the format of the extended SRv6 SID, and each satellite generates an initial extended SRv6 SID; the extended SRv6 SID is composed of six fields of a network prefix, an orbital plane number, a node number, a service type, a resource margin and a must-be-zero;

[0010] S2: the satellite network is divided into different autonomous domains;

[0011] S3: each satellite obtains the initial extended SRv6 SID of other satellites in the orbital layer to which it belongs, and determines the number of satellite orbital planes in the orbital layer and the number of satellite nodes in each orbital plane;

[0012] S4: each satellite establishes a SID matrix of the orbital layer to which it belongs, and records the initial extended SRv6 SID of the satellite in the SID matrix;

[0013] S5: each satellite establishes a satellite node adjacency matrix of the orbital layer to which it belongs, which is used to represent the communication link connection relationship between satellites in the orbital layer;

[0014] S6: Each satellite perceives its own resource state, updates the field content of the service type and resource margin in its extended SRv6 SID, and updates the extended SRv6 SID of the satellite in the SID matrix; at the same time, each satellite periodically performs link state perception between adjacent satellites within the orbit layer to which the satellite belongs, and updates the value of the element in the satellite node adjacency matrix;

[0015] S7: Each satellite synchronizes the SID matrix and the satellite node adjacency matrix with adjacent satellites within the autonomous domain until the values of the elements in the SID matrix and the satellite node adjacency matrix of all satellites within the autonomous domain are the same;

[0016] S8: Within the same orbit layer, first, the boundary satellite nodes within each autonomous domain synchronize and update the SID matrix and the satellite node adjacency matrix across regions with the boundary satellite nodes of other autonomous domains, and then the boundary satellite nodes within each autonomous domain synchronize and update the SID matrix and the satellite node adjacency matrix of other autonomous domains to the satellite nodes within the autonomous domain, so that the values of the elements in the SID matrix and the satellite node adjacency matrix of all satellites within the orbit layer are consistent;

[0017] S9: The boundary satellite nodes of each orbit layer synchronize the SID matrix and the satellite node adjacency matrix within different orbit layers with the boundary satellite nodes of other orbit layers.

[0018] Further, the total length of the extended SRv6 SID is 128 bits;

[0019] The service type field in the step S6 is used to indicate that the satellite node suspends service, or to indicate that the satellite node provides a single service including forwarding, storage, and calculation, or a heterogeneous service composed of forwarding, storage, and calculation.

[0020] Further, in the step S4, the initial extended SRv6 SID of each satellite in the orbit layer to which the satellite belongs can be obtained in the following two ways:

[0021] Method one: the controller sends a message to each satellite to announce the initial extended SRv6 SID of other satellites in the orbit layer to which the satellite belongs;

[0022] Method two: each satellite discovers neighbors through an internal gateway protocol and an external gateway protocol within the orbit layer to which the satellite belongs, and exchanges the initial extended SRv6 SID.

[0023] Further, in the step S5, the method for each satellite to establish the satellite node adjacency matrix of the orbit layer to which the satellite belongs is as follows:

[0024] (1) According to the number m of columns and the number n of rows of the SID matrix, the number N ele of elements in the SID matrix is calculated;

[0025] (2) Establish a satellite node adjacency matrix with N ele columns and rows, each element in the satellite node adjacency matrix stores a tuple, namely (link, ttl), where link is 0, indicating a broken link, link is a positive real number, indicating link cost, and ttl indicates the predicted communication link connection duration; the initial value of each element tuple is set to (0, 0).

[0026] Further, in step S6, in order to reduce frequent network signaling interaction, only when the predicted service duration is greater than the preset time threshold or the resource residual amount changes greater than the preset capacity threshold, update the corresponding extended SRv6 SID in the SID matrix of the satellite.

[0027] Further, in step S6, the determination method of the element in the corresponding satellite node adjacency matrix is:

[0028] (1) First, number the elements in the SID matrix in ascending order of row number and column number;

[0029] (2) Convert the value of the node number field of the extended SRv6 SID to decimal to determine the row i of the extended SRv6 SID in the SID matrix; convert the value of the orbital plane number field to decimal to determine the column j of the extended SRv6 SID in the SID matrix;

[0030] (3) Finally, calculate the number a of the extended SRv6 SID in the SID matrix element according to (j-1)*n+i, where n is the total number of rows of the SID matrix.

[0031] Further, in step S6, updating the value of the element in the corresponding satellite node adjacency matrix specifically includes:

[0032] When two satellites are linked, if the ttl of the corresponding element tuple (link, ttl) in the current satellite node adjacency matrix is equal to 0, indicating that the link is in a failed state, the link value is modified to a positive real number, and the value of the positive real number represents the link cost, and ttl is set to the predicted communication link connection duration;

[0033] When two satellites are linked, if the ttl of the corresponding element tuple (link, ttl) in the current satellite node adjacency matrix is greater than 0, indicating that the link is in a valid state, i.e., the link connection state of the two satellites is in a state of oscillation and re-linking after breaking, then only the link value is modified to a positive real number, and the value of the positive real number represents the link cost;

[0034] Once the link between two satellites is established, if the link status is continuously monitored multiple times, the value of the corresponding element tuple (link, ttl) will not be changed.

[0035] Once the link between two satellites is established, if the link is broken in three consecutive link status sensing events, the value of link in the corresponding element tuple (link, ttl) will be modified to 0 to mark the link as broken.

[0036] Once the link between the two satellites is established, when the TTL countdown reaches 0, the value of link in the corresponding tuple (link, ttl) is modified to 0 to mark the link as broken, and ttl is set to 0 to indicate that the link is in a failed state.

