Lightweight in-band telemetry method for space-ground integrated computing network

Abstract

The present invention discloses a lightweight in-band telemetry method for an integrated space-ground computing network, including integrating user needs at the application layer; calculating the total amount of node detection data at the controller layer; solving the shortest path between node pairs in the network topology; solving the optimal logical traversal path of the required node list; obtaining the single best detection path through a greedy cutting algorithm; sending the detection path flow table item to the path source node; judging the data packet type, the detection data packet continues to be processed, otherwise it is forwarded according to the routing table; if it is a source node, a segment routing extension header is added; the default bitmap is saved in an on-demand bitmap register; judging the current node type, the endpoint node continues to be processed, otherwise it is forwarded according to the routing table; updating the non-zero customized bitmap to the on-demand bitmap register; dynamically adding metadata items according to the on-demand bitmap; if the node is a sink node, the summary information is reported to the controller. The present invention has the advantages of low cost and high flexibility of network measurement.

Description

Technical Field

[0001] The present invention relates to the field of network measurement information, and in particular to a lightweight in-band telemetry method for an air-space-ground integrated computing network. Background Art

[0002] Measurement solutions based on traditional network architectures are limited by network devices. When deploying a network, the measurement capabilities of the network need to be designed and supported. Such solutions are highly customized, resulting in poor protocol scalability and deployment flexibility. In response to the limitations of fixed mechanisms for flexible network measurement, researchers have begun to resort to emerging software-defined networking (SDN) and programmable protocol-independent packet processors (P4) and corresponding forwarding models that implement programmable data planes, giving the data plane higher flexibility and programmability. In-band network telemetry (INT) is a typical application of p4. As a new hybrid network measurement technology, it allows network status information to be added to data packets in a piggyback manner by using INT headers and additional INT metadata, and accompanies data packets as they traverse the network, enabling the data plane to passively collect and report real-time fine-grained packet network information without the intervention of the network control plane.

[0003] However, traditional in-band telemetry has some limitations as follows: (1) poor scalability; (2) poor network compatibility; (3) network measurement requirements cannot be selected on demand; (4) lack of reasonable data packet navigation strategy. Segment Routing (SR) technology is a routing technology based on source routing. Nodes use the ordered list in the Segment Routing Header (SRH) to guide the forwarding of data packets, which can well meet the path planning requirements. SRv6 is an outstanding representative of SR technology. SRv6 combines SR and IPv6 by adding an extension header called Segment Routing Extension Header (Segment Routing Header) to IPv6 data packets. It implements network programming capabilities through these flexible IPv6 extension headers and uses a segment routing node identifier (Segment ID, SID) that is similar to an IPv6 address to identify a node. Some existing studies focus on the packet processing logic of in-band telemetry but ignore the packet forwarding design, resulting in the lack of scalability in the telemetry solution. Another part uses SR technology to guide packet forwarding but does not consider the various overheads brought about by the process. Therefore, the existing solutions cannot cope with the dynamic and large-scale characteristics of the network in large-scale air-ground integrated networks, thereby achieving efficient network measurement. Summary of the invention

[0004] To solve the above problems, the present invention discloses a lightweight in-band telemetry method for an integrated air-ground-space computing network. On-demand nodes are measured through SRv6. At the same time, for each required node, the programmability of the software switch is used to use the default bitmap in the detection packet header and the ARG field of the SID corresponding to the node to realize on-demand measurement of metadata items. Combined with low-overhead detection path planning, lightweight and customized in-band telemetry is achieved.

[0005] Specific solutions include:

[0006] The lightweight in-band telemetry method for the integrated space-ground computing network includes the following steps:

[0007] Step 1: The application layer integrates user requirements, stores them in the form of key-value pairs, and transmits them to the controller layer;

[0008] Step 2: The controller layer calculates the amount of data required to be measured for each node in the on-demand measurement list. The amount of data refers to the sum of the lengths of all metadata required to be measured for the node, and stores it in the form of a key-value pair together with the node number. Specifically, the node number is used as the key and the amount of data is used as the value. It is stored in the database for subsequent path cutting. The specific calculation formula for the total amount of detection data for a single node is:

[0009]

[0010] in is the detection data length of node i, is the length of resource j, is the value of the jth bit in the bitmap. A value of 1 indicates that the resource is needed, and a value of 0 indicates that the resource is not needed.

