Computing power routing advertisement method and apparatus, and device

Abstract

The present application provides a computing power routing advertisement method and apparatus, and a device. The method comprises: receiving a first service packet, the first service packet comprising a first service feature; if there is no flow table matching the first service feature, and a mapping table comprises a first mapping entry matching the first service feature, sending a second-type computing power route to a computing power routing egress node, the second-type computing power route comprising a first CS-ID corresponding to the first service feature; receiving a third-type computing power route sent by the computing power routing egress node, the third-type computing power route comprising a computing power metric value of each computing power service node among a plurality of computing power service nodes; selecting a target computing power service node on the basis of the computing power metric value of each computing power service node, and establishing a flow table matching the first service feature, the flow table comprising path information; and sending the first service packet to the target computing power service node by means of the path information. The solution of the present application improves the utilization efficiency of computing resources.

Description

A computing power routing advertisement method, device and equipment TECHNICAL FIELD

[0001] The present application relates to the technical field of communication, in particular to a computing power routing advertisement method, device and equipment. BACKGROUND

[0002] With the rapid development of new technologies such as supercomputing, artificial intelligence and big data, the application of computing power services is becoming more and more widespread. Computing power service refers to providing computing power as a service to enterprises or individuals through cloud computing, edge computing and other technical means. Through computing power service, the problems of insufficient computing resources, high cost and low efficiency can be solved, thereby providing flexible computing power for enterprises or individuals, meeting different business needs, and coping with different scale computing tasks, so as to improve computing power and efficiency.

[0003] For the same computing power service (such as video computing power service, artificial intelligence computing power service, etc.), multiple computing power service nodes are usually deployed, and these computing power service nodes jointly implement the computing power service.

[0004] However, when one computing power service corresponds to multiple computing power service nodes, there is no effective implementation way to send the message of a terminal device to which computing power service node when the terminal device accesses the computing power service. This makes it impossible to fully utilize the computing resources of each computing power service node, resulting in waste of computing resources.

[0005] SUMMARY

[0006] The present application provides a computing power routing advertisement method applied to a computing power routing entry node, wherein a mapping table already exists in the computing power routing entry node, the mapping table includes a corresponding relationship between service characteristics and CS-ID, and the mapping table is generated by a first type of computing power routing sent by a computing power routing exit node, including:

[0007] receiving a first service message, wherein the first service message includes a first service characteristic;

[0008] if there is no flow table matching the first service characteristic, and there is a first mapping table item in the mapping table matching the first service characteristic, sending a second type of computing power routing to the computing power routing exit node, wherein the second type of computing power routing includes a first CS-ID corresponding to the first service characteristic;

[0009] receiving a third type of computing power routing sent by the computing power routing exit node, wherein the third type of computing power routing includes a computing power metric value of each computing power service node in a plurality of computing power service nodes, and the plurality of computing power service nodes correspond to the same first CS-ID;

[0010] select a target computing power service node based on the computing power metric value of each computing power service node, and establish a flow table matching the first service feature, the flow table including path information;

[0011] send the first service message to the target computing power service node through the path information.

[0012] The application provides a computing power routing advertisement method applied to a computing power routing exit node, the method comprising:

[0013] sending a first type of computing power routing to a computing power routing entry node, the first type of computing power routing including a service feature and a CS-ID, so that the computing power routing entry node generates a mapping table including a corresponding relationship between the service feature and the CS-ID;

[0014] receiving a second type of computing power routing sent by the computing power routing entry node, the second type of computing power routing including a first CS-ID;

[0015] obtaining a computing power metric value of each computing power service node in a plurality of computing power service nodes corresponding to the first CS-ID;

[0016] sending a third type of computing power routing to the computing power routing entry node, the third type of computing power routing including the computing power metric value of each computing power service node, the third type of computing power routing being used to make the computing power routing entry node select a target computing power service node based on the computing power metric value of each computing power service node.

[0017] The application provides a computing power routing advertisement device applied to a computing power routing entry node, a mapping table already existing in the computing power routing entry node, the mapping table including a corresponding relationship between a service feature and a CS-ID, the mapping table being generated by a first type of computing power routing sent by a computing power routing exit node, the device comprising:

[0018] a receiving module configured to receive a first service message, the first service message including a first service feature;

[0019] a sending module configured to, if there is no flow table matching the first service feature and there is a first mapping table item matching the first service feature in the mapping table, send a second type of computing power routing to the computing power routing exit node, the second type of computing power routing including a first CS-ID corresponding to the first service feature;

[0020] The receiving module is further configured to receive a third type of computing power routing sent by the computing power routing exit node, the third type of computing power routing including a computing power metric value of each computing power service node in a plurality of computing power service nodes, the plurality of computing power service nodes corresponding to the same first CS-ID.

[0021] a processing module configured to select a target computing power service node based on the computing power metric value of each computing power service node, and establish a flow table matching the first service feature, the flow table comprising path information;

[0022] The sending module is further configured to send the first service packet to the target computing power service node through the path information.

[0023] The present application provides a computing power routing advertisement device applied to a computing power routing exit node, the device comprising:

[0024] a sending module configured to send a first type of computing power routing to a computing power routing entry node, the first type of computing power routing comprising a service feature and a CS-ID, so that the computing power routing entry node generates a mapping table comprising a corresponding relationship between the service feature and the CS-ID;

[0025] a receiving module configured to receive a second type of computing power routing sent by the computing power routing entry node, the second type of computing power routing comprising a first CS-ID;

[0026] a obtaining module configured to obtain a computing power metric value of each computing power service node in a plurality of computing power service nodes corresponding to the first CS-ID;

[0027] The sending module is further configured to send a third type of computing power routing to the computing power routing entry node, the third type of computing power routing comprising the computing power metric value of each computing power service node, the third type of computing power routing being used to enable the computing power routing entry node to select a target computing power service node based on the computing power metric value of each computing power service node.

[0028] The present application provides a computing power routing entry node comprising a processor and a machine readable storage medium, the machine readable storage medium storing machine executable instructions capable of being executed by the processor; the processor is configured to execute the machine executable instructions to implement the computing power routing advertisement method of the above examples.

[0029] The present application provides a computing power routing exit node comprising a processor and a machine readable storage medium, the machine readable storage medium storing machine executable instructions capable of being executed by the processor; the processor is configured to execute the machine executable instructions to implement the computing power routing advertisement method of the above examples.

[0030] From the above technical solutions, in the embodiments of the present application, through the interaction of the first type of computing power routing, the second type of computing power routing and the third type of computing power routing, the computing power routing entry node can obtain the computing power metric value of each computing power service node, select the target computing power service node based on the computing power metric value, and send the message to the target computing power service node. In this way, the message can be sent to the computing power service node with the optimal computing power metric value, the computing resources of each computing power service node can be fully utilized, the utilization efficiency of the computing resources can be improved, and the waste of computing resources can be avoided. Through the distributed on-demand subscription computing power routing, the announcement of the computing power metric value can be efficiently realized. BRIEF DESCRIPTION OF DRAWINGS

[0031] FIGS. 1A and 1B are flow diagrams of a computing power routing announcement method in an embodiment of the present application;

[0032] FIG. 2 is a structural diagram of a computing power service system in an embodiment of the present application;

[0033] FIG. 3 is a schematic diagram of a distributed on-demand subscription computing power routing in an embodiment of the present application;

[0034] FIG. 4 is a flow diagram of a computing power routing announcement method in an embodiment of the present application;

[0035] FIG. 5A is a diagram of a computing power metric value corresponding to a subscription CS-ID without RR in the present application;

[0036] FIG. 5B is a diagram of a computing power metric value corresponding to a subscription CS-ID with RR in the present application;

[0037] FIG. 5C is a diagram of a computing power metric value corresponding to a de-subscription CS-ID without RR in the present application;

[0038] FIG. 5D is a diagram of a computing power metric value corresponding to a de-subscription CS-ID with RR in the present application;

[0039] FIGS. 6A-6E are format diagrams of computing power routing with RR in the present application;

[0040] FIGS. 7A and 7B are structural diagrams of a computing power routing announcement device in an embodiment of the present application;

[0041] FIG. 8 is a hardware structure diagram of an electronic device in an embodiment of the present application. DETAILED DESCRIPTION

[0042] The application embodiment proposes a computing power routing advertisement method, which can be applied to a computing power routing entry node. A mapping table exists in the computing power routing entry node, the mapping table includes a corresponding relationship between a service feature and a CS-ID (service identifier), and the mapping table is generated by a first computing power routing sent by a computing power routing exit node. Referring to FIG. 1A, a flowchart of the method is shown, and the method can include:

[0043] Step 111, a first service message is received, and the first service message includes a first service feature.

[0044] Step 112, if there is no flow table matched with the first service feature, and there is a first mapping table item matched with the first service feature in the mapping table, a second computing power routing is sent to the computing power routing exit node, and the second computing power routing includes a first CS-ID corresponding to the first service feature.

[0045] Step 113, a third computing power routing sent by the computing power routing exit node is received, and the third computing power routing includes a computing power metric value of each computing power service node in a plurality of computing power service nodes, and the plurality of computing power service nodes correspond to the same first CS-ID.

[0046] Step 114, based on the computing power metric value of each computing power service node, a target computing power service node is selected, and a flow table matched with the first service feature is established, and the flow table includes path information.

[0047] Step 115, the first service message is sent to the target computing power service node through the path information.

[0048] As can be seen from the above technical solution, in the application embodiment, by interacting the first computing power routing, the second computing power routing and the third computing power routing, the computing power routing entry node can obtain the computing power metric value of each computing power service node, select the target computing power service node based on the computing power metric value, and send the message to the target computing power service node. In this way, the message can be sent to the computing power service node with the optimal computing power metric value, the computing resources of each computing power service node can be fully utilized, the utilization efficiency of the computing resources can be improved, and the waste of the computing resources can be avoided. Through the distributed on-demand subscription of the computing power routing, the advertisement of the computing power metric value can be efficiently implemented.

[0049] The application embodiment proposes a computing power routing advertisement method, which can be applied to a computing power routing exit node, and referring to FIG. 1B, a flowchart of the method is shown, and the method can include:

[0050] Step 121, a first computing power routing is sent to a computing power routing entry node, and the first computing power routing includes a service feature and a CS-ID, so that the computing power routing entry node generates a mapping table, and the mapping table includes a corresponding relationship between the service feature and the CS-ID.

[0051] Step 122, receiving the second type of computing power routing sent by the computing power routing entry node, the second type of computing power routing including the first CS-ID.

[0052] Step 123, obtaining the computing power metric value of each computing power service node in the plurality of computing power service nodes corresponding to the first CS-ID.

[0053] Step 124, sending the third type of computing power routing to the computing power routing entry node, the third type of computing power routing including the computing power metric value of each computing power service node, the third type of computing power routing being used for the computing power routing entry node to select a target computing power service node based on the computing power metric value of each computing power service node.