[0037] A method for constructing routing paths in a satellite network, which is based on a network topology constructed using a satellite network topology discovery method, includes the following steps:

[0038] Step 1: The source satellite node queries the SID matrix for target satellite nodes that meet the service type requirements and remaining resource requirements through its own node or the controller;

[0039] Step 2: The source satellite node determines the source satellite node and target satellite node numbers in the SID matrix through its own node or the controller;

[0040] Step 3: The source satellite node calculates the routing path based on the source and target satellite node numbers in the SID matrix and the satellite node adjacency matrix through its own node or the controller;

[0041] Step 4: The source satellite node obtains the extended SRv6 SID of the corresponding node based on the node number in the routing path calculation result through this node or the controller;

[0042] Step 5: The source satellite node fills the extended SRv6 SID of each node into the SRH extension header of its own node in sequence through its own node or the controller.

[0043] Furthermore, in step four, the method for obtaining the extended SRv6 SID of the corresponding node is as follows:

[0044] (1) Calculate the row number of the corresponding element in the SID matrix according to (y-1)mod n+1, where y is the node number in the path calculation result, n is the total number of rows in the SID matrix, and mod is the modulo operation;

[0045] The column number of the element in the corresponding SID matrix is ​​calculated based on floor((y-1) / n)+1, where floor() is the floor function.

[0046] (2) According to the row number and the column number of the element in the SID matrix, an extended SRv6 SID of the node is obtained.

[0047] A satellite network topology discovery device, characterized by comprising a memory and one or more processors, the memory has stored executable code, and the one or more processors execute the executable code to implement a satellite network topology discovery method.

[0048] The beneficial effects of the present application are:

[0049] The present application improves the format of the SRv6 SID to become an extended SRv6 SID, keeps the total length of the SRv6 SID unchanged, and enables the satellite node to master the global dynamic satellite network topology structure and perceive the global on-board heterogeneous computing power resources. Meanwhile, the method of the present application enables the satellite node to construct a routing path locally without relying on the controller. The present application can also effectively reduce the signaling overhead and improve the accuracy of satellite network topology discovery and routing efficiency. BRIEF DESCRIPTION OF DRAWINGS

[0050] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0051] Figure 1 An extended SRv6 SID format diagram of the present application;

[0052] Figure 2 An Orbit Plane ID (Orbit Plane ID) field format diagram in the extended SRv6 SID format of the present application;

[0053] Figure 3 An Orbital Layers (Orbital Layers) field value and meaning diagram in the extended SRv6 SID format of the present application;

[0054] Figure 4 A ResourceAvailability (ResourceAvailability) field format diagram in the extended SRv6 SID format of the present application when the satellite node suspends service;

[0055] Figure 5 A ResourceAvailability (ResourceAvailability) field format diagram in the extended SRv6 SID format of the present application when the satellite node provides a service type;

[0056] Figure 6 ResourceAvailability (resource availability) field format diagram when two service types are provided for the satellite node in the SRv6 SID format extended by the present application;

[0057] Figure 7 ResourceAvailability (resource availability) field format diagram when three service types are provided for the satellite node in the SRv6 SID format extended by the present application;

[0058] Figure 8 Flow diagram of a satellite network topology discovery method provided by an embodiment of the present application;

[0059] Figure 9 Satellite orbit, satellite node number and SID matrix diagram;

[0060] Figure 10 Flow diagram of a satellite network routing path construction method provided by another embodiment of the present application.

[0061] Figure 11 SRH extension header format and SRH extension header abstract format diagram. DETAILED DESCRIPTION

[0062] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the scope of protection of the present application.

[0063] Explanation of technical terms:

[0064] MBZ, Must Be Zero, must be zero;

[0065] BFD, Bidirectional Forwarding Detection, bidirectional forwarding detection;

[0066] FPGA, Field Programmable Gate Array, field programmable gate array;

[0067] CPU, Central Processing Unit, central processing unit;

[0068] GPU, Graphics Processing Unit, graphics processing unit;

[0069] NPU, Neural Processing Unit, neural processing unit

[0070] PLD, Programmable Logic Device, programmable logic device

[0071] EGP, Exterior Gateway Protocol, exterior gateway protocol

[0072] IGP, Interior Gateway Protocol, interior gateway protocol

[0073] TB, Terabyte, terabyte

[0074] MB, Megabyte, megabyte

[0075] GB, Gigabyte, gigabyte

[0076] MHz, Megahertz, megahertz

[0077] GHz, Gigahertz, gigahertz

[0078] MFLOPS, Million Floating-Point Operations Per Second, million floating-point operations per second

[0079] GFLOPS, Giga Floating-point Operations Per Second, giga floating-point operations per second

[0080] TFLOPS, Tera Floating-Point Operations Per Second, tera floating-point operations per second

[0081] PFLOPS, Peta Floating-Point Operations Per Second, peta floating-point operations per second

[0082] MOPS, Million Operations Per Second, million operations per second

[0083] GOPS, Giga Operations Per Second, giga operations per second

[0084] TOPS, Tera Operations Per Second, tera operations per second

[0085] POPS, Peta Operations Per Second, one trillion operations per second;

[0086] SRH, Segment Routing Header, segment routing header;

[0087] TTL, Time to Live, time to live.

[0088] It should be noted that the features in the following embodiments and implementation manners can be combined with each other without conflict.

[0089] Figure 1 is an extended SRv6 SID format diagram. The traditional SRv6 SID format includes four fields of Locator (location identifier), Function (function instruction), Arguments (argument) and MBZ (must be zero), and it is stipulated that the total length of the four fields of Locator (location identifier), Function (function instruction), Arguments (argument) and MBZ (must be zero) is 128 bits.

[0090] The extended SRv6 SID divides the Locator (location identifier) field into three parts of Network Prefix (network prefix), Orbit Plane ID (orbit plane number) and Node ID (node number), defines Service Type (service type) in the Function (function instruction) field, and defines Resource Availability (resource availability) in the Arguments (argument) field. The MBZ (must be zero) field is used to pad with 0 for the lowest bits when the sum of the bit numbers of the other three fields is less than 128 bits, so that the total length of the extended SRv6 SID is 128 bits, consistent with the length of the traditional SRv6 SID.