[0011] Step 3: construct a logical auxiliary graph for calculating the optimal detection path. Specifically, first access the topology information database to obtain the topology and link parameter information in the network, mainly the connection relationship of the link and the link cost. The specific calculation formula of the link cost is:

[0012]

[0013] in is the link cost of link m, and are the proportions of delay and bandwidth in link cost respectively, and are the delay and bandwidth of link m, and are the maximum and minimum delays in the topology database, respectively. and They are the maximum and minimum bandwidths in the topology database, taking into account the link delay and bandwidth. Then, the Dijistra algorithm is used to solve the shortest path between two nodes based on the link cost. The shortest path is used as the logical link of the logical auxiliary graph. The two ends of each logical link are the starting node and the ending node of the physical path, and the logical link cost is calculated. The logical link cost is the sum of the costs of each physical link on the physical path. The specific calculation formula is:

[0014]

[0015] in represents the logical link cost from node i to node j, For each physical direct path in the shortest path between nodes ij obtained in step 3, represents the nth physical link, Represents the link cost of the nth physical link. Construct a logical complete graph composed of logical links as a logical auxiliary graph to assist path planning calculation;

[0016] Step 4, extract the subgraph consisting of the required node list from the logic auxiliary graph, and use the TSP solving algorithm to solve the optimal logic traversal path;

[0017] Step 5: Based on the node measurement demand data volume, the greedy cutting algorithm is used to cut the logical traversal path. The cutting needs to meet the following two conditions: the total amount of data added by the data packet cannot exceed the network maximum transmission unit limit, and secondly, the amount of data added is as much as possible to obtain the best detection path that meets the network maximum transmission unit limit. The mathematical constraint formula for greedy cutting is:

[0018]

[0019] Where num represents the number of nodes in a single path. Representation Node The detection data length is , MTU is the maximum transmission unit value of the network, and max(num) represents the maximum number of detection nodes for a single path.

[0020] Step 6: construct the obtained multiple detection paths into SRv6 flow table entries respectively and send them to the source nodes of the respective paths;

[0021] Step 7: Parse the data packet layer by layer. After receiving the data packet, the programmable node determines whether it is a detection packet. The judgment is based on the traffic class field value of the IPv6 header. If the value is 1, it means that the current data packet is a detection packet and continues to be processed. Otherwise, it is a normal data packet and is directly forwarded according to the routing table without any other processing.

[0022] Step 8: Determine whether the current node is the first node on the path. The determination is based on whether the metadata stack in the data packet is empty. If it is empty, it means that metadata has not been added, indicating that the current node is the first node. Perform SRv6 flow table operations according to the destination address and add the corresponding segment routing extension header, which contains a segment list consisting of a subset of the on-demand node list.

[0023] Step 9, save the default bitmap in the detection data packet to a predefined on-demand bitmap register list, which consists of 8 1-bit registers, each bit corresponding to a resource, 1 indicates that it needs to be added, and 0 indicates that it does not need to be added;

[0024] Step 10: determine whether the LOC field of the current node SID matches the IPv6 destination address. If so, it means that the current node is a node in the segment list and is an on-demand measurement node. It is necessary to add metadata, update the corresponding field of the SRv6 header in the data packet, update the destination address, and then continue to step 11. Otherwise, the node only acts as an intermediate transmission node and directly forwards according to the IPv6 destination address.

[0025] The format of the node segment identifier in step 10 is as follows:

[0026] The length of a single node identifier is 128 bits, the same as an IPv6 address. The first 64 bits are the LOC field, which is used to locate the node. The middle 56 bits are the FUNC field, which is used to guide the switch behavior. The last 8 bits are the ARG field, which is used to store a customized bitmap.

[0027] Step 11, obtain the SID of the current node from the IPv6 destination address field, obtain the last 8 bits of the SID, that is, the ARG field, by intercepting, and determine whether it is greater than 0. If it is greater than 0, it means that the current node has a customized bitmap, and the 8-bit content of the ARG field is used as a customized bitmap and updated to the on-demand bitmap register;

[0028] Step 12: Dynamically add metadata items according to the content in the on-demand bitmap. Access each bit register in the on-demand bitmap register list one by one. Each register represents whether the corresponding resource needs to be added. If it is 1, the corresponding resource information of the current node is added to the metadata stack. When adding, the metadata information is wrapped with the node ID and the metadata item index as the basis for tracing after the controller layer receives the summary information. If it is 0, it means that this resource does not need to be added and no operation is required.