[0054] From the above technical solutions, in the embodiments of the present application, by interacting the first type of computing power routing, the second type of computing power routing and the third type of computing power routing, the computing power routing entry node can obtain the computing power metric value of each computing power service node, and select a target computing power service node based on the computing power metric value, and send the message to the target computing power service node. In this way, the message can be sent to the computing power service node with the optimal computing power metric value, the computing resources of each computing power service node can be fully utilized, the utilization efficiency of the computing resources can be improved, and the waste of computing resources can be avoided. Through the distributed on-demand subscription computing power routing, the announcement of the computing power metric value can be efficiently implemented.

[0055] In one example, the first type of computing power routing further includes the node position of the computing power routing exit node. The mapping table further includes the node position, i.e., the mapping table includes the correspondence between the service characteristics, the CS-ID and the node position of the computing power routing exit node. Based on this, the computing power routing entry node sending the second type of computing power routing to the computing power routing exit node can include: obtaining the first node position corresponding to the first service characteristics from the first mapping table item; sending the second type of computing power routing to the computing power routing exit node through the first node position.

[0056] In one example, the computing power routing entry node further includes a metric value update table. The computing power routing entry node sending the second type of computing power routing to the computing power routing exit node can include: querying the metric value update table based on the first CS-ID; if the metric value update table does not exist the computing power metric value of the plurality of computing power service nodes corresponding to the first CS-ID, sending the second type of computing power routing to the computing power routing exit node.

[0057] In an example, the computing power routing egress node further comprises a metric value updating table. After the computing power routing egress node receives the second computing power routing, the computing power routing egress node can record the first CS-ID and the computing power metric value of each computing power service node in the metric value updating table. When the computing power metric value of any computing power service node corresponding to the first CS-ID changes, the computing power routing egress node sends a third computing power routing to the computing power routing ingress node, and records the changed computing power metric value of the computing power service node in the metric value updating table.

[0058] Since the computing power metric value of the computing power service node changes multiple times, each change triggers the computing power routing egress node to send the third computing power routing. Therefore, the computing power routing ingress node can receive multiple third computing power routings.

[0059] The computing power routing ingress node further comprises a metric value updating table. After the computing power routing ingress node receives each third computing power routing, the computing power routing ingress node records the correspondence between the first CS-ID and the computing power metric value of each computing power service node in the metric value updating table, and records the update time of each computing power metric value (e.g., the current time as the update time).

[0060] In an example, the computing power routing ingress node can receive a second service message, and the second service message comprises a second service feature. If there is no flow table matching the second service feature, and there is a first mapping table entry matching the second service feature in the mapping table, the computing power routing ingress node queries the metric value updating table based on the first CS-ID.

[0061] If the metric value updating table contains the computing power metric value of multiple computing power service nodes corresponding to the first CS-ID, and the difference between the current time and the update time is less than a threshold value, the computing power routing ingress node selects a target computing power service node based on the computing power metric value of each computing power service node, and establishes a flow table matching the second service feature.

[0062] If the metric value updating table does not contain the computing power metric value of multiple computing power service nodes corresponding to the first CS-ID, or if the metric value updating table contains the computing power metric value of multiple computing power service nodes corresponding to the first CS-ID, and the difference between the current time and the update time is not less than the threshold value, the computing power routing ingress node sends a second computing power routing to the computing power routing egress node.

[0063] In an example, if all flow tables corresponding to the first CS-ID are aged, the computing power routing ingress node sends a revocation routing to the computing power routing egress node, and the revocation routing can comprise the first CS-ID.

[0064] If the computing power routing egress node receives the revocation route, the first CS-ID and the computing power metric value corresponding to the first CS-ID are deleted from the metric value update table. The computing power routing egress node stops sending the third type of computing power route for the first CS-ID to the computing power routing ingress node.

[0065] When the computing power routing ingress node sends the revocation route to the computing power routing egress node, the first node position corresponding to the first CS-ID can be obtained from the first mapping table entry. The computing power routing ingress node sends the revocation route to the computing power routing egress node through the first node position.

[0066] In one example, the revocation route is a revocation route for the second type of computing power route, indicating that the second type of computing power route needs to be revoked. For example, the function of the second type of computing power route is to trigger the computing power routing egress node to send the third type of computing power route for the first CS-ID to the computing power routing ingress node. Therefore, the function of the revocation route of the second type of computing power route is to trigger the computing power routing egress node to stop sending the third type of computing power route for the first CS-ID to the computing power routing ingress node.

[0067] In one example, when the computing power routing egress node sends the first type of computing power route to the computing power routing ingress node, the first type of computing power route can be sent directly by the computing power routing egress node to the computing power routing ingress node; or the first type of computing power route can be forwarded by the computing power routing egress node to the computing power routing ingress node through RR.

[0068] In one example, when the computing power routing ingress node sends the second type of computing power route to the computing power routing egress node, the second type of computing power route can be sent directly by the computing power routing ingress node to the computing power routing egress node; or the second type of computing power route can be forwarded by the computing power routing ingress node to the computing power routing egress node through RR.

[0069] In one example, when the computing power routing egress node sends the third type of computing power route to the computing power routing ingress node, the third type of computing power route can be sent directly by the computing power routing egress node to the computing power routing ingress node; or the third type of computing power route can be forwarded by the computing power routing egress node to the computing power routing ingress node through RR.

[0070] In one example, an address family, i.e., a computing power routing address family, can be added. The computing power routing address family supports the first type of computing power route, the second type of computing power route, and the third type of computing power route. In the embodiments of the present application, the routing information of the computing power routing address cluster can be implemented through BGP protocol packets.

[0071] For example, the BGP protocol message includes an NLRI field. The NLRI field can include an address family information field, a route type field, and a route type specific key field. If the address family information field is a first specified identifier, it indicates that the BGP protocol message is a BGP protocol message for computing power routing.

[0072] If the route type field is a first numerical value, it indicates that the BGP protocol message is a first type of computing power routing, and the route type specific key field carries a first CS-ID. The BGP protocol message further includes a path attribute field related to the NLRI field, which carries service characteristics corresponding to the first CS-ID and a node location of a computing power routing egress node.

[0073] If the route type field is a second numerical value, it indicates that the BGP protocol message is a second type of computing power routing, and the route type specific key field carries the first CS-ID.

[0074] If the route type field is a third numerical value, it indicates that the BGP protocol message is a third type of computing power routing, and the route type specific key field carries the first CS-ID. If the BGP protocol message is a third type of computing power routing, the NLRI field further includes a route type specific non-key field, which carries a computing power metric value of each computing power service node corresponding to the first CS-ID.

[0075] In one example, computing power routing can be extended in an EVPN address cluster, and the EVPN address cluster supports EVPN routing, first type of computing power routing, second type of computing power routing, and third type of computing power routing.

[0076] For example, the BGP protocol message includes an NLRI field, which includes an address family information field, a route type field, and a route type specific key field. If the address family information field is a second specified identifier, it indicates that the BGP protocol message is a BGP protocol message for EVPN routing.

[0077] If the route type field is a fourth numerical value in a preset set, it indicates that the BGP protocol message is a first type of computing power routing, and the route type specific key field carries a first CS-ID. The BGP protocol message can further include a path attribute field related to the NLRI field, which carries service characteristics corresponding to the first CS-ID and a node location of a computing power routing egress node.

[0078] If the route type field is a fifth numerical value in the preset set, it indicates that the BGP protocol message is a second type of computing power routing, and the route type specific key carries the first CS-ID.

[0079] If the route type field is the sixth value in the preset set, it indicates that the BGP protocol packet is a third type of computing power route, and the route type specific key field carries a first CS-ID; wherein, if the BGP protocol packet is the third type of computing power route, the NLRI field further includes a route type specific non-key field, and the route type specific non-key field carries a computing power metric value of each computing power service node corresponding to the first CS-ID.

[0080] The values in the preset set are values not used by the EVPN route.

[0081] The above technical solutions of the embodiments of the present application are described below in combination with specific application scenarios.

[0082] Referring to FIG. 2, it is a structural schematic diagram of a computing power service system, which can include a computing power routing entry node (such as an entry gateway device), a computing power routing exit node (such as an exit gateway device) and a plurality of computing power service nodes, and the computing power service is realized by these computing power service nodes. The computing power routing exit node can be a router, such as a BGP (Border Gateway Protocol) router, and the computing power routing entry node can also be a router, such as a BGP router.

[0083] The computing power routing entry node can be connected with a terminal device, which can be a mobile device (such as a smart phone, etc.), a household device (such as a personal computer, a router, a notebook computer, a television, a video camera, etc.), an enterprise device (such as a personal computer, a router, a notebook computer, etc.).

[0084] The computing power routing exit node can be connected with a cloud platform (cloud management platform), and the cloud platform can include a plurality of computing power service nodes, that is, the computing power routing exit node can be connected with a plurality of computing power service nodes.

[0085] In FIG. 2, taking the plurality of computing power service nodes as three computing power service nodes for example, denoted as computing power service node a1, computing power service node a2 and computing power service node a3. Of course, the number of computing power service nodes can be more. The computing power service node can be a server, or other devices capable of providing computing power resources.

[0086] The plurality of computing power service nodes can collectively realize computing power services (such as video computing power services, artificial intelligence computing power services), and these computing power service nodes are used to provide computing power resources (such as computing resources, etc.) for the computing power services.

[0087] In some scenarios, the computing power service system can include a RR (Route Reflector), and in some scenarios, the computing power service system can not include a RR. The RR, as a kind of router, can serve as a central point for storage and forwarding of routing entries, and can redistribute routing information to other routers.

[0088] In one example, when one computing power service corresponds to multiple computing power service nodes (computing power service node a1, computing power service node a2, and computing power service node a3), there is no effective implementation way for sending a message of a terminal device to which computing power service node when the terminal device accesses the computing power service. This makes it impossible to fully utilize the computing resources of each computing power service node, resulting in waste of computing resources. For example, all messages of terminal devices can be sent to computing power service node a1, while computing power service node a2 and computing power service node a3 are idle, resulting in waste of computing resources of computing power service node a2 and computing power service node a3.

[0089] In view of the above discovery, in the embodiments of the present application, the first type of computing power routing, the second type of computing power routing, and the third type of computing power routing are interacted, so that the computing power routing entry node obtains the computing power metric value of each computing power service node, thereby sending the message to the computing power service node with the optimal computing power metric value, fully utilizing the computing resources of each computing power service node, and improving the utilization efficiency of computing resources.

[0090] In one example, the efficient announcement of the computing power metric value can be implemented by means of distributed on-demand subscription computing power routing. Referring to FIG. 3, which is a schematic diagram of distributed on-demand subscription computing power routing.

[0091] First, the service registration process of the distributed on-demand subscription computing power routing is performed. For example, the computing power routing exit node sends the first type of computing power routing to the computing power routing entry node. The first type of computing power routing includes service characteristics, a CS-ID, and the node position of the computing power routing exit node. The computing power routing exit node registers the correspondence between the service characteristics, the CS-ID, and the node position of the computing power routing exit node at the computing power routing entry node.

[0092] Then, the subscription process of the distributed on-demand subscription computing power routing is performed. For example, when the computing power routing entry node needs to obtain the forwarding path corresponding to the CS-ID (i.e., needs to obtain the computing power metric value), the computing power routing entry node sends the second type of computing power routing to the computing power routing exit node. The second type of computing power routing includes the CS-ID, and the second type of computing power routing is used to on-demand subscribe to the computing power metric value corresponding to the CS-ID.