[0091] As shown in Figure 1 , in this embodiment, the length of the Network Prefix (network prefix) field in the extended SRv6 SID format is 32 bits, the length of the Orbit Plane ID (orbit plane number) field is 12 bits, the length of the Node ID (node number) field is 8 bits, the length of the Service Type (service type) field is 8 bits, and the length of the Resource Availability (resource availability) field is 60 bits. The length of each field in the above extended SRv6 SID format is only an example in this embodiment, and the length of each field can be adjusted according to requirements.

[0092] Figure 2 is a format diagram of extending the Orbit Plane ID field in the SRv6 SID format. The Orbit Plane ID field is divided into Orbital Layers and Orbit Plane No. fields. Among them, the Orbital Layers field is 4 bits long, used to indicate the orbit height of the satellite. The Orbit Plane No. field is 8 bits long, used to indicate the orbit plane number in the same orbit layer.

[0093] Figure 3 is a diagram of the value and meaning of the Orbital Layers field in the SRv6 SID format of the application. As shown in Figure 3 0000 in the Orbital Layers field is used to indicate low orbit, 0001 is used to indicate medium orbit, and 0010 is used to indicate high orbit.

[0094] The Service Type field is used to represent the forwarding, storage, computing and other services provided by the satellite node, for example: when the satellite node provides a single service type, the Service Type field uses 0x00 to represent suspended service, 0x01 to represent forwarding service, 0x02 to represent storage service, 0x03 to represent CPU computing service, 0x04 to represent GPU computing service, 0x05 to represent NPU computing service, 0x06 to represent FPGA computing service; when the satellite node provides no less than two service types, the Service Type field uses 0x07 to represent CPU and GPU heterogeneous computing service, 0x08 to represent CPU and FPGA heterogeneous computing service, 0x09 to represent CPU and NPU heterogeneous computing service, 0x0A to represent storage, CPU and GPU heterogeneous computing service, 0x0B to represent storage, CPU and FPGA heterogeneous computing service, and 0x0C to represent storage, CPU and NPU heterogeneous computing service.

[0095] Figure 4is a format diagram of the Resource Availability (resource availability) field in the extended SRv6 SID format when the satellite node suspends service. When the Service Type (service type) field indicates that the satellite node suspends service, that is, the value of the Service Type (service type) field is 0x00, the Resource Availability (resource availability) field is divided into an Error Code (error code) and an MBZ (must be zero) field. The Error Code (error code) field is 8 bits long, used to indicate the reason for the satellite node suspending service, such as 0x00 for insufficient resources, 0x01 for CPU hardware failure, and the like. The MBZ (must be zero) field is padded with 0s to the lowest number of bits, so that the total length of the Resource Availability (resource availability) field is 60 bits.

[0096] When the Service Type (service type) field indicates that the satellite node provides one / multiple service types, the Resource Availability (resource availability) field format is [Service Type 1 (service type 1) | Unit 1 (unit 1) | Resource Availability 1 (resource availability 1) | Service Type 2 (service type 2) | Unit 2 (unit 2) | Resource Availability 2 (resource availability 2) | …… | MBZ (must be zero)], where the ellipsis indicates that there are more service types, which are added in the format of [Service Type n (service type n) | Unit n (unit n) | Resource Availability n (resource availability n)].

[0097] Figure 5is an example of the format of the Resource Availability field in the extended SRv6 SID format when the satellite node provides a single service type. When the Service Type field indicates that the satellite node provides a single service type, i.e., the value of the Service Type field is 0x01, 0x02, 0x03, 0x04, 0x05, or 0x06, the Resource Availability field is divided into the Service Type 1, Unit 1, Resource Availability 1, and MBZ fields. The Service Type 1 field has a length of 4 bits, the Unit 1 field has a length of 4 bits, and the Resource Availability 1 field has a length of 12 bits. The MBZ field is used to pad the lowest bits with 0 when the sum of the number of bits of the Unit 1 and Resource Availability 1 fields is less than 60 bits, so that the total length of the Resource Availability field is 60 bits.

[0098] The Service Type 1 field is used to indicate the forwarding, storage, computing, and other services provided by the satellite node, for example, 0x1 represents a forwarding service, 0x2 represents a storage service, 0x3 represents a CPU computing service, 0x4 represents a GPU computing service, 0x5 represents an NPU computing service, and 0x6 represents an FPGA computing service. The difference between the Service Type 1 field and the Service Type field is that the Service Type field has a length of 8 bits and can be used to represent multiple heterogeneous service types. The Service Type 1 field has a length of 4 bits and is only used to represent a single service type.

[0099] The Unit 1 field is used to represent the unit of different resources. When the Service Type 1 field indicates 0x1 forwarding service or 0x2 storage service, the Unit 1 field uses 0x0 to represent MB, 0x1 to represent GB, and 0x2 to represent TB. When the Service Type 1 field indicates 0x3 CPU computing service, the Unit 1 field uses 0x0 to represent MHz and 0x1 to represent GHz. When the Service Type 1 field indicates 0x4 GPU computing service or 0x6 FPGA computing service, the Unit 1 field uses 0x0 to represent MFLOPS, 0x1 to represent GFLOPS, 0x2 to represent TFLOPS, and 0x3 to represent PFLOPS. When the Service Type 1 field indicates 0x5 NPU computing service, the Unit 1 field uses 0x0 to represent MOPS, 0x1 to represent GOPS, 0x2 to represent TOPS, and 0x3 to represent POPS.