[0029] Step 13, determine whether the remaining segments field of SRv6 is equal to 0. If so, it means that the data packet has arrived at the last node in the segment list, that is, the end node of this path. At this time, the Ethernet frame and metadata stack part of the data packet are retained, and the rest of the data packet is discarded. The obtained new data packet is sent as a detection task summary data packet to the in-band telemetry controller for data analysis and storage.

[0030] Compared with the prior art, the present invention adopts the above technical solution and has the following technical effects:

[0031] 1. It can distinguish the different roles of nodes in the network and realize on-demand measurement of a specific set of nodes. The nodes designed for specific network tasks are usually not directly connected. At this time, there is an intermediate transmission node between two on-demand nodes. Such nodes do not need to collect network information. Traditional in-band telemetry does not distinguish between different node types on the path, and treats them equally. The present invention uses SRv6 technology and its SID matching mechanism to selectively select the required nodes for measurement.

[0032] 2. It can realize personalized measurement between different nodes. By utilizing the high programmability of the P4 programming language and designing the default bitmap field in the detection packet header and the ARG field of the SID in the SRv6 header, it can realize on-demand measurement of metadata for each node, ensuring the high flexibility of the detection task.

[0033] 3. High network compatibility. The present invention sets the INT-related design in the UDP payload part according to the principle of "INT over UDP". When the traditional equipment in the network receives the detection data packet but cannot identify it, it will transmit the INT-related information as a normal UDP payload without affecting the parsing of the remaining network protocols. At the same time, the feature of SRv6 updating the IPv6 address is used to ensure that the data packet always maintains the most basic forwarding behavior validity during the forwarding process, that is, forwarding according to the destination address according to the routing table.

[0034] 4. Lightweight mechanism. In terms of bandwidth occupancy, traditional in-band telemetry solutions do not distinguish between types of nodes and metadata, and collect all nodes and all metadata on the path. Although the effective information content is the same, non-required information occupies a lot of valuable bandwidth resources in both data plane bandwidth occupancy and bandwidth occupancy between numerical control planes. The present invention achieves highly on-demand measurement, greatly improves bandwidth utilization efficiency, and can undertake more detection tasks under the same bandwidth conditions. In terms of control plane processing overhead, the present invention adds metadata traceability information while ensuring that the cost is controllable during the metadata addition process. When the control plane receives the metadata stack, it can save the overhead generated by metadata tracing and improve the work efficiency of the control plane. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 The topology diagram of the integrated air-space-ground network constructed;

[0036] Figure 2 To detect the packet structure of the data packet;

[0037] Figure 3 It is the packet processing flow of the programmable switch;

[0038] Figure 4 This is a screenshot of the INT header information after the data packet passes through ladar1;

[0039] Figure 5 This is a screenshot of the INT metadata stack information after the data packet passes through ladar1;

[0040] Figure 6 This is a screenshot of the INT header information after the data packet passes through sal2;

[0041] Figure 7 This is a screenshot of the INT metadata stack information after the data packet passes through sal2;

[0042] Figure 8 This is a screenshot of the data packet after it passes through b3;

[0043] Fig. 9 This is a screenshot of the INT header information after the packet passes through b3;

[0044] Fig.10 This is a screenshot of the INT metadata stack information after passing b3. DETAILED DESCRIPTION

[0045] The present invention is further explained below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the following specific embodiments are only used to illustrate the present invention and are not used to limit the scope of the present invention. It should be noted that the words "front", "rear", "left", "right", "upper" and "lower" used in the following description refer to directions in the accompanying drawings, and the words "inner" and "outer" refer to directions toward or away from the geometric center of a specific component, respectively.

[0046] A lightweight in-band telemetry method for an air-ground integrated computing network in this embodiment includes the following steps:

[0047] Step 1: The application layer integrates user requirements, stores them in the form of key-value pairs, and transmits them to the controller layer;

[0048] Step 2: The controller layer calculates the amount of data that needs to be measured for each node in the on-demand measurement list. The amount of data refers to the sum of the lengths of all metadata that need to be measured for the node, and stores it in the form of a key-value pair together with the node number. Specifically, the node number is used as the key and the amount of data is used as the value. It is stored in the database for subsequent path cutting. The specific calculation formula for the total amount of detection data for a single node is:

[0049]

[0050] in is the detection data length of node i, is the length of resource j, is the value of the jth bit in the bitmap. A value of 1 indicates that the resource is needed, and a value of 0 indicates that the resource is not needed.