[0093] Then, the announcement process of the distributed on-demand subscription computing power routing is performed. For example, after the computing power routing egress node receives the second type of computing power routing, it learns that the computing power metric value corresponding to the CS-ID needs to be announced. Therefore, the computing power routing egress node obtains the computing power metric value of each computing power service node corresponding to the CS-ID, and sends the third type of computing power routing to the computing power routing ingress node. The third type of computing power routing includes the computing power metric value of each computing power service node corresponding to the CS-ID, thereby announcing the computing power metric value to the computing power routing ingress node.

[0094] In one example, referring to FIG. 4, a flowchart of a computing power routing announcement method is shown. The method includes:

[0095] Step 401: The computing power routing egress node sends the first type of computing power routing to the computing power routing ingress node. The first type of computing power routing includes the service feature, the CS-ID, and the node location of the computing power routing egress node.

[0096] Step 402: The computing power routing ingress node receives the first type of computing power routing, obtains the service feature, the CS-ID, and the node location of the computing power routing egress node from the first type of computing power routing, and records the corresponding relationship between the service feature, the CS-ID, and the node location of the computing power routing egress node in the mapping table.

[0097] In one example, for the scenario where the RR exists, the computing power routing egress node can send the first type of computing power routing to the RR, and the RR sends the first type of computing power routing to the computing power routing ingress node. For the scenario where the RR does not exist, the computing power routing egress node can directly send the first type of computing power routing to the computing power routing ingress node.

[0098] In one example, the computing power routing egress node and the computing power routing ingress node both start the service of the distributed on-demand subscription computing power routing. The service is used to enable the computing power routing egress node and the computing power routing ingress node to perform the announcement process of the distributed on-demand subscription computing power routing, i.e., to perform the computing power routing announcement method shown in FIG. 4.

[0099] After the computing power routing egress node starts the service of the distributed on-demand subscription computing power routing, the computing power routing egress node can send the first type of computing power routing to the computing power routing ingress node. For example, after the computing power routing egress node starts up, it sends the first type of computing power routing to the computing power routing ingress node.

[0100] In the first type of computing power routing, the CS-ID represents the service identifier, which is a unique identifier for the computing power service (such as a video computing power service or an artificial intelligence computing power service). Different computing power services correspond to different CS-IDs.

[0101] The CS-ID can be used to find a plurality of computing power service nodes for providing computing power resources for the computing power service, that is, the CS-ID corresponds to the computing power service, and the CS-ID corresponds to a plurality of computing power service nodes.

[0102] In the first type of computing power routing, the service feature represents the feature corresponding to the computing power service. For example, if the message includes the service feature corresponding to the computing power service, it means that the message accesses the computing power service. If the message does not include the service feature corresponding to the computing power service, it means that the message does not access the computing power service.

[0103] The service feature can be a five-tuple information (such as source IP address, destination IP address, source port, destination port, and protocol type), a three-tuple information (such as source IP address, destination IP address, and protocol type), a destination IP address, or a message payload feature, and the service feature is not limited.

[0104] In the first type of computing power routing, the node location of the computing power routing exit node represents the location of the computing power routing exit node, which can be an IP address or a device identifier of the computing power routing exit node, and is not limited as long as the message can be sent to the computing power routing exit node based on the node location.

[0105] After the computing power routing entry node receives the first type of computing power routing, the correspondence between the service feature, the CS-ID, and the node location of the computing power routing exit node is recorded in the mapping table, so that the service feature, the CS-ID, and the node location are registered in the computing power routing entry node, and the service feature can be mapped to the CS-ID.

[0106] Step 403, the computing power routing entry node receives the first service message. The first service message can include a service feature, and the service feature in the first service message is denoted as a first service feature.

[0107] In one example, the computing power routing entry node can be connected to a plurality of terminal devices, and the terminal devices can send messages to the computing power routing entry node, which are denoted as service messages (also referred to as data messages) in the following.

[0108] Step 404, if there is no flow table matching the first service feature, and there is a first mapping table item in the mapping table matching the first service feature, the computing power routing entry node sends a second type of computing power routing to the computing power routing exit node, and the second type of computing power routing includes a first CS-ID corresponding to the first service feature.

[0109] In one example, if there is a first mapping table item in the mapping table matching the first service feature, the first CS-ID corresponding to the first service feature can be obtained from the first mapping table item.

[0110] If the mapping table has a first mapping table entry matching the first service feature, the first node position corresponding to the first service feature can be obtained from the first mapping table entry; and the first node position is used by the computing power routing entry node to send the second type of computing power routing to the computing power routing exit node.

[0111] In one example, the computing power routing entry node can query the flow table based on the first service feature (such as five-tuple information, three-tuple information, destination IP address, etc.). If there is a flow table matching the first service feature (see the flow table establishment process in the subsequent steps), the first service message can be directly forwarded through the flow table.

[0112] Obviously, if the first service message accesses the computing power service, the first service message can be sent to the computing power service node. All service messages for the same service feature, i.e., all service messages of the same terminal device accessing the computing power service, can be sent to the same computing power service node based on the flow table.

[0113] In one example, if there is no flow table matching the first service feature, the computing power routing entry node queries the mapping table based on the first service feature. If there is no mapping table entry matching the first service feature in the mapping table, it means that the first service feature is not the feature corresponding to the computing power service, and the first service message is not a message accessing the computing power service. In this case, the computing power routing entry node can send the first service message.

[0114] If the mapping table has a first mapping table entry matching the first service feature (the first mapping table entry includes a first CS-ID corresponding to the first service feature and a first node position), it means that the first service feature is the feature corresponding to the computing power service, and the first service message is a message accessing the computing power service.

[0115] In this case, it is assumed that the first service message is the first message accessing the computing power service, i.e., the first CS-ID corresponding computing power metric value has not been subscribed through the second type of computing power routing. The computing power routing entry node sends the second type of computing power routing to the computing power routing exit node through the first node position (such as the IP address or device identifier of the computing power routing exit node), thereby subscribing the first CS-ID corresponding computing power metric value.

[0116] In one example, the computing power routing entry node also has a metric value update table. If there is no flow table matching the first service feature, and the mapping table has a first mapping table entry matching the first service feature, the computing power routing entry node can query the metric value update table based on the first CS-ID.

[0117] If the first service message is the first message accessing the computing power service, the metric value update table does not exist the computing power metric value corresponding to the first CS-ID. Based on this, if the metric value update table does not exist the computing power metric value of the plurality of computing power service nodes corresponding to the first CS-ID, the computing power routing entry node sends the second type of computing power routing to the computing power routing exit node.

[0118] In one example, for the scenario where the RR exists, the computing power routing entry node sends the second type of computing power routing to the RR, and the RR sends the second type of computing power routing to the computing power routing exit node. For the scenario where the RR does not exist, the computing power routing entry node can directly send the second type of computing power routing to the computing power routing exit node.

[0119] In step 405, the computing power routing exit node receives the second type of computing power routing, which includes the first CS-ID. The computing power routing exit node obtains the computing power metric value of each computing power service node in the plurality of computing power service nodes corresponding to the first CS-ID.

[0120] In one example, after receiving the second type of computing power routing, the computing power routing exit node learns that the computing power routing entry node subscribes to the computing power metric value corresponding to the first CS-ID. Therefore, the computing power routing exit node obtains the computing power metric value of each computing power service node corresponding to the first CS-ID.

[0121] For example, the computing power routing exit node sends a request message to the computing power service node. The computing power service node sends a response message to the computing power routing exit node, which can include the computing power state information of the computing power service node, such as CPU usage, memory usage, capacity usage, bandwidth usage, and the like.

[0122] The computing power routing exit node calculates the computing power metric value (Metric) of the computing power service node based on the computing power state information. For example, if the computing power metric value is a coarse-grained metric value, each state parameter in the computing power state information is normalized to the same numerical interval.

[0123] Then, the normalized plurality of state parameters are subjected to weighted operation, and the weighting coefficients of different state parameters can be the same or different, and the weighted operation result represents the computing power metric value.

[0124] For another example, if the computing power metric value is a fine-grained metric value, each state parameter in the computing power state information is normalized to the same numerical interval.

[0125] The normalized plurality of state parameters can be used as the computing power metric value, such as the normalized CPU usage, memory usage, capacity usage, bandwidth usage, and the like.

[0126] Of course, the above is only an example of obtaining the computing power metric value, and the computing power metric value is not limited.

[0127] Step 406, the computing power routing exit node sends a third type of computing power routing to the computing power routing entry node, the third type of computing power routing including the computing power metric value of each computing power service node in the plurality of computing power service nodes corresponding to the first CS-ID. The computing power routing entry node receives the third type of computing power routing.

[0128] In one example, for the scenario where the RR exists, the computing power routing exit node can send the third type of computing power routing to the RR. The RR sends the third type of computing power routing to the computing power routing entry node. For the scenario where the RR does not exist, the computing power routing exit node can directly send the third type of computing power routing to the computing power routing entry node.

[0129] In the third type of computing power routing, in addition to the computing power metric value, the third type of computing power routing can include the first CS-ID and the unique identifier of each computing power service node. The unique identifier is the IP address or CIS-ID of the computing power service node, which can uniquely represent the computing power service node. Taking the CIS-ID as an example for illustration.

[0130] The CIS-ID represents the service instance identifier, and the CIS-ID is a unique identifier for the computing power service node. Different computing power service nodes can correspond to different CIS-IDs. Through the CIS-ID, the computing power service node for providing computing power resources for the computing power service can be found, that is, the CIS-ID corresponds to the computing power service node.

[0131] For example, the third type of computing power routing can include the CS-ID, the CIS-ID of the computing power service node a1, the computing power metric value of the computing power service node a1, the CIS-ID of the computing power service node a2, the computing power metric value of the computing power service node a2, the CIS-ID of the computing power service node a3, and the computing power metric value of the computing power service node a3.

[0132] Step 407, the computing power routing entry node selects one computing power service node as a target computing power service node from all computing power service nodes based on the computing power metric value of each computing power service node.

[0133] After receiving the third type of computing power routing, the computing power routing entry node can obtain the computing power metric value of each computing power service node corresponding to the first CS-ID from the third type of computing power routing.

[0134] If the computing power metric value is a coarse-grained metric value, the computing power routing entry node takes the computing power service node with the optimal computing power metric value as the target computing power service node. For example, assuming that the smaller the computing power metric value is, the more optimal the computing power metric value is, the computing power routing entry node takes the computing power service node with the smallest computing power metric value as the target computing power service node. Assuming that the larger the computing power metric value is, the more optimal the computing power metric value is, the computing power routing entry node takes the computing power service node with the largest computing power metric value as the target computing power service node.

[0135] If the computing power metric value is a fine-grained metric value, assuming that a CPU priority strategy is adopted, the computing power routing entry node takes the computing power service node with the smallest CPU usage rate as the target computing power service node (if the CPU usage rates of multiple computing power service nodes are all the smallest, other state parameters are considered).