[0100] Figure 6 is an example of the format of the Resource Availability field in the extended SRv6 SID format when the satellite node provides two service types. When the Service Type field indicates that the satellite node provides two service types, that is, the value of the Service Type field is 0x07, 0x08, or 0x09, the Resource Availability field is divided into the Service Type 1, Unit 1, Resource Availability 1, Service Type 2, Unit 2, Resource Availability 2, and MBZ fields. The Service Type 1 and Service Type 2 fields each have a length of 4 bits, the Unit 1 and Unit 2 fields each have a length of 4 bits, and the Resource Availability 1 and Resource Availability 2 fields each have a length of 12 bits. The MBZ field is used to pad the lowest bits with 0 to make the total length of the Resource Availability field 60 bits when the sum of the number of bits of the above-mentioned fields is less than 60 bits.

[0101] When the Service Type field indicates that the satellite node provides two service types, the Service Type 1 and Service Type 2 fields are used to indicate the first and second service types, respectively, the Unit 1 and Unit 2 fields are used to indicate the units of the first and second service types, respectively, and the Resource Availability 1 and Resource Availability 2 fields are used to indicate the resource availability of the first and second service types, respectively. The values and meanings of the Service Type 1 and Service Type 2 fields are the same as those of the Service Type 1 field when the satellite node provides one service type. Figure 5 The values and meanings of the Unit 1 and Unit 2 fields are the same as those of the Unit 1 field when the satellite node provides one service type. Figure 5 The values and meanings of the Unit 1 and Unit 2 fields are the same as those of the Unit 1 field when the satellite node provides one service type. For example, when the Service Type field is 0x07 indicating a CPU and GPU heterogeneous computing service, the Service Type 1 field takes a value of 0x3 indicating a CPU computing service, and the Service Type 2 field takes a value of 0x4 indicating a GPU computing service.

[0102] Figure 7is an example of the format of the Resource Availability field in the extended SRv6 SID format when the satellite node provides three service types. When the Service Type field indicates that the satellite node provides three service types, i.e., the value of the Service Type field is 0x0A, 0x0B or 0x0C, the Resource Availability field is divided into Service Type 1, Unit 1, Resource Availability 1, Service Type 2, Unit 2, Resource Availability 2, Service Type 3, Unit 3 and Resource Availability 3. The lengths of the Service Type 1, Service Type 2 and Service Type 3 fields are all 4 bits, the lengths of the Unit 1, Unit 2 and Unit 3 fields are all 4 bits, and the lengths of the Resource Availability 1, Resource Availability 2 and Resource Availability 3 fields are all 12 bits.

[0103] When the Service Type field indicates that the satellite node provides three service types, the Service Type 1, Service Type 2 and Service Type 3 fields are respectively used to indicate the first, second and third service types, the Unit 1, Unit 2 and Unit 3 fields are respectively used to indicate the units of the first, second and third service types, and the Resource Availability 1, Resource Availability 2 and Resource Availability 3 fields are respectively used to indicate the resource availabilities of the first, second and third service types. The values and meanings of the Service Type 1, Service Type 2 and Service Type 3 fields are the same as those of the Service Type 1, Service Type 2 and Service Type 3 fields in the extended SRv6 SID formatFigure 5 The value and meaning of the Service Type 1 field when the middle satellite node provides a service type are the same as those of the Service Type field. The values and meanings of the Unit 1, Unit 2, and Unit 3 fields are the same as those of the Unit 1, Unit 2, and Unit 3 fields in the SRv6 SID format. Figure 5 The value and meaning of the Unit 1 field when the middle satellite node provides a service type are the same as those of the Service Type field. For example, when the Service Type field is 0x0A, indicating a storage, CPU, and GPU heterogeneous computing service, the Service Type 1 value is 0x2, indicating a storage service, the Service Type 2 value is 0x3, indicating a CPU computing service, and the Service Type 3 value is 0x4, indicating a GPU computing service.

[0104] Figure 8 A flowchart of a satellite network topology discovery method provided by an embodiment of the present application is shown in FIG. 1. The flow in FIG. 1 needs to be re-run whenever the number of satellite nodes in the satellite orbital layer changes. For example, when a batch of satellites are released or injected into orbit, a plurality of satellites can be deployed in the same orbit in batches. At this time, the overall satellite network basic topology changes, and the flow in FIG. 1 needs to be re-run to discover the satellite network topology. Figure 8 Figure 8

[0105] As shown in FIG. 2, a satellite network topology discovery method according to an embodiment of the present application includes the following steps: Figure 8

[0106] S101, a controller assigns an extended SRv6 SID to each satellite according to the format of the extended SRv6 SID, and each satellite generates an initial extended SRv6 SID; the extended SRv6 SID is composed of six fields, including a network prefix, an orbital plane number, a node number, a service type, a resource margin, and a must-be-zero field.

[0107] The controller pre-assigns a Network Prefix, an Orbit Plane ID (including Orbital Layers and Orbit Plane No.), a Node ID, and a Service Type field to each satellite. Each satellite has a unique number composed of an Orbit Plane No. and a Node ID.

[0108] ​​​For example, the satellite node 5 running in low orbit with orbital plane number 4 is assigned a Network Prefix field of 2001:db8:: / 32, an Orbital Layers field of 0000 indicating low orbit, an Orbit Plane No. field of 0x04, and a Node ID field of 0x05. The satellite node provides CPU and GPU heterogeneous computing services, so the Service Type field is configured as 0x07.

[0109] Each satellite generates an initial extended SRv6 SID according to the fields assigned by the controller. The Resource Availability field is filled with 0.

[0110] For example, the initial extended SRv6 SID of the satellite node 5 running in low orbit with orbital plane number 4 in S101 is 2001:db8:4:507::.

[0111] S102, divide the satellite network into different autonomous domains.