[0051] Step 3: construct a logical auxiliary graph to calculate the optimal detection path. Specifically, first access the topology information database to obtain the topology and link parameter information in the network. The example network topology graph is as follows: Figure 1 The node-related parameter settings are shown in Table 1, which mainly include the link connection relationship and link cost. The specific calculation formula of the link cost is:

[0052]

[0053] in is the link cost of link m, and are the proportions of delay and bandwidth in link cost respectively, and are the delay and bandwidth of link m, and are the maximum and minimum delays in the topology database, respectively. and They are the maximum and minimum bandwidths in the topology database, respectively. The delay and bandwidth of the link are comprehensively considered, and then the Dijkstra algorithm is used to solve the shortest path between two nodes based on the link cost. The shortest path is used as the logical link of the logical auxiliary graph. The two ends of each logical link are the starting node and the end node of the physical path, and the logical link cost is calculated. The logical link cost is the sum of the costs of each physical link on the physical path. The specific calculation formula for the logical link cost is:

[0054]

[0055] in represents the logical link cost from node i to node j, For each physical direct path in the shortest path between nodes ij obtained in step 3, represents the nth physical link, Represents the link cost of the nth physical link. A logical complete graph consisting of logical links is constructed as a logical auxiliary graph to assist path planning calculation.

[0056] Table 1 Node parameter settings

[0057]

[0058] Step 4, extract the subgraph consisting of the required node list from the logic auxiliary graph, and use the TSP solving algorithm to solve the optimal logic traversal path;

[0059] Step 5: Based on the node measurement demand data volume, the greedy cutting algorithm is used to cut the logical traversal path. The cutting needs to meet the following two conditions: the total amount of data added by the data packet cannot exceed the network maximum transmission unit limit, and secondly, the amount of data added is as much as possible to obtain the best detection path that meets the network maximum transmission unit limit. The mathematical constraint formula for greedy cutting is:

[0060]

[0061] Where num represents the number of nodes in a single path. Representation Node The length of the detection data, MTU is the maximum transmission unit value of the network, and max (num) indicates the maximum number of nodes detected on a single path. In this example, the nodes to be detected include ladar1, sal1 / sal2, ladar2, b3, ladar1->sal1 / sal2->ladar2->b3->gb2; in this example, the number of nodes is small, and the maximum transmission unit limit does not exceed the limit during the network measurement process, so there is no need to cut, and there is only one detection path.

[0062] Step 6: construct the obtained multiple detection paths into SRv6 flow table entries respectively and send them to the source nodes of the respective paths;

[0063] Step 7: Parse the data packet layer by layer. The data packet structure is as follows: Figure 2 As shown in the figure, after receiving the data packet, the programmable node determines whether it is a detection packet. The judgment is based on the traffic type field value of the IPv6 header. A value of 1 indicates that the current data packet is a detection data packet and continues to be processed. Otherwise, it is a normal data packet and is directly forwarded according to the routing table without other processing. The packet processing flow inside the switch is as follows Figure 3 shown.

[0064] Step 8: Determine whether the current node is the first node on the path. The determination is based on whether the metadata stack in the data packet is empty. If it is empty, it means that metadata has not been added, indicating that the current node is the first node. Perform SRv6 flow table operations according to the destination address and add the corresponding segment routing extension header, which contains a segment list consisting of a subset of the on-demand node list.

[0065] Step 9, save the default bitmap in the detection data packet to a predefined on-demand bitmap register list, which consists of 8 1-bit registers, each bit corresponding to a resource, 1 means it needs to be added, 0 means it does not need to be added, and the bit index of the metadata is shown in Table 2.

[0066] Table 2 Metadata item bitmap index

[0067]

[0068] Step 10, determine whether the LOC field of the current node SID matches the IPv6 destination address. If it matches, it means that the current node is a node in the segment list and is an on-demand measurement node. It is necessary to add metadata, update the corresponding field of the SRv6 header in the data packet, update the destination address, and then continue to step 11. Otherwise, the node only acts as an intermediate transmission node and forwards directly according to the IPv6 destination address. The format of the node segment identifier in step 10 is as follows:

[0069] The length of a single node identifier is 128 bits, the same as an IPv6 address. The first 64 bits are the LOC field, which is used to locate the node. The middle 56 bits are the FUNC field, which is used to guide the switch behavior. The last 8 bits are the ARG field, which is used to store a customized bitmap.