[0136] Assuming that a memory priority strategy is adopted, the computing power routing entry node takes the computing power service node with the smallest memory usage rate as the target computing power service node.

[0137] Assuming that a balanced measurement is adopted, the computing power routing entry node performs weighted operation on the state parameters such as CPU usage rate, memory usage rate, capacity usage rate, and bandwidth usage rate to obtain a weighted operation result. The computing power routing entry node takes the computing power service node with the smallest weighted operation result as the target computing power service node.

[0138] In one example, in addition to the computing power metric value, the computing power routing entry node can also consider a network state value when selecting the target computing power service node. For example, a computing power service node can be selected as the target computing power service node from all computing power service nodes based on the computing power metric value and the network state value.

[0139] For example, the computing power service system can also include a network controller. The network controller can obtain network state information from the computing power routing entry node to the computing power service node, which can include available bandwidth, delay, packet loss rate, etc. The network controller sends the network state information to the computing power routing entry node.

[0140] The computing power routing entry node can determine a network state value based on the network state information. If the network state information includes available bandwidth, the network state value can be proportional to the available bandwidth, that is, the larger the available bandwidth is, the more optimal the network state value is. If the network state information includes delay, the network state value can be inversely proportional to the delay, that is, the smaller the delay is, the more optimal the network state value is. If the network state information includes packet loss rate, the network state value can be inversely proportional to the packet loss rate, that is, the smaller the packet loss rate is, the more optimal the network state value is.

[0141] The computing power routing entry node can perform weighted operation on the computing power metric value and the network state value to obtain a comprehensive metric value. The computing power routing entry node takes the computing power service node with the optimal comprehensive metric value as the target computing power service node.

[0142] At step 408, the computing power routing entry node establishes a flow table matching the first service feature based on the target computing power service node. The match item of the flow table includes the first service feature, and the action item of the flow table includes path information, and the path information is used to send the first service message to the target computing power service node.

[0143] At step 409, the computing power routing entry node sends the first service message to the target computing power service node through the path information. For example, the flow table matching the first service feature of the first service message is queried, and the first service message is sent to the target computing power service node through the path information in the flow table.

[0144] When the computing power routing entry node sends the first service message to the target computing power service node, the computing power routing entry node first sends the first service message to the computing power routing exit node, and then the computing power routing exit node sends the first service message to the target computing power service node.

[0145] In one example, in order to send the first service message to the computing power routing exit node, the computing power routing entry node can perform tunnel encapsulation on the first service message. That is, the computing power routing entry node encapsulates a tunnel header for the first service message. The destination IP address of the tunnel header is the IP address of the computing power routing exit node (such as obtained from the mapping table). For this purpose, when the flow table matching the first service feature is established, the path information includes the IP address of the computing power routing exit node.

[0146] Alternatively, the computing power routing entry node can send the first service message to the computing power routing exit node in SRv6 mode. For this purpose, when the flow table matching the first service feature is established, the path information can include a SID list of the forwarding path of SRv6, and the SID list is used to send the first service message to the computing power routing exit node.

[0147] Of course, the above is only two examples of path information, and no limitation is made thereto, as long as the first service message can be sent to the computing power routing exit node based on the path information.

[0148] In one example, in order to send the first service message from the computing power routing exit node to the target computing power service node, the computing power routing entry node can modify the destination IP address of the first service message to the IP address of the target computing power service node. In this way, after receiving the first service message, the computing power routing exit node forwards the first service message to the target computing power service node based on the destination IP address of the first service message.

[0149] Alternatively, the computing power routing ingress node can encapsulate the CIS-ID of the target computing power service node (obtained from the third type of computing power routing) in the first service message. In this way, after receiving the first service message, the computing power routing egress node sends the first service message to the target computing power service node corresponding to the CIS-ID. For example, the first service message can be sent to the target computing power service node based on the IP address of the target computing power service node, or the first service message can be sent to the target computing power service node in an SRv6 manner, without limitation.

[0150] To this end, when establishing the flow table matched with the first service feature, the path information can include the CIS-ID of the target computing power service node, which is encapsulated in the first service message.

[0151] Of course, the above is only an example of path information, without limitation, as long as the first service message can be sent from the computing power routing egress node to the target computing power service node based on the path information.

[0152] Obviously, after the flow table matched with the first service feature is established in the above manner, based on the path information of the flow table, the computing power routing ingress node can send the first service message to the computing power routing egress node, and the computing power routing egress node can send the first service message to the target computing power service node.

[0153] For subsequent service messages of the first service feature, after the computing power routing ingress node receives the service message, since there is a flow table matched with the first service feature, the computing power routing ingress node can send the service message to the target computing power service node through the path information of the flow table.

[0154] In one example, for step 405, after the computing power routing egress node receives the second type of computing power routing and obtains the computing power metric value of each computing power service node corresponding to the first CS-ID, the computing power routing egress node can also record the first CS-ID and the computing power metric value of each computing power service node in the metric value update table. Referring to Table 1, which shows an example of the metric value update table.

[0155] Table 1

[0156] On this basis, for each CS-ID in the metric value update table, the computing power metric value of each computing power service node corresponding to the CS-ID can be obtained periodically, such as once every 1 minute.

[0157] After obtaining the computing power metric value each time, if the computing power metric value of all computing power service nodes has not changed, the metric value update table remains unchanged, and the third type of computing power routing is not re-sent.

[0158] When the computing power metric value of any computing power service node corresponding to the first CS-ID (any CS-ID in the metric value update table) changes, such as the computing power metric value A changes to the computing power metric value A', the computing power routing exit node re-sends the third type of computing power routing to the computing power routing entry node. The third type of computing power routing can include the first CS-ID, the CIS-ID of each computing power service node, and the computing power metric value of each computing power service node.

[0159] In addition, the computing power routing exit node records the changed computing power metric value of the computing power service node in the metric value update table, such as updating the computing power metric value A in the metric value update table to the computing power metric value A'.

[0160] In one example, for step 406, after the computing power routing entry node receives the third type of computing power routing, the computing power routing entry node can also record the first CS-ID and the computing power metric value of each computing power service node corresponding to the first CS-ID in the metric value update table, and record the update time of the computing power metric value (such as the current time can be used as the update time). Referring to Table 2, which shows an example of the metric value update table.

[0161] Table 2

[0162] On this basis, the computing power routing entry node updates the metric value update table each time it receives the third type of computing power routing, that is, updates the computing power metric value of each computing power service node corresponding to the first CS-ID in the metric value update table, and records the update time of the computing power metric value.

[0163] In one example, for step 404, after the computing power routing entry node receives the first service message, if there is no flow table matching the first service feature, and there is a first CS-ID corresponding to the first service feature in the mapping table, the computing power routing entry node can also query the metric value update table through the first CS-ID. If the metric value update table does not have the computing power metric value corresponding to the first CS-ID, the computing power routing entry node sends the second type of computing power routing to the computing power routing exit node.

[0164] In one example, the computing power routing entry node can also receive a second service message (such as a service message sent by another terminal device), and the second service message includes a second service feature.

[0165] For step 404, if there is no flow table matching the second service feature, and there is a first mapping table item (the first mapping table item includes a first CS-ID and a first node location) matching the second service feature in the mapping table, the second service feature is the feature corresponding to the computing power service.

[0166] In this case, the second service packet is not the first packet accessing the computing power service, i.e., the computing power routing entry node has subscribed to the computing power metric value corresponding to the first CS-ID through the second type of computing power routing. Therefore, the computing power routing entry node can query the metric value update table (see Table 2) through the first CS-ID.

[0167] If the metric value update table contains the computing power metric values of the plurality of computing power service nodes corresponding to the first CS-ID, and the difference between the current time and the update time is less than the threshold value, the computing power routing entry node can obtain the computing power metric values of the plurality of computing power service nodes corresponding to the first CS-ID from the metric value update table.

[0168] Based on the computing power metric value of each computing power service node, the computing power routing entry node selects one computing power service node from all the computing power service nodes as a target computing power service node (see step 407). A flow table for matching the second service feature can be established based on the target computing power service node (see step 408). The computing power routing entry node sends the second service packet to the target computing power service node through the flow table.

[0169] If the metric value update table does not contain the computing power metric values of the plurality of computing power service nodes corresponding to the first CS-ID, or if the metric value update table contains the computing power metric values of the plurality of computing power service nodes corresponding to the first CS-ID, and the difference between the current time and the update time is not less than the threshold value (i.e., the content of the metric value update table is not trusted, and may be incorrect data), the computing power routing entry node sends the second type of computing power routing to the computing power routing exit node, and the second type of computing power routing includes the first CS-ID, i.e., steps 404-409 are repeated.

[0170] In one example, flow table aging refers to: the computing power routing entry node updates the last hit time of the flow table each time a service packet for the flow table is received. If the difference between the current time and the last hit time of the flow table is greater than the time threshold, the flow table is aged, i.e., the flow table needs to be deleted.

[0171] The flow table corresponding to the CS-ID refers to: if the service feature corresponds to the CS-ID (the correspondence between the service feature and the CS-ID is obtained from the mapping table), the flow table corresponding to the service feature is the flow table corresponding to the CS-ID.

[0172] If all the flow tables corresponding to the first CS-ID are aged, the computing power routing entry node sends a route revocation (which can also be referred to as a route unsubscribing) to the computing power routing exit node, and the route revocation includes the first CS-ID.

[0173] For example, the computing power routing entry node obtains the node position of the computing power routing exit node corresponding to the first CS-ID from the mapping table, and sends the revocation route to the computing power routing exit node based on the node position.

[0174] For example, the revocation route is a revocation route for the second type of computing power routing, indicating that the second type of computing power routing needs to be revoked, and the revocation route is used to make the computing power routing exit node stop sending the third type of computing power routing.

[0175] For the scenario where the RR exists, the computing power routing entry node sends the revocation route to the RR, and the RR sends the revocation route to the computing power routing exit node. For the scenario where the RR does not exist, the computing power routing entry node can directly send the revocation route to the computing power routing exit node.

[0176] After receiving the revocation route, the computing power routing exit node obtains the first CS-ID from the revocation route. Since the revocation route is used to revoke the second type of computing power routing, and the revocation route includes the first CS-ID, the computing power routing exit node can delete the first CS-ID and the computing power metric value of the Megalith computing power service node corresponding to the first CS-ID from the metric value update table (see Table 1).

[0177] After deleting the entry corresponding to the first CS-ID from the metric value update table, the computing power routing exit node can stop sending the third type of computing power routing for the first CS-ID to the computing power routing entry node.

[0178] In one example, referring to FIG. 5A, a diagram for subscribing to the computing power metric value of the CS-ID without RR is shown. In step 1, Egress1 (computing power routing exit node) sends a 1-type registration route (first type of computing power routing) to Ingress (computing power routing entry node), and the 1-type registration route includes a CS-ID, which is 1.

[0179] In step 2, Ingress receives service traffic (i.e., service packet), and maps the CS-ID as 1 according to the characteristics in the 1-type registration route. For example, the 1-type registration route also includes service characteristics, and the CS-ID is mapped as 1 according to the service characteristics in the service traffic and the service characteristics in the 1-type registration route.