[0112] For example, the satellite network within the same orbital plane number in the same orbital layer can be divided into an autonomous domain according to the Orbital Layers field and the Orbit Plane No. field of the initial extended SRv6 SID. EGP protocol is run between autonomous domains, and IGP protocol is run within an autonomous domain.

[0113] S103, each satellite obtains the initial extended SRv6 SID of other satellites in its own orbital layer, and determines the number of satellite orbital planes in its own orbital layer and the number of satellite nodes in each orbital plane.

[0114] Each satellite obtaining the initial extended SRv6 SID of other satellites in its own orbital layer includes the following two methods:

[0115] Method 1: the controller (such as a ground control center, a high-orbit satellite, etc.) sends a message to each satellite to notify it of the initial extended SRv6 SID of other satellites in its own orbital layer;

[0116] Method 2: each satellite discovers neighbors and exchanges initial extended SRv6 SIDs through IGP and EGP protocols within its own orbital layer.

[0117] Each satellite determines the number of satellite orbital planes in its own orbital layer and the number of satellite nodes in each orbital plane according to the initial extended SRv6 SID of all satellites in its own orbital layer.

[0118] Each satellite checks the Orbital Layers field in the initial extended SRv6 SID of all satellites, and selects the initial extended SRv6 SID of other satellites in the same orbital layer as itself. According to the Orbit Plane No. field and the Node ID field in the initial extended SRv6 SID of all satellites in the same orbital layer, the number of satellite orbit planes m in the orbital layer to which the satellite belongs and the number of satellite nodes n(i) in each orbit plane (i = 1, 2, …, m) are determined.

[0119] S104, each satellite establishes a SID matrix of the orbital layer to which the satellite belongs, and records the initial extended SRv6 SID of the satellite in the SID matrix.

[0120] Let m represent the number of satellite orbit planes in the orbital layer to which the satellite belongs, and n(i) represent the number of satellite nodes in orbit plane i (i = 1, 2, …, m) in the orbital layer to which the satellite belongs. Then each satellite takes m as the number of columns of the SID matrix, takes the maximum value n of n(i), i.e., n = max n(i), (i = 1, 2, …, m), as the number of rows of the SID matrix, and establishes a SID matrix locally, where the initial value of all elements in the SID matrix is 0.

[0121] At the same time, each satellite sorts the Orbit Plane No. field in the initial extended SRv6 SID of all satellites in the orbital layer to which the satellite belongs in ascending order to correspond to the columns in the SID matrix, sorts the Node ID field in ascending order to correspond to the rows in the SID matrix, and records the initial extended SRv6 SID in the corresponding SID matrix element.

[0122] For example, assume that the number of satellite orbit planes in the orbital layer to which the satellite belongs is 6, and that the 6 Orbit Plane No. fields sorted in ascending order are 0x01, 0x02, 0x03, 0x04, 0x05, and 0x06. At the same time, there are 5 satellites in the first 5 satellite orbit planes, and the 5 Node ID fields sorted in ascending order are 0x01, 0x02, 0x03, 0x04, and 0x05. There are 4 satellites in the sixth satellite orbit plane, and the 4 Node ID fields sorted in ascending order are 0x01, 0x02, 0x03, and 0x04. The SID matrix is established as shown in Figure 9 .

[0123] Figure 9 is a schematic diagram of a satellite orbit, a satellite node number, and a SID matrix. In Figure 9In the SID matrix, each element records the initial extended SRv6 SID value of the corresponding satellite. For example... Figure 9 In the SID matrix, SID5,4 represents the element value in the 5th row and 4th column. Here, 4 indicates the Orbit Plane No. (orbit plane number) field, which is the 4th position in ascending order within the satellite's orbital layer (0x04). 5 indicates the Node ID (orbit number) field, which is the 5th position in ascending order within the satellite's orbital layer (0x05). Based on the initial extended SRv6 SID generated in S101, the initial extended SRv6 SID of satellite node 5, operating in low Earth orbit with orbital plane number 4, is 2001:db8:4:507::. Therefore, the value of SID5,4 is 2001:db8:4:507::. Figure 9 The element value of 0 in the 5th row and 6th column of the SID matrix indicates that there is no satellite in the 6th position of the OrbitPlane No. (orbit plane number) that is sorted in ascending order and has the 5th position of the Node ID (node ​​number).

[0124] S105. Each satellite establishes a satellite node adjacency matrix for its orbital layer, which is used to represent the communication link connection relationship between satellites within its orbital layer.

[0125] Given the number of columns *m* and rows *n* of the SID matrix, calculate the number of elements in the SID matrix: N_ele = m * n. Construct a satellite node adjacency matrix with N_ele columns and rows. Each element in this adjacency matrix stores a tuple, (link, ttl), where link = 0 indicates a broken link, a positive real number represents the link cost, and ttl represents the predicted duration of the communication link connection. The initial value of each element tuple is set to (0, 0).

[0126] S106. Each satellite senses its own resource status, updates the contents of the Service Type and Resource Availability fields in its extended SRv6 SID, and updates its extended SRv6 SID in the SID matrix. At the same time, each satellite periodically senses the link status with neighboring satellites within its orbital layer and updates the values ​​of the elements in the corresponding satellite node adjacency matrix.

[0127] When each satellite perceives its own resource state, the satellite node sets the value of the Service Type field in the extended SRv6 SID to 0x00 if it suspends service, and indicates the reason for suspending service in the Error Code field of the Resource Availability field. When the satellite node resumes service, the value of the Service Type field in the extended SRv6 SID is set to the initial value. In addition, when the satellite node provides normal service, according to the value of the Service Type field in the extended SRv6 SID, the format of the Resource Availability field is determined as shown in Figure 5 、 Figure 6 、 Figure 7 and the values in the corresponding Resource Availability field format in the extended SRv6 SID are modified according to the results of the perception of its own resource state.