[0070] Step 11, obtain the SID of the current node from the IPv6 destination address field, obtain the last 8 bits of the SID, that is, the ARG field, by intercepting, and determine whether it is greater than 0. If it is greater than 0, it means that the current node has a customized bitmap, and the 8-bit content of the ARG field is used as a customized bitmap and updated to the on-demand bitmap register;

[0071] Step 12: Dynamically add metadata items according to the content in the on-demand bitmap. Access each bit register in the on-demand bitmap register list one by one. Each register represents whether the corresponding resource needs to be added. If it is 1, the corresponding resource information of the current node is added to the metadata stack. When adding, the metadata information is wrapped with the node ID and the metadata item index as the basis for tracing after the controller layer receives the summary information. If it is 0, it means that this resource does not need to be added and no operation is required.

[0072] When the data packet after ladar1 reaches sal2, Figure 4As shown in the figure, we can see that the hop count in the INT detection packet header is already 1, the default bitmap is 0Xa0, that is, 10100000, and the length of the metadata stack collected is 0X0c, that is, 12 bytes. Figure 5 By dividing the elements in the stack, we can see that the first and third items of data have been added to the metadata stack of the data packet.

[0073] When the data packet arrives at sal2, metadata is collected according to the ARG field (8 bits) of the SID. The SID is A3:33::33, and its ARG field is 0X33, that is, 00110011. It is necessary to detect metadata 3, 4, 7, and 8, such as Figure 6 As shown in the figure, the number of hops has increased to 2, and the metadata stack length has also increased to 0X24, that is, 36 bytes. Figure 7 As shown, by dividing the metadata stack, it can be seen that the above four metadata information have been added to the data packet. This part proves that the measurement of on-demand metadata has been achieved.

[0074] Since b3 is not in the segment routing extension header, although the data packet is forwarded through b3, the metadata is not collected. Figure 8 As can be seen, the data packet is transmitted from b3 (mac address 08:00:00:00:55:00) to gb2 (mac address 08:00:00:00:44:00), such as Fig. 9 As shown in the figure, the number of hops in the INT header remains unchanged, and is still 3 before passing b3. Fig.10 As shown, the metadata stack has not changed. The latest metadata collected are the 3rd and 7th items from ladar2, which means that b3 only acts as a transmission node and completes the forwarding task, which shows that the scheme realizes the measurement of on-demand nodes.

[0075] Step 13, determine whether the segment remaining field is equal to 0. If so, it means that the data packet has arrived at the last node in the segment list, that is, the end node of this path. At this time, the Ethernet frame and metadata stack part of the data packet are retained, and the rest of the data packet is discarded. The obtained new data packet is sent as a detection task summary data packet to the in-band telemetry controller for data analysis and storage.

[0076] The technical means disclosed in the scheme of the present invention are not limited to the technical means disclosed in the above-mentioned implementation mode, but also include technical schemes composed of any combination of the above technical features.