[0180] In step 3, Ingress sends a 2-type subscription route (the 2-type subscription route is the second type of computing power routing) to Egress1, and the 2-type subscription route can include a CS-ID, which is 1.

[0181] Step 4, Egress 1 sends a 3-type metric update route (3-type metric update route is a third-type computing power route) to Ingress, the 3-type metric update route can include CS-ID, CIS-ID and metric value of computing power, and the metric value of computing power can be denoted as metric. For example, CS-ID is 1, CIS-ID is 1, and metric is 10.

[0182] Step 5, Egress 2 sends a 1-type registration route to Ingress, and CS-ID is 1.

[0183] Step 6, Ingress sends a 2-type subscription route to Egress 2, and CS-ID is 1.

[0184] Step 7, Egress 2 sends a 3-type metric update route to Ingress, and the 3-type metric update route can include CS-ID, CIS-ID and metric, CS-ID is 1, CIS-ID is 2, and metric is 10.

[0185] Referring to FIG. 5B, it is a schematic diagram of the computing power metric value corresponding to the subscribed CS-ID with RR.

[0186] Step 1, Egress 1 sends a 1-type registration route (first-type computing power route) to RR, and RR sends a 1-type registration route to Ingress, and the 1-type registration route includes CS-ID, and CS-ID is 1.

[0187] Step 2, Ingress receives service traffic (i.e. service packet), and according to the characteristics in the 1-type registration route, maps out that CS-ID is 1. For example, the 1-type registration route further includes service characteristics, and according to the service characteristics in the service traffic and the service characteristics in the 1-type registration route, maps out that CS-ID is 1.

[0188] Step 3, Ingress sends a 2-type subscription route (second-type computing power route) to RR, and RR sends a 2-type subscription route to Egress 1, and the 2-type subscription route can include CS-ID, and CS-ID is 1.

[0189] Step 4, Egress 1 sends a 3-type metric update route (third-type computing power route) to RR, and RR sends a 3-type metric update route to Ingress, and the 3-type metric update route can include CS-ID, CIS-ID and metric. For example, CS-ID is 1, CIS-ID is 1, and metric is 10.

[0190] Step 5, Egress 2 sends a 1-type registration route to RR, and RR sends a 1-type registration route to Ingress, and the 1-type registration route can include CS-ID, for example, CS-ID is 1.

[0191] Step 6, Ingress sends a Class 2 subscribe route to RR, RR sends a Class 2 subscribe route to Egress 2, the Class 2 subscribe route can include CS-ID, say, CS-ID is 1.

[0192] Step 7, Egress 2 sends a Class 3 metric update route to RR, RR sends a Class 3 metric update route to Ingress, the Class 3 metric update route can include CS-ID is 1, CIS-ID is 2, metric is 10.

[0193] Referring to FIG. 5C, a diagram of unsubscribing the computing power metric value corresponding to the CS-ID is shown.

[0194] Step 1, Ingress determines that all flow tables of this service are aged, and initiates unsubscribing.

[0195] Step 2, Ingress sends a Class 2 unsubscribe route (i.e., a revocation route for the second computing power route) to Egress 1, the Class 2 unsubscribe route can include CS-ID, and CS-ID is 1. Ingress sends a Class 2 unsubscribe route to Egress 2, the Class 2 unsubscribe route can include CS-ID, and CS-ID is 1.

[0196] Step 3, Egress 1 determines that there is no subscriber, and sends a revocation Class 3 metric update route to Ingress, the revocation Class 3 metric update route can include CS-ID is 1, CIS-ID is 1.

[0197] Step 4, Egress 2 determines that there is no subscriber, and sends a revocation Class 3 metric update route to Ingress, the revocation Class 3 metric update route can include CS-ID is 1, CIS-ID is 2.

[0198] Referring to FIG. 5D, a diagram of unsubscribing the computing power metric value corresponding to the CS-ID is shown.

[0199] Step 1, Ingress determines that all flow tables of this service are aged, and initiates unsubscribing.

[0200] Step 2, Ingress sends a Class 2 unsubscribe route (i.e., a revocation route for the second computing power route) to RR, RR sends a Class 2 unsubscribe route to Egress 1, and RR sends a Class 2 unsubscribe route to Egress 2, the Class 2 unsubscribe route can include CS-ID, and CS-ID is 1.

[0201] Step 3, Egress1 determines that there are no subscribers, sends a withdraw Class-3 metric update route to RR, RR sends a withdraw Class-3 metric update route to Ingress, the route includes CS-ID 1, CIS-ID 1.

[0202] Step 4, Egress2 determines that there are no subscribers, sends a withdraw Class-3 metric update route to RR, RR sends a withdraw Class-3 metric update route to Ingress, the route includes CS-ID 1, CIS-ID 2.

[0203] In one example, a computing power routing address cluster (i.e., an address cluster is added as a computing power routing address cluster) can be added in a BGP protocol message, and the computing power routing address cluster can support a first type of computing power routing (Class-1 computing power routing), a second type of computing power routing (Class-2 computing power routing), and a third type of computing power routing (Class-3 computing power routing). Adding a computing power routing address cluster means that the added computing power routing address cluster is only used to transmit computing power routing and is not used to transmit other types of BGP routing except computing power routing.

[0204] For example, taking the UPDATE message of the BGP protocol message as an example, the BGP protocol message can include an Unfeasible Routes Length field, a Withdraw Routes field, a Total Path Attribute Length field, a Path Attribute field, and an NLRI (Network Layer Reachability Information) field. By improving these fields, a computing power routing address cluster can be added.

[0205] In order to support the first type of computing power routing, the following improvements can be made to the BGP protocol message:

[0206] The BGP protocol message can include an NLRI field and a path attribute field related to the NLRI field. The NLRI field can include an address family information field, a route type field, and a route type specific key field.

[0207] The address family information (Address Family Information) field can be a first designated identifier, and the first designated identifier indicates that the BGP protocol message is a BGP protocol message for computing power routing.

[0208] For example, the address family information field is used to represent the address cluster of the BGP protocol message, IPv6 address family, multicast address family, EVPN address family, etc. On this basis, the BGP protocol message can support the newly added computing power routing address cluster, and the newly added computing power routing address cluster is represented by the first specified identifier.

[0209] The address family information field can be composed of 2 bytes of AFI (Address Family Identifier) and 1 byte of SAFI (Subsequent Address Family Identifier). Therefore, by setting AFI to value A and SAFI to value B, the first specified identifier is represented. As for the value A and the value B, they can be configured arbitrarily, and the values of the existing address cluster should not be repeated.

[0210] In addition to the address family information field, the NLRI field can also include a route type field, a length field, and a route type specific key field. Referring to FIG. 6A, it is a schematic diagram of the NLRI field.

[0211] The Route Type field represents the route type field. If the route type field is the first value, it means that the BGP protocol message is the first type of computing power routing. For example, if the Route Type field is 1, it means that the BGP protocol message is the first type of computing power routing. The length of the Route Type field can be 1 byte.

[0212] The Length field represents the length field. The Length field represents the length of the first type of computing power routing. The length of the Length field can be 1 byte, and the length of the Length field can also be 2 bytes.

[0213] The Route Type specific Key field represents the route type specific key field, and the route type specific key field carries the first CS-ID. Referring to FIG. 6B, it is a schematic diagram of the route type specific key field of the NLRI field. The route type specific key field can include an RD field, a CS-ID field, and a CIS-ID field.

[0214] The RD field is used to carry the route distinguisher, and the length of the RD field can be 8 bytes. This route distinguisher is not limited. The CS-ID field is used to carry the CS-ID (service identifier), that is, to carry the first CS-ID described above. The length of the CS-ID field can be 4 bytes. The CIS-ID field is used to carry the CIS-ID (that is, the service instance identifier). The length of the CIS-ID field can be 4 bytes.

[0215] The BGP protocol packet can include a path attribute field related to the NLRI field, and the path attribute field is used to carry the node location of the service feature and the computing power routing egress node corresponding to the first CS-ID. For example, the path attribute field can include a Locator TLV, and the Locator TLV can include the node location of the computing power routing egress node, or the node location of the computing power routing egress node and the service feature.

[0216] In order to support the second type of computing power routing, the following improvements can be made to the BGP protocol packet:

[0217] The BGP protocol packet can include an NLRI field, and the NLRI field can include an address family information field, a route type field, and a route type specific key field. The address family information field can be a first specified identifier, and the first specified identifier indicates that the BGP protocol packet is a BGP protocol packet for computing power routing.

[0218] In addition to the address family information field, the NLRI field can also include a route type field, a length field, and a route type specific key field, as shown in FIG. 6A, which is a schematic diagram of the NLRI field.

[0219] The Route Type field indicates the route type field, and if the route type field is a second value, it indicates that the BGP protocol packet is a second type of computing power routing. For example, if the Route Type field is 2, it indicates that the BGP protocol packet is a second type of computing power routing. The length of the Route Type field can be 1 byte.

[0220] The Length field indicates the length field, and the Length field indicates the length of the second type of computing power routing. The length of the Length field can be 1 byte, and the length of the Length field can also be 2 bytes.

[0221] The Route Type specific Key field indicates the route type specific key field, and the route type specific key field carries the first CS-ID. As shown in FIG. 6C, which is a schematic diagram of the route type specific key field of the NLRI field. The route type specific key field can include an RD field and a CS-ID field.

[0222] The RD field is used to carry a route distinguisher, and the length of the RD field can be 8 bytes, without limitation on the route distinguisher. The CS-ID field is used to carry the CS-ID (service identifier), i.e., to carry the first CS-ID described above, and the length of the CS-ID field can be 4 bytes.

[0223] The BGP protocol packet can also include a path attribute field related to the NLRI field, and the path attribute field can include a Subscribe Option TLV, which is referred to as a Flag TLV. Of course, the Flag TLV is optional, and the path attribute field can also not include the Flag TLV.

[0224] The value of the Flag TLV indicates whether aggregation is supported, whether the original metric is requested, and the like.

[0225] If the value of the Flag TLV indicates that aggregation is supported, it means that the computing power routing entry node subscribes to a coarse-grained computing power metric value, that is, the original metric needs to be aggregated to obtain a coarse-grained computing power metric value.

[0226] If the value of the Flag TLV indicates that the original metric is requested, it means that the computing power routing entry node subscribes to a fine-grained computing power metric value, that is, the computing power routing exit node does not need to aggregate the original metric.

[0227] In order to support the third type of computing power routing, the following improvements can be made to the BGP protocol packet:

[0228] The BGP protocol packet includes an NLRI field, and the NLRI field includes an address family information field, a route type field, a route type specific key field, and a route type specific non-key field. The address family information field can be a first specified identifier, indicating that the BGP protocol packet is a BGP protocol packet for computing power routing. In addition to the address family information field, the NLRI field also includes a route type field, a length field, a route type specific key field, and a route type specific non-key field. Referring to FIG. 6D, which is a schematic diagram of the NLRI field.