[0128] In addition, each satellite performs link state perception within the orbit layer to which it belongs according to the configured BFD detection time, and periodically performs link state perception with adjacent satellites through the BFD protocol. In order to reduce frequent network signaling interaction, only when the predicted service duration is greater than a preset time threshold (such as a task lasting 10 minutes), or the resource availability changes greater than a preset capacity threshold, the corresponding extended SRv6 SID in the SID matrix of the satellite is updated.

[0129] Taking two adjacent satellites in the same orbit layer (assuming that the SIDs of the two satellites are extended SRv6 SID 1 and extended SRv6 SID 2) as an example, the determination method of the elements in the satellite node adjacency matrix is as follows:

[0130] (1) The elements in the SID matrix are numbered in ascending order of row number and ascending order of column number.

[0131] As shown in Figure 9 , the elements in the established SID matrix are numbered from ele1 to ele30 in ascending order of row number and ascending order of column number.

[0132] (2) The value of the Node ID field in the extended SRv6 SID (i.e., SRv6 SID 1) is converted to decimal to determine the row i of the extended SRv6 SID (i.e., SRv6 SID 1) in the SID matrix; the value of the Orbit Plane No. field is converted to decimal to determine the column j of the extended SRv6 SID (i.e., SRv6 SID 1) in the SID matrix.

[0133] (3) Calculate the number a of the extended SRv6 SID 1 in the SID matrix element according to (j-1)*n+i, where n is the total number of rows of the SID matrix.

[0134] Similarly, the number b of the extended SRv6 SID 2 in the SID matrix element is calculated according to the same steps.

[0135] Updating the value of the element in the corresponding satellite node adjacency matrix specifically includes:

[0136] When the two satellites are chained, if the ttl of the corresponding element tuple (link, ttl) in the current satellite node adjacency matrix is equal to 0, indicating that the link is in a failure state, the link value is modified to a positive real number, and the value of the positive real number represents the link cost, and the ttl is set to the predicted communication link connection duration;

[0137] When the two satellites are chained, if the ttl of the corresponding element tuple (link, ttl) in the current satellite node adjacency matrix is greater than 0, indicating that the link is in an effective state, i.e., the link connection state of the two satellites is in a state of oscillation, disconnection and reconnection within the predicted communication link connection duration, only the link value is modified to a positive real number, and the value of the positive real number represents the link cost;

[0138] When the two satellites are chained, if the link state perception is continuously connected for multiple times, the value of the corresponding element tuple (link, ttl) is not changed;

[0139] When the two satellites are chained, if the link state perception is continuously disconnected for three times, only the value of the link in the corresponding element tuple (link, ttl) is modified to 0, marking the link disconnection;

[0140] When the two satellites are chained, and the TTL countdown is 0, the value of the link in the corresponding element tuple (link, ttl) is modified to 0, marking the link disconnection, and the ttl is set to 0, indicating that the link is in a failure state.

[0141] To prevent the frequent on-off of inter-satellite links from causing the values of the satellite node adjacency matrix to change frequently, each satellite configures a dynamic BFD detection time according to the Orbital Layers field in the extended SRv6 SID of the satellite itself. For example, when the Orbital Layers field is 0000, it indicates a low-orbit satellite, and the BFD detection time is configured to be 100 milliseconds; when the Orbital Layers field is 0001, it indicates a medium-orbit satellite, and the BFD detection time is configured to be 250 milliseconds; and when the Orbital Layers field is 0010, it indicates a high-orbit satellite, and the BFD detection time is configured to be 1000 milliseconds. Meanwhile, for the case of continuous triggering of link changes within a period of time, the BFD detection time is extended by a multiple, and is restored after the link situation stabilizes.

[0142] S107, each satellite synchronizes the SID matrix and the satellite node adjacency matrix with the adjacent satellites in the autonomous domain, until the values of the elements in the SID matrix and the satellite node adjacency matrix of all satellites in the autonomous domain are the same.

[0143] To reduce communication overhead, an incremental update mechanism is adopted. Each satellite only sends the value of the element tuple in the updated satellite node adjacency matrix and the matrix row and column where the element tuple is located to the adjacent satellites in the same autonomous domain, for example, (5, 4, 0, 0) indicates that the value of the 5th row and the 4th column of the satellite node adjacency matrix is updated to the tuple (0, 0), instead of sending the entire updated satellite node adjacency matrix to the adjacent satellites. Each node that receives the value update of the elements in the SID matrix and the satellite node adjacency matrix continues to forward the information to other neighbors until the values of the elements in the SID matrix and the satellite node adjacency matrix of all satellites in the autonomous domain are synchronized.

[0144] S108, in the same orbital layer, the boundary satellite nodes in each autonomous domain synchronize and update the SID matrix and the satellite node adjacency matrix with the boundary satellite nodes of other autonomous domains across regions (through the EGP protocol), and then the boundary satellite nodes in each autonomous domain synchronize and update the SID matrix and the satellite node adjacency matrix of other autonomous domains to the satellite nodes in the autonomous domain, so that the values of the elements in the SID matrix and the satellite node adjacency matrix of all satellites in the orbital layer are consistent.

[0145] S109, the boundary satellite nodes of each orbital layer synchronize the SID matrix and the satellite node adjacency matrix in different orbital layers with the boundary satellite nodes of other orbital layers.

[0146] Figure 10This is a flowchart illustrating a satellite network routing path construction method according to another embodiment of the present invention. This embodiment provides a method for constructing satellite network routing paths within the same orbital layer, but it can also be used for constructing satellite network routing paths within different orbital layers.

[0147] like Figure 10 As shown in the figure, a satellite network routing path construction method in an embodiment of the present invention includes the following steps:

[0148] S201. The source satellite node queries the SID matrix for target satellite nodes that meet the service type requirements and remaining resource requirements through this node or the controller.