Claims

1. A lightweight in-band telemetry method for an integrated space-ground computing network, characterized by: The following steps are involved: Step 1: The application layer integrates user requirements, stores them in the form of key-value pairs, and transmits them to the controller layer; Step 2: The controller layer calculates the total amount of detection data of a single node in the on-demand measurement list, where the amount of data refers to the sum of the lengths of all metadata that need to be measured for the node, and stores it in the form of a key-value pair together with the node identifier. Specifically, the node identifier is used as the key and the amount of data is used as the value. It is stored in the database for subsequent path cutting. Step 3, construct a logical auxiliary graph for calculating the optimal detection path; specifically, first access the topology information database to obtain the topology and link parameter information in the network, including the connection relationship of the link and the link cost, comprehensively consider the link delay and bandwidth, and then use the link cost as a basis to solve the shortest path between two nodes in the network topology, and use the shortest path as the logical link of the logical auxiliary graph. The two ends of each logical link are the starting node and the end node of the physical path, and the logical link cost is calculated. A logical complete graph composed of logical links is constructed as a logical auxiliary graph for auxiliary path planning calculation; Step 4, extracting a subgraph consisting of a required node list from the logic auxiliary graph, and using a traveling salesman problem solving algorithm to solve the optimal logic traversal path; Step 5: Based on the amount of data required for node measurement, the greedy cutting algorithm is used to cut the logical traversal path. The cutting needs to meet the following two conditions: the total amount of data added by the data packet cannot exceed the maximum transmission unit limit of the network, and the number of traversed nodes is maximized, so as to obtain the best detection path that meets the maximum transmission unit limit of the network; Step 6: construct the obtained detection path into an SRv6 flow table entry and send it to each path source node; Step 7: The data packet is parsed layer by layer. After receiving the data packet, the programmable node determines whether it is a detection data packet. The judgment is based on the traffic type field value of the IPv6 header. If the value is 1, it means that the current data packet is a detection data packet and continues to be processed. Otherwise, it is a normal data packet and is directly forwarded according to the routing table without other processing. Step 8: Determine whether the current node is the first node on the path. The determination is based on whether the metadata stack in the data packet is empty. If it is empty, it means that metadata has not been added, indicating that the current node is the first node. Perform SRv6 flow table operations according to the destination address and add the corresponding segment routing extension header, which contains an SRv6 segment list consisting of a subset of the on-demand node list. Step 9, save the default bitmap in the detection data packet to a predefined on-demand bitmap register list, which consists of 8 1-bit registers, each bit corresponding to a resource, 1 indicates that it needs to be added, and 0 indicates that it does not need to be added; Step 10, determine whether the LOC field of the current node segment identifier matches the IPv6 destination address. If it matches, it means that the current node is a node in the segment list and is an on-demand measurement node. It is necessary to add metadata, update the corresponding field of the SRv6 header in the data packet, update the destination address, and then continue to step 11. Otherwise, the node only acts as an intermediate transmission node and forwards directly according to the IPv6 destination address; Step 11, obtain the segment identifier of the current node from the IPv6 destination address field, obtain the last 8 bits of the segment identifier, that is, the ARG field, by intercepting, and determine whether it is greater than 0. If it is greater than 0, it means that the current node has a customized bitmap, and the 8-bit content of the ARG field is used as the customized bitmap and updated to the on-demand bitmap register; Step 12: Dynamically add metadata items according to the content in the on-demand bitmap, access each bit register in the on-demand bitmap register list one by one, each register represents whether the corresponding resource needs to be added. If it is 1, the corresponding resource information of the current node is added to the metadata stack. When adding, the metadata information is wrapped with the node identifier and the metadata item index as the basis for tracing after the controller layer receives the summary information. If it is 0, it means that this resource does not need to be added and no operation is required; Step 13: Check whether the current segment is the last one in the segment list. If so, it means that the data packet has arrived at the last node, that is, the end node of this path. At this time, the Ethernet frame and metadata stack of the data packet are retained, and the rest of the data packet is discarded. The new data packet obtained is sent as a detection task summary data packet to the in-band telemetry controller for data analysis and storage.

2. The lightweight in-band telemetry method for the space-ground integrated computing network according to claim 1 is characterized in that: The specific calculation formula for the total amount of detection data of a single node in step 2 is: ,in is the detection data length of node i, is the length of resource j, is the value of the jth bit in the bitmap. A value of 1 indicates that the resource is needed, and a value of 0 indicates that the resource is not needed.

3. The lightweight in-band telemetry method for the space-ground integrated computing network according to claim 1 is characterized in that: In step 3, the optimal detection path solution algorithm is the Dijkstra algorithm; the specific calculation formula of the link cost is: ;in is the link cost of link m, and are the proportions of delay and bandwidth in link cost respectively, and are the delay and bandwidth of link m, and are the maximum and minimum delays in the topology database, respectively. and They are the maximum and minimum bandwidth values ​​in the topology database.

4. The lightweight in-band telemetry method for the space-ground integrated computing network according to claim 1 is characterized in that: The specific calculation formula of the logical link cost in step 3 is: ,in represents the logical link cost from node i to node j, For each physical direct path in the shortest path between nodes ij obtained in step 3, represents the nth physical link, Indicates the link cost of the nth physical link.

5. The lightweight in-band telemetry method for the space-ground integrated computing network according to claim 1 is characterized in that: The mathematical constraint formula of the greedy cutting algorithm in step 5 is: ; where num represents the number of nodes in a single path, Representation Node The detection data length is , MTU is the maximum transmission unit value of the network, and max(num) represents the maximum number of detection nodes for a single path.

6. The lightweight in-band telemetry method for the space-ground integrated computing network according to claim 1 is characterized in that: The format of the node segment identifier in step 10 is as follows: The length of a single node identifier is 128 bits, the same as an IPv6 address. The first 64 bits are the LOC field, which is used to locate the node. The middle 56 bits are the FUNC field, which is used to guide the switch behavior. The last 8 bits are the ARG field, which is used to store a customized bitmap.

Images