[0229] The Route Type field indicates the route type field, and if the route type field is a third value, it means that the BGP protocol packet is a third type of computing power routing. For example, if the Route Type field is 3, it means that the BGP protocol packet is a third type of computing power routing. The length of the Route Type field can be 1 byte.

[0230] The Length field indicates the length field, and the Length field indicates the length of the second type of computing power routing. The length of the Length field can be 1 byte, and the length of the Length field can also be 2 bytes.

[0231] The Route Type specific Key field represents a route type specific key field, and the route type specific key field carries the first CS-ID. Referring to FIG. 6B, which is a schematic diagram of the route type specific key field of the NLRI field. The route type specific key field can include an RD field, a CS-ID field, and a CIS-ID field.

[0232] The RD field is used to carry a route identifier, and the length of the RD field can be 8 bytes. The CS-ID field is used to carry the CS-ID, i.e., to carry the first CS-ID, and the length of the CS-ID field can be 4 bytes. The CIS-ID field is used to carry the CIS-ID, and the length of the CIS-ID field can be 4 bytes.

[0233] The Route Type specific Non-Key Fields (Route Type Non-Key for short) field represents a route type specific non-key field, and the route type specific non-key field is used to carry the computing power metric value of each computing power service node corresponding to the first CS-ID. For example, referring to FIG. 6E, which is a schematic diagram of the route type specific non-key field of the NLRI field.

[0234] The route type specific non-key field can include a Metric TLV and a Raw Metric TLV. The Metric TLV carries a coarse-grained computing power metric value, and the Raw Metric TLV carries a fine-grained metric value. If the computing power routing entry node subscribes to a coarse-grained computing power metric value, the coarse-grained computing power metric value is carried through the Metric TLV, and the Raw Metric TLV is empty. If the computing power routing entry node subscribes to a fine-grained computing power metric value, the fine-grained computing power metric value is carried through the Raw Metric TLV, and the Metric TLV is empty.

[0235] The BGP protocol message can also include a path attribute field related to the NLRI field, and the path attribute field can include a Priority TLV. The Priority TLV represents the priority of the third type of computing power routing. In this way, a higher priority can be set in the Priority TLV, so as to set a higher priority for the third type of computing power routing, so that the network can preferentially transmit the third type of computing power routing.

[0236] In one example, computing power routing can be extended into the EVPN (Ethernet Virtual Private Network) address cluster of BGP protocol messages (i.e., the EVPN address cluster is used as a computing power routing address cluster). The EVPN address cluster supports EVPN routing, Type 1 computing power routing (Type 1 routing of computing power routing), Type 2 computing power routing (Type 2 routing of computing power routing), and Type 3 computing power routing (Type 3 routing of computing power routing).

[0237] Extending computing power routing within an EVPN address cluster refers to expanding an existing EVPN address cluster that is used exclusively for transmitting EVPN routes, such as EVPN Type 2, Type 3, and Type 5 routes. In addition, the EVPN address cluster is also used for transmitting computing power routes, such as Type 1, Type 2, and Type 3 computing power routes. Therefore, the EVPN address cluster needs to be extended to support EVPN routes, Type 1 computing power routes, Type 2 computing power routes, and Type 3 computing power routes simultaneously.

[0238] To support Type I computing power routing, the following improvements can be made to BGP protocol messages:

[0239] BGP protocol messages may include an NLRI field and path attribute fields associated with the NLRI field. The NLRI field may include an address family information field, a route type field, and a route type specific key field.

[0240] The address family information field can be a second designated identifier, which indicates that the BGP protocol message is a BGP protocol message for EVPN routing. That is, the second designated identifier represents the EVPN address family in the BGP protocol message. For example, the address family information field can consist of 2 bytes of AFI and 1 byte of SAFI, and the combined value of AFI and SAFI represents the second designated identifier.

[0241] In addition to the address family information field, the NLRI field may also include a route type field, a length field, and a route type specific key field, as shown in Figure 6A, which is a schematic diagram of the NLRI field.

[0242] The Route Type field indicates the route type. If the Route Type field is the fourth value in the preset set, it indicates that the BGP protocol message is a Type 1 computing route. For example, the values ​​in the preset set are values ​​not used by EVPN routes; that is, at least three values ​​not used by EVPN routes are added to the preset set, such as 15, 16, 17, etc. Based on this, Type 1 computing routes occupy a certain value in the preset set (denoted as the fourth value). For example, if the Route Type field is 15, it indicates that the BGP protocol message is a Type 1 computing route. The length of the Route Type field can be 1 byte.

[0243] The Length field represents the length of the first type of computing power route. The length of the Length field can be 1 byte or 2 bytes.

[0244] The Route Type Specific Key field represents the route type specific key field, and this field carries the first CS-ID. Referring to Figure 6B, the Route Type Specific Key field can include an RD field, a CS-ID field, and a CIS-ID field. The RD field carries the route identifier and can be 8 bytes long. The CS-ID field carries the CS-ID and can be 4 bytes long. The CIS-ID field carries the CIS-ID and can also be 4 bytes long.

[0245] BGP protocol messages may include path attribute fields related to the NLRI field. The path attribute fields are used to carry the service characteristics corresponding to the first CS-ID and the node location of the computing power routing exit node.

[0246] To support Type II computing power routing, the following improvements can be made to BGP protocol messages:

[0247] BGP protocol messages may include an NLRI field, which may include an address family information field, a route type field, and a route type specific key field. The address family information field can be a second designated identifier, which represents the EVPN address family in the BGP protocol message.

[0248] In addition to the address family information field, the NLRI field may also include a route type field, a length field, and a route type specific key field, as shown in Figure 6A, which is a schematic diagram of the NLRI field.

[0249] The Route Type field indicates the route type. If the Route Type field is the fifth value in the preset set, it indicates that the BGP protocol message is a Type 2 computing route. For example, if the values ​​in the preset set are values ​​not used by EVPN routes, then a Type 2 computing route occupies a value in the preset set (denoted as the fifth value). For instance, if the Route Type field is 16, it indicates that the BGP protocol message is a Type 2 computing route. The Route Type field can be 1 byte long.

[0250] The Length field represents the length of the second type of computing power route. The length of the Length field can be 1 byte or 2 bytes.

[0251] The Route Type Specific Key field represents a route type-specific key field, and this field carries the first CS-ID. For example, the Route Type Specific Key field can include an RD field and a CS-ID field. The RD field carries the route identifier and can be 8 bytes long. The CS-ID field carries the CS-ID (Service Identifier) ​​and can be 4 bytes long.

[0252] BGP protocol messages may also include path attribute fields related to the NLRI field. Path attribute fields may include Subscribe Option TLV, which is denoted as Flag TLV. The value of Flag TLV indicates whether aggregation is supported, whether raw metrics are requested, etc.

[0253] To support Type 3 computing power routing, the following improvements can be made to BGP protocol messages:

[0254] BGP protocol messages may include an NLRI field, which can include an address family information field, a route type field, and a route type-specific key field. The address family information field can be a second designated identifier, representing the EVPN address family in the BGP protocol message. In addition to the address family information field, the NLRI field also includes a route type field, a length field, a route type-specific key field, and a route type-specific non-key field. See Figure 6D for a schematic diagram of the NLRI field.

[0255] The Route Type field indicates the route type. If the Route Type field is the sixth value in the preset set, it indicates that the BGP protocol message is a Type 3 computing route. For example, if the values ​​in the preset set are values ​​not used by EVPN routes, then a Type 3 computing route occupies a value in the preset set (denoted as the sixth value). For instance, if the Route Type field is 17, it indicates that the BGP protocol message is a Type 3 computing route. The Route Type field can be 1 byte long.

[0256] The Length field represents the length of the second type of computing power route. The length of the Length field can be 1 byte or 2 bytes.

[0257] The Route Type Specific Key field represents a route type-specific key field, and this field carries the first CS-ID. For example, the Route Type Specific Key field can include an RD field, a CS-ID field, and a CIS-ID field. The RD field carries the route identifier and can be 8 bytes long. The CS-ID field carries the CS-ID, specifically the first CS-ID mentioned above, and can be 4 bytes long. The CIS-ID field carries the CIS-ID and can also be 4 bytes long.

[0258] The Route Type Specific Non-Key Fields represent non-key fields specific to the route type. These fields are used to carry the computing power value for each computing power service node corresponding to the first CS-ID. For example, as shown in Figure 6E, the Route Type Specific Non-Key Fields may include Metric TLV and Raw Metric TLV. The Metric TLV carries the coarse-grained computing power value, while the Raw Metric TLV carries the fine-grained metric value.

[0259] As can be seen from the above technical solutions, in this embodiment, messages can be sent to the computing power service node with the optimal computing power value, making full use of the computing resources of each computing power service node and improving the utilization efficiency of computing resources. Through distributed on-demand subscription computing power routing, the announcement of computing power values ​​can be efficiently achieved. Expanding to new address clusters or EVPN address clusters to announce computing power values ​​on demand, with the computing power values ​​defined in NLRI, avoids the inefficiency of packaging computing power values.

[0260] Based on the same application concept as the above method, this application proposes a computing power routing announcement device, applied to a computing power routing ingress node. The computing power routing ingress node already contains a mapping table, which includes the correspondence between service features and CS-IDs. The mapping table is generated by a first type of computing power routing sent by a computing power routing egress node. Referring to Figure 7A, the device may include:

[0261] The receiving module 711 is used to receive a first service message, the first service message including a first service feature;

[0262] The sending module 712 is configured to send a second type of computing power route to the computing power routing exit node if there is no flow table matching the first service feature and there is a first mapping table entry matching the first service feature in the mapping table. The second type of computing power route includes the first CS-ID corresponding to the first service feature.

[0263] The receiving module 711 is further configured to receive a third type of computing power route sent by the computing power route exit node, the third type of computing power route including the computing power value of each of a plurality of computing power service nodes, the plurality of computing power service nodes corresponding to the same first CS-ID;

[0264] The processing module 713 is used to select a target computing power service node based on the computing power value of each computing power service node, and to establish a flow table that matches the first service feature, wherein the flow table includes path information.

[0265] The sending module 712 is further configured to send the first service message to the target computing power service node through the path information.

[0266] In one example, the first type of computing power routing also includes the node location of the computing power routing exit node, and the mapping table also includes the node location; when the sending module 712 sends the second type of computing power routing to the computing power routing exit node, it is specifically used to: obtain the first node location corresponding to the first service feature from the first mapping table entry; and send the second type of computing power routing to the computing power routing exit node through the first node location.

[0267] In one example, the computing power routing entry node also contains a metric update table; when the sending module 712 sends the second type of computing power route to the computing power routing exit node, it is specifically used to: query the metric update table based on the first CS-ID; if the metric update table does not contain computing power metric values ​​of multiple computing power service nodes corresponding to the first CS-ID, then send the second type of computing power route to the computing power routing exit node.