[0149] For example Figure 9 Satellite node 2 in orbital plane 6 requires heterogeneous computing services from CPU and GPU. This node determines the target satellite node that meets the requirements by querying the Service Type and Resource Availability fields in each extended SRv6 SID in the SID matrix through its own node or the controller. That is, satellite node 5 in orbital plane 4.

[0150] S202. The source satellite node determines the source satellite node and target satellite node numbers in the SID matrix through this node or the controller.

[0151] For example, the source satellite node is Figure 9 Satellite node 2 in orbital plane 6 is identified by its SID matrix number as (6-1)*5+2=27 according to the method described in S106. The target satellite node is... Figure 9 Satellite node 5 in orbital plane 4 of the satellite is numbered (4-1)*5+5=20 in the SID matrix according to the method described in S106.

[0152] S203. The source satellite node calculates the routing path through this node or the controller based on the source satellite node and the target satellite node's numbers in the SID matrix and the satellite node adjacency matrix.

[0153] One method for calculating routes is to minimize the overall link cost while ensuring that the time-to-live (TTL) of all links in the route path is within its validity period. When multiple target satellite nodes can meet the requirements of the source satellite node, the minimum overall link cost from the source satellite node to each target satellite node is calculated separately, and the route path with the minimum overall link cost is selected.

[0154] For example, Figure 9 The intermediate source satellite node is satellite node numbered 27 in the SID matrix, and the target satellite node is satellite node numbered 20 in the SID matrix. For example...Figure 9 As shown, assuming the routing path calculation result is ele27→ele22→ele17→ele18→ele19→ele20.

[0155] S204, the source satellite node obtains the corresponding node extended SRv6 SID according to the node number in the routing path calculation result through the node or the controller.

[0156] Assuming the node number in the routing path calculation result is y, and n is the total number of rows of the SID matrix, then the row number of the element in the corresponding SID matrix is calculated according to (y-1) mod n+1, wherein mod is the modulus operation. The column number of the element in the corresponding SID matrix is calculated according to floor((y-1) / n)+1, wherein floor() is the floor operation. Then, the extended SRv6 SID of the node is obtained according to the row number and the column number of the element in the SID matrix.

[0157] For example, the target satellite node is numbered 20, i.e. Figure 9 the satellite node numbered ele20 in the above table, the row of the element in the corresponding SID matrix is (20-1) mod 5+1=5, and the column of the element in the corresponding SID matrix is floor((20-1) / 5)+1=4. The value of the element in the 5th row and the 4th column of the SID matrix is queried to obtain the extended SRv6 SID of the node, which is 2001:db8:4:507::.

[0158] S205, the source satellite node fills each node extended SRv6 SID into the SRH extension header of the node in order through the node or the controller.

[0159] Figure 11 The SRH extension header format and the SRH extension header abstract format are shown in the following table. According to the routing path calculation result ele27→ele22→ele17→ele18→ele19→ele20, the path passes through 6 nodes. Therefore Figure 11The value of n is 4 (counting from 0 and excluding the source node ele27), and the Segment List includes SID[0], SID[1], SID[2], SID[3], and SID[4]. Among them, SID[4] is the SID of ele22, SID[3] is the SID of ele17, SID[2] is the SID of ele18, SID[1] is the SID of ele19, and SID[0] is the SID of ele20. For example, the extended SRv6 SID of the satellite node numbered ele20 obtained according to S204 is 2001:db8:4:507::, and 2001:db8:4:507:: is filled into SID[0].

[0160] The application also provides a satellite network topology discovery device corresponding to Figure 8 At the hardware level, the satellite network topology discovery device includes a processor, an internal bus, a network interface, a memory, and a non-volatile memory, and can also include other hardware required by a business. The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs to implement the above-mentioned Figure 8 inter-satellite computing satellite network topology discovery method. Of course, in addition to the software implementation, the application does not exclude other implementation manners, such as a logic device or a combination of software and hardware, and the like, that is, the execution subject of the following processing flow is not limited to each logic unit, and can also be hardware or a logic device.

[0161] The above only describes the embodiments of the application and is not used to limit the application. The application can have various changes and modifications for those skilled in the art. Any modification, equivalent replacement, improvement, and the like within the spirit and principle of the application shall be included in the scope of claims of the application.

Claims

1. A method for discovering satellite network topology, characterized in that, Includes the following steps: S1: The controller assigns an extended SRv6 SID to each satellite according to the format of the extended SRv6 SID, and each satellite generates an initial extended SRv6 SID; the extended SRv6 SID consists of six fields: network prefix, orbital plane number, node number, service type, resource margin, and must be set to zero. S2: Divide the satellite network into different autonomous domains; S3: Each satellite obtains the initial extended SRv6 SID of other satellites in its orbital layer and determines the number of satellite orbital planes and the number of satellite nodes in each orbital plane in its orbital layer; S4: Establish a SID matrix for each satellite's orbital layer, and record the satellite's initial extended SRv6 SID in the SID matrix; S5: Each satellite establishes a satellite node adjacency matrix for its orbital layer, which is used to represent the communication link connection relationship between satellites within its orbital layer; S6: Each satellite senses its own resource status, updates the service type and resource balance fields in its extended SRv6 SID, and updates its extended SRv6 SID in the SID matrix; at the same time, each satellite periodically senses the link status with neighboring satellites within its orbital layer and updates the values ​​of the elements in the corresponding satellite node adjacency matrix. S7: Each satellite synchronizes its SID matrix and satellite node adjacency matrix with neighboring satellites within its autonomous system until the values ​​of all elements in the SID matrix and satellite node adjacency matrix of all satellites within the autonomous system are the same. S8: Within the same orbital layer, the boundary satellite nodes within each autonomous region first synchronize and update the SID matrix and satellite node adjacency matrix with the boundary satellite nodes of other autonomous regions across regions. Then, the boundary satellite nodes within each autonomous region synchronize and update the SID matrix and satellite node adjacency matrix of other autonomous regions with the satellite nodes within their own autonomous region, so that the values ​​of each element in the SID matrix and satellite node adjacency matrix of all satellites within the orbital layer are consistent. S9: The boundary satellite nodes of each orbital layer synchronize the SID matrix and satellite node adjacency matrix within different orbital layers with the boundary satellite nodes of other orbital layers.

