Computing power distribution method and system, nonvolatile storage medium and electronic equipment

Abstract

The invention discloses a computing power distribution method and system, a nonvolatile storage medium and electronic equipment. The method comprises the steps that a first router obtains network state information, the first router is an edge router connected with a client, the network state information comprises network performance data between the first router and a second router, the second router is an edge router connected with a computing power server, and the edge router is connected with the computing power server; the network performance comprises at least one of the following: packet loss rate, time delay and time delay jitter; receiving computing power resource state data sent by the second router; determining a computing power routing table according to the network state information and the computing power resource state data; and forwarding a computing power request message sent by the client to the target computing power server according to the computing power routing table. The technical problems of unbalanced computing power resource allocation and low utilization rate caused by low computing power resource scheduling efficiency and message lagging of a traditional computing power allocation technology are solved.

Description

Technical Field

[0001] The present application relates to the field of computing power networks, and more particularly, to a computing power allocation method, system, non-volatile storage medium, and electronic device. Background Art

[0002] In today's increasingly complex and diverse computing power networks, the importance of balanced computing power allocation has become even more prominent. Traditional computing power