[0268] In one example, the computing power routing entry node also contains a metric update table; when the receiving module 711 receives the third type of computing power route sent by the computing power routing exit node, it is specifically used to: receive multiple third type of computing power routes sent by the computing power routing exit node; wherein, when the computing power metric value of any computing power service node corresponding to the first CS-ID changes, the computing power routing exit node sends a third type of computing power route to the computing power routing entry node;

[0269] The processing module 713 is further configured to, after the receiving module receives each third type of computing power route, record the correspondence between the first CS-ID and the computing power value of each of the plurality of computing power service nodes in the metric update table, and record the update time of each computing power value.

[0270] In one example, the receiving module 711 is further configured to receive a second service message, the second service message including a second service feature; the processing module 713 is further configured to, if no flow table matching the second service feature exists, and the mapping table contains a first mapping table entry matching the second service feature, query the metric update table based on the first CS-ID; if the metric update table contains the computing power values ​​of multiple computing power service nodes corresponding to the first CS-ID, and the difference between the current time and the update time is less than a threshold, select a target computing power service node based on the computing power value of each computing power service node, and establish a flow table matching the second service feature;

[0271] The sending module 712 is further configured to send a second type of computing power route to the computing power route exit node if the computing power value of the multiple computing power service nodes corresponding to the first CS-ID does not exist in the metric update table, or if the computing power value of the multiple computing power service nodes corresponding to the first CS-ID exists in the metric update table, and the difference between the current time and the update time is not less than a threshold.

[0272] In one example, the sending module 712 is further configured to send a cancellation route to the computing power routing egress node if all flow tables corresponding to the first CS-ID are aged, wherein the cancellation route includes the first CS-ID, so that after receiving the cancellation route, the computing power routing egress node stops sending the third type of computing power route for the first CS-ID to the computing power routing ingress node.

[0273] In one example, the first type of computing power routing is forwarded from the computing power routing egress node to the computing power routing ingress node via a route reflector (RR); the second type of computing power routing is forwarded from the computing power routing egress node via the RR; and the third type of computing power routing is forwarded from the computing power routing egress node to the computing power routing ingress node via the RR.

[0274] Based on the same concept as the above method, this application proposes a computing power routing announcement device, applied to a computing power routing exit node, as shown in Figure 7B. The device may include:

[0275] The sending module 721 is used to send a first type of computing power route to the computing power routing entry node. The first type of computing power route includes service features and CS-ID, so that the computing power routing entry node generates a mapping table, the mapping table including the correspondence between the service features and the CS-ID;

[0276] The receiving module 722 is used to receive a second type of computing power route sent by the computing power routing entry node, wherein the second type of computing power route includes a first CS-ID;

[0277] The acquisition module 723 is used to acquire the computing power value of each of the multiple computing power service nodes corresponding to the first CS-ID;

[0278] The sending module 721 is further configured to send a third type of computing power route to the computing power routing entry node. The third type of computing power route includes the computing power value of each computing power service node. The third type of computing power route is used to enable the computing power routing entry node to select a target computing power service node based on the computing power value of each computing power service node.

[0279] In one example, the computing power routing exit node also contains a metric update table; the receiving module 722 is further configured to record the first CS-ID and the computing power metric value of each computing power service node in the metric update table after receiving the second type of computing power route sent by the computing power routing entry node; the sending module 721 is further configured to send a third type of computing power route to the computing power routing entry node when the computing power metric value of any computing power service node corresponding to the first CS-ID changes.

[0280] The receiving module 722 is further configured to, if it receives a route cancellation message sent by the computing power routing entry node, wherein the route cancellation message includes the first CS-ID, delete the first CS-ID and the computing power value corresponding to the first CS-ID from the metric update table; and stop sending the third type of computing power route for the first CS-ID to the computing power routing entry node.

[0281] In one example, the first type of computing power routing is forwarded to the computing power routing ingress node via a route reflector (RR); the second type of computing power routing is forwarded from the computing power routing ingress node to the computing power routing egress node via the RR; and the third type of computing power routing is forwarded to the computing power routing ingress node via the RR.

[0282] Based on the same application concept as the above method, this application proposes an electronic device (such as the above-mentioned computing power routing entry node or computing power routing exit node) in one example. Referring to Figure 8, the electronic device may include a processor 811 and a machine-readable storage medium 812. The machine-readable storage medium 812 stores machine-executable instructions that can be executed by the processor 811. The processor 811 is used to execute the machine-executable instructions to implement the computing power routing announcement method disclosed in the above example of this application.

[0283] In one example, processor 811 may include one or more processing cores, such as a quad-core processor, an octa-core processor, etc. Processor 811 may be implemented using at least one hardware form selected from DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), and PLA (Programmable Logic Array). Processor 811 may also include a main processor and a coprocessor. The main processor, also known as a CPU (Central Processing Unit), is used to process data in the wake-up state; the coprocessor is a low-power processor used to process data in the standby state. In some embodiments, processor 811 may integrate a GPU (Graphics Processing Unit), which is responsible for rendering and drawing the content required to be displayed on the screen.

[0284] In one example, the electronic device may optionally include a peripheral device interface 813 and at least one peripheral device. The processor 811 and the peripheral device interface 813 can be connected via a bus or signal line. Each peripheral device can be connected to the peripheral device interface 813 via a bus, signal line, or circuit board. The peripheral device may include at least one of a radio frequency circuit 814 and a power supply 815.

[0285] The radio frequency (RF) circuit 814 is used to receive and transmit RF (Radio Frequency) signals, also known as electromagnetic signals. The RF circuit 814 communicates with communication networks and other communication devices via electromagnetic signals. The RF circuit 814 converts electrical signals into electromagnetic signals for transmission, or converts received electromagnetic signals back into electrical signals. Optionally, the RF circuit 814 includes: an antenna system, an RF transceiver, one or more amplifiers, a tuner, an oscillator, a digital signal processor, a user identity module card, etc. The RF circuit 814 can communicate with user equipment through at least one wireless communication protocol. This wireless communication protocol includes, but is not limited to: the World Wide Web, metropolitan area networks, intranets, various generations of mobile communication networks (2G, 3G, 4G, and 5G), wireless local area networks, and / or WiFi (Wireless Fidelity) networks.

[0286] The power supply 815 is used to power the various components in the base station. The power supply 815 can be AC ​​power, DC power, a disposable battery or a rechargeable battery.

[0287] Based on the same application concept as the above method, this application embodiment also provides a machine-readable storage medium storing a plurality of computer instructions, which, when executed by a processor, can implement the computing power routing announcement method disclosed in the above examples of this application.

[0288] The aforementioned machine-readable storage medium can be any electronic, magnetic, optical, or other physical storage device that can contain or store information, such as executable instructions, data, etc. For example, machine-readable storage media can be: RAM (Random Access Memory), volatile memory, non-volatile memory, flash memory, storage drives (such as hard disk drives), solid-state drives, any type of storage disk (such as optical discs, DVDs, etc.), or similar storage media, or combinations thereof.

[0289] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, embodiments of this application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0290] The above description is merely an embodiment of this application and is not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for announcing computing power routing, characterized in that, This is applied to the computing power routing ingress node, where a mapping table already exists. This mapping table includes the correspondence between service characteristics and CS-IDs. The mapping table is generated by the first type of computing power routing sent by the computing power routing egress node, and includes: Receive a first service message, the first service message including a first service feature; If no flow table matches the first service feature, and the mapping table contains a first mapping table entry that matches the first service feature, then a second type of computing power route is sent to the computing power routing exit node. The second type of computing power route includes the first CS-ID corresponding to the first service feature. Receive a third type of computing power route sent by the computing power routing exit node, wherein the third type of computing power route includes the computing power value of each of a plurality of computing power service nodes, and the plurality of computing power service nodes correspond to the same first CS-ID; Based on the computing power value of each computing power service node, a target computing power service node is selected, and a flow table matching the first service feature is established, wherein the flow table includes path information. Using the path information, the first service message is sent to the target computing power service node.

2. The method according to claim 1, characterized in that, The first type of computing power routing also includes the node location of the computing power routing exit node, and the mapping table also includes the node location; Sending the second type of computing power route to the computing power routing exit node includes: Obtain the position of the first node corresponding to the first service feature from the first mapping table entry; The second type of computing power route is sent to the computing power routing exit node based on the location of the first node.

3. The method according to claim 1, characterized in that, The computing power routing entry node also contains a metric update table; Sending the second type of computing power route to the computing power routing exit node includes: Query the metric update table based on the first CS-ID; If the metric update table does not contain the computing power values ​​of multiple computing power service nodes corresponding to the first CS-ID, then the second type of computing power route is sent to the computing power routing exit node.

4. The method according to claim 1, characterized in that, The computing power routing entry node also contains a metric update table; The receipt of the third type of computing power route sent by the computing power routing exit node includes: Receive multiple third-type computing power routes sent by the computing power routing exit node; wherein, when the computing power value of any computing power service node corresponding to the first CS-ID changes, the computing power routing exit node sends a third-type computing power route to the computing power routing entry node; After receiving each third-type computing power route, the method further includes: The metric update table records the correspondence between the first CS-ID and the computing power metric value of each of the plurality of computing power service nodes, and records the update time of each computing power metric value.

5. The method according to claim 4, characterized in that, The method further includes: Receive a second service message, the second service message including a second service feature; If no flow table matches the second service feature, and the mapping table contains a first mapping table entry that matches the second service feature, then the metric update table is queried based on the first CS-ID. If the measurement update table contains the computing power values ​​of multiple computing power service nodes corresponding to the first CS-ID, and the difference between the current time and the update time is less than a threshold, then a target computing power service node is selected based on the computing power value of each computing power service node, and a flow table matching the second service feature is established. If the measurement update table does not contain the computing power values ​​of the multiple computing power service nodes corresponding to the first CS-ID, or if the measurement update table contains the computing power values ​​of the multiple computing power service nodes corresponding to the first CS-ID, and the difference between the current time and the update time is not less than a threshold, then a second type of computing power route is sent to the computing power route exit node.

6. The method according to claim 4, characterized in that, The method further includes: If all flow tables corresponding to the first CS-ID are aged, a cancellation route is sent to the computing power routing egress node. The cancellation route includes the first CS-ID, so that after the computing power routing egress node receives the cancellation route, it stops sending the third type of computing power route for the first CS-ID to the computing power routing ingress node.

7. The method according to claim 1, characterized in that, The first type of computing power routing involves the computing power routing egress node forwarding data to the computing power routing ingress node via a route reflector (RR). The second type of computing power routing involves forwarding data to the computing power routing exit node via the RR. The third type of computing power routing is forwarding from the computing power routing exit node to the computing power routing entry node through the RR.

8. The method according to any one of claims 1-7, characterized in that, BGP protocol messages include an NLRI field, which includes an address family information field, a route type field, and a route type specific key field; wherein, if the address family information field is a first designated identifier, it indicates that the BGP protocol message is a BGP protocol message for computing power routing; If the route type field is a first value, it indicates that the BGP protocol message is a first type of computing power route, and the route type specific key field carries the first CS-ID; wherein, the BGP protocol message also includes a path attribute field related to the NLRI field, and the path attribute field carries the service characteristics corresponding to the first CS-ID and the node location of the computing power route exit node; If the route type field is the second value, it indicates that the BGP protocol message is a second type of computing power route, and the route type specific key field carries the first CS-ID; If the route type field is a third value, it indicates that the BGP protocol message is a third type of computing power route, and the route type specific key field carries the first CS-ID; wherein, if the BGP protocol message is a third type of computing power route, the NLRI field also includes a route type specific non-key domain field, and the route type specific non-key domain field carries the computing power value of each computing power service node corresponding to the first CS-ID.