2. The satellite network topology discovery method according to claim 1, characterized in that, The total length of the extended SRv6 SID is 128 bits; The service type field in step S6 is used to indicate whether the satellite node is suspending its service, or to indicate whether the satellite node provides a single service including forwarding, storage, and computing, or a heterogeneous service consisting of forwarding, storage, and computing.

3. The satellite network topology discovery method according to claim 1, characterized in that, In step S4, each satellite obtains the initial extended SRv6 SID of other satellites in its orbital layer in the following two ways: Method 1: The controller sends a message to each satellite to announce the initial extended SRv6 SID of other satellites in its orbital layer; Method 2: Each satellite discovers its neighbors within its orbital layer using internal gateway protocols and external gateway protocols, and exchanges initial extended SRv6 SIDs.

4. The satellite network topology discovery method according to claim 1, characterized in that, In step S5, the method for establishing the satellite node adjacency matrix for each satellite in its orbital layer is as follows: (1) Calculate the number of elements in the SID matrix N_ele=m*n based on the number of columns m and rows n of the SID matrix; (2) Establish a satellite node adjacency matrix with N_ele columns and rows. Each element in the satellite node adjacency matrix stores a tuple, namely (link, ttl), where link is 0 to indicate a broken link, link is a positive real number to indicate link cost, and ttl indicates the predicted duration of the communication link connection. The initial value of each element tuple is set to (0,0).

5. The satellite network topology discovery method according to claim 1, characterized in that, In step S6, in order to reduce frequent network signaling interactions, the corresponding extended SRv6 SID is updated in the SID matrix of the satellite only when the predicted service duration is greater than a preset time threshold or the resource reserve change is greater than a preset capacity threshold.

6. The satellite network topology discovery method according to claim 1, characterized in that, In step S6, the method for determining the elements in the corresponding satellite node adjacency matrix is ​​as follows: (1) First, number the elements in the SID matrix in ascending order of row number and then column number; (2) Convert the value of the node number field of the extended SRv6 SID to decimal and determine the row i of the extended SRv6 SID in the SID matrix; convert the value of the track surface number field to decimal and determine the column j of the extended SRv6 SID in the SID matrix; (3) Finally, calculate the number a of the extended SRv6 SID in the SID matrix element according to (j-1)*n+i, where n is the total number of rows in the SID matrix.

7. The satellite network topology discovery method according to claim 1, characterized in that, In step S6, updating the values ​​of the elements in the corresponding satellite node adjacency matrix specifically includes: When two satellites establish a link, if the ttl of the corresponding element tuple (link, ttl) in the current satellite node's adjacency matrix is ​​equal to 0, it indicates that the link is in a failed state. In this case, the link value is modified to a positive real number, where the value of the positive real number represents the link cost, and ttl is set to the predicted duration of the communication link connection. When two satellites establish a link, if the ttl of the corresponding element tuple (link, ttl) in the current satellite node adjacency matrix is ​​greater than 0, it indicates that the link is in a valid state. That is, the link connection state of the two satellites oscillates within the predicted duration of the communication link connection. If the link is broken and then re-established, the link value is only modified to a positive real number. The value of the positive real number represents the link cost. Once the link between two satellites is established, if the link status is continuously monitored multiple times, the value of the corresponding element tuple (link, ttl) will not be changed. Once the link between two satellites is established, if the link is broken in three consecutive link status sensing events, the value of link in the corresponding element tuple (link, ttl) will be modified to 0 to mark the link as broken. Once the link between the two satellites is established, when the TTL countdown reaches 0, the value of link in the corresponding tuple (link, ttl) is modified to 0 to mark the link as broken, and ttl is set to 0 to indicate that the link is in a failed state.

8. A method for constructing satellite network routing paths, characterized in that, This method is based on the network topology constructed by the satellite network topology discovery method described in claim 1, and includes the following steps: Step 1: The source satellite node queries the SID matrix for target satellite nodes that meet the service type requirements and remaining resource requirements through its own node or the controller; Step 2: The source satellite node determines the source satellite node and target satellite node numbers in the SID matrix through its own node or the controller; Step 3: The source satellite node calculates the routing path based on the source and target satellite node numbers in the SID matrix and the satellite node adjacency matrix through its own node or the controller; Step 4: The source satellite node obtains the extended SRv6 SID of the corresponding node based on the node number in the routing path calculation result through this node or the controller; Step 5: The source satellite node fills the extended SRv6 SID of each node into the SRH extension header of its own node in sequence through its own node or the controller.

9. The satellite network routing path construction method according to claim 8, characterized in that, In step four, the method for obtaining the extended SRv6 SID of the corresponding node is as follows: (1) Calculate the row number of the corresponding element in the SID matrix according to (y-1)mod n+1, where y is the node number in the path calculation result, n is the total number of rows in the SID matrix, and mod is the modulo operation; The column number of the element in the corresponding SID matrix is ​​calculated based on floor((y-1) / n)+1, where floor() is the floor function. (2) Obtain the extended SRv6 SID of the node based on the row and column number of the element in the SID matrix.

10. A satellite network topology discovery device, characterized in that, The system includes a memory and one or more processors, wherein the memory stores executable code, and the one or more processors execute the executable code to implement the satellite network topology discovery method according to any one of claims 1-7.

Images