9. The method according to any one of claims 1-7, characterized in that, BGP protocol messages include an NLRI field, which includes an address family information field, a route type field, and a route type specific key field; wherein, if the address family information field is a second specified identifier, it indicates that the BGP protocol message is a BGP protocol message for EVPN routes; If the route type field is the fourth value in the preset set, it indicates that the BGP protocol message is a first type of computing power route, and the route type specific key field carries the first CS-ID; wherein, the BGP protocol message also includes a path attribute field related to the NLRI field, and the path attribute field carries the service characteristics corresponding to the first CS-ID and the node location of the computing power route exit node; If the route type field is the fifth value in the preset set, it indicates that the BGP protocol message is a second type of computing power route, and the route type specific key carries the first CS-ID; If the route type field is the sixth value in the preset set, it indicates that the BGP protocol message is a third type of computing power route, and the route type specific key field carries the first CS-ID; wherein, if the BGP protocol message is a third type of computing power route, the NLRI field also includes a route type specific non-key domain field, and the route type specific non-key domain field carries the computing power value of each computing power service node corresponding to the first CS-ID. The values ​​in the preset set are those not used by EVPN routing.

10. A method for announcing computing power routing, characterized in that, The method, applied to the computing power routing egress node, includes: A first type of computing power route is sent to the computing power routing entry node. The first type of computing power route includes service characteristics and CS-ID, so that the computing power routing entry node generates a mapping table, which includes the correspondence between the service characteristics and the CS-ID. Receive a second type of computing power route sent by the computing power routing entry node, wherein the second type of computing power route includes a first CS-ID; Obtain the computing power value of each of the multiple computing power service nodes corresponding to the first CS-ID; A third type of computing power route is sent to the computing power routing entry node. The third type of computing power route includes the computing power value of each computing power service node. The third type of computing power route is used to enable the computing power routing entry node to select a target computing power service node based on the computing power value of each computing power service node.

11. The method according to claim 10, characterized in that, The computing power routing exit node also contains a metric update table; After receiving the second type of computing power route sent by the computing power routing entry node, the method further includes: The first CS-ID and the computing power value of each computing power service node are recorded in the metric update table; When the computing power value of any computing power service node corresponding to the first CS-ID changes, a third type of computing power route is sent to the computing power routing entry node.

12. The method according to claim 11, characterized in that, After recording the first CS-ID and the computing power value of each computing power service node in the metric update table, the method further includes: If a route cancellation message is received from the computing power routing entry node, and the cancelled route includes the first CS-ID, then the first CS-ID and the computing power value corresponding to the first CS-ID are deleted from the metric update table; and the sending of the third type of computing power route for the first CS-ID to the computing power routing entry node is stopped.

13. The method according to claim 10, characterized in that, The first type of computing power routing is forwarded to the computing power routing entry node through the route reflector (RR); The second type of computing power routing involves the computing power routing ingress node forwarding data to the computing power routing egress node via the RR. The third type of computing power routing is forwarding to the computing power routing entry node via RR.

14. The method according to any one of claims 10-13, characterized in that, BGP protocol messages include an NLRI field, which includes an address family information field, a route type field, and a route type specific key field; wherein, if the address family information field is a first designated identifier, it indicates that the BGP protocol message is a BGP protocol message for computing power routing; If the route type field is a first value, it indicates that the BGP protocol message is a first type of computing power route, and the route type specific key field carries the first CS-ID; wherein, the BGP protocol message also includes a path attribute field related to the NLRI field, and the path attribute field carries the service characteristics corresponding to the first CS-ID and the node location of the computing power route exit node; If the route type field is the second value, it indicates that the BGP protocol message is a second type of computing power route, and the route type specific key field carries the first CS-ID; If the route type field is a third value, it indicates that the BGP protocol message is a third type of computing power route, and the route type specific key field carries the first CS-ID; wherein, if the BGP protocol message is a third type of computing power route, the NLRI field also includes a route type specific non-key domain field, and the route type specific non-key domain field carries the computing power value of each computing power service node corresponding to the first CS-ID.

15. The method according to any one of claims 10-13, characterized in that, BGP protocol messages include an NLRI field, which includes an address family information field, a route type field, and a route type specific key field; wherein, if the address family information field is a second specified identifier, it indicates that the BGP protocol message is a BGP protocol message for EVPN routes; If the route type field is the fourth value in the preset set, it indicates that the BGP protocol message is a first type of computing power route, and the route type specific key field carries the first CS-ID; wherein, the BGP protocol message also includes a path attribute field related to the NLRI field, and the path attribute field carries the service characteristics corresponding to the first CS-ID and the node location of the computing power route exit node; If the route type field is the fifth value in the preset set, it indicates that the BGP protocol message is a second type of computing power route, and the route type specific key carries the first CS-ID; If the route type field is the sixth value in the preset set, it indicates that the BGP protocol message is a third type of computing power route, and the route type specific key field carries the first CS-ID; wherein, if the BGP protocol message is a third type of computing power route, the NLRI field also includes a route type specific non-key domain field, and the route type specific non-key domain field carries the computing power value of each computing power service node corresponding to the first CS-ID. The values ​​in the preset set are those not used by EVPN routing.

16. A computing power routing announcement device, characterized in that, An apparatus applied to a computing power routing ingress node, wherein a mapping table already exists within the computing power routing ingress node, the mapping table including the correspondence between service characteristics and CS-IDs, the mapping table being generated by a first type of computing power routing sent from a computing power routing egress node, the apparatus comprising: A receiving module is configured to receive a first service message, wherein the first service message includes a first service feature; The sending module is configured to send a second type of computing power route to the computing power routing exit node if there is no flow table matching the first service feature and there is a first mapping table entry matching the first service feature in the mapping table. The second type of computing power route includes the first CS-ID corresponding to the first service feature. The receiving module is further configured to receive a third type of computing power route sent by the computing power routing exit node, the third type of computing power route including the computing power value of each of a plurality of computing power service nodes, the plurality of computing power service nodes corresponding to the same first CS-ID; The processing module is used to select a target computing power service node based on the computing power value of each computing power service node, and to establish a flow table that matches the first service feature, wherein the flow table includes path information; The sending module is further configured to send the first service message to the target computing power service node using the path information.

17. The apparatus according to claim 16, characterized in that, The first type of computing power routing also includes the node location of the computing power routing exit node, and the mapping table also includes the node location; When the sending module sends the second type of computing power route to the computing power routing exit node, it is specifically used for: Obtain the position of the first node corresponding to the first service feature from the first mapping table entry; The second type of computing power route is sent to the computing power routing exit node based on the location of the first node.

18. The apparatus according to claim 16, characterized in that, The computing power routing entry node also contains a metric update table; When the sending module sends the second type of computing power route to the computing power routing exit node, it is specifically used for: Query the metric update table based on the first CS-ID; If the metric update table does not contain the computing power values ​​of multiple computing power service nodes corresponding to the first CS-ID, then the second type of computing power route is sent to the computing power routing exit node.

19. The apparatus according to claim 16, characterized in that, The computing power routing entry node also contains a metric update table; When the receiving module receives the third type of computing power route sent by the computing power routing exit node, it is specifically used to: receive multiple third type of computing power routes sent by the computing power routing exit node; wherein, when the computing power value of any computing power service node corresponding to the first CS-ID changes, the computing power routing exit node sends a third type of computing power route to the computing power routing entry node. The processing module is further configured to, after the receiving module receives each third type of computing power route, record the correspondence between the first CS-ID and the computing power value of each of the plurality of computing power service nodes in the metric update table, and record the update time of each computing power value.

20. The apparatus according to claim 19, characterized in that, The receiving module is further configured to receive a second service message, the second service message including a second service feature; The processing module is further configured to query the metric update table based on the first CS-ID if there is no flow table matching the second service feature and the mapping table contains a first mapping table entry matching the second service feature. If the metric update table contains the computing power values ​​of multiple computing power service nodes corresponding to the first CS-ID, and the current time is different from the update... If the time difference is less than the threshold, then the target computing power service node is selected based on the computing power value of each computing power service node, and a flow table matching the second service feature is established. The sending module is further configured to send a second type of computing power route to the computing power route exit node if the computing power value of the multiple computing power service nodes corresponding to the first CS-ID does not exist in the metric update table, or if the computing power value of the multiple computing power service nodes corresponding to the first CS-ID exists in the metric update table, and the difference between the current time and the update time is not less than a threshold.

21. The apparatus according to claim 19, characterized in that, The sending module is further configured to send a cancellation route to the computing power routing egress node if all flow tables corresponding to the first CS-ID are aged, wherein the cancellation route includes the first CS-ID, so that after the computing power routing egress node receives the cancellation route, it stops sending the third type of computing power route for the first CS-ID to the computing power routing ingress node.

22. A computing power routing notification device, characterized in that, The device, applied to a computing power routing egress node, includes: The sending module is used to send a first type of computing power route to the computing power routing entry node. The first type of computing power route includes service features and CS-ID, so that the computing power routing entry node generates a mapping table, the mapping table including the correspondence between the service features and the CS-ID; The receiving module is configured to receive a second type of computing power route sent by the computing power routing entry node, wherein the second type of computing power route includes a first CS-ID; The acquisition module is used to acquire the computing power value of each of the multiple computing power service nodes corresponding to the first CS-ID; The sending module is further configured to send a third type of computing power route to the computing power routing entry node. The third type of computing power route includes the computing power value of each computing power service node. The third type of computing power route is used to enable the computing power routing entry node to select a target computing power service node based on the computing power value of each computing power service node.

23. The apparatus according to claim 22, characterized in that, The computing power routing exit node also contains a metric update table; The receiving module is further configured to record the first CS-ID and the computing power value of each computing power service node in the metric update table after receiving the second type of computing power route sent by the computing power routing entry node; The sending module is further configured to send a third type of computing power route to the computing power routing entry node when the computing power value of any computing power service node corresponding to the first CS-ID changes.

24. The method according to claim 23, characterized in that, The receiving module is further configured to, if it receives a route cancellation message sent by the computing power routing entry node, wherein the route cancellation message includes the first CS-ID, delete the first CS-ID and the computing power value corresponding to the first CS-ID from the metric update table. Stop sending the third type of computing power route for the first CS-ID to the computing power routing entry node.

25. A computing power routing entry node, characterized in that, include: A processor and a machine-readable storage medium, the machine-readable storage medium storing machine-executable instructions that can be executed by the processor; The processor is configured to execute machine-executable instructions to implement the method of any one of claims 1-9.

26. A computing power routing exit node, characterized in that, include: A processor and a machine-readable storage medium, the machine-readable storage medium storing machine-executable instructions that can be executed by the processor; The processor is configured to execute machine-executable instructions to implement the method of any one of claims 10-15.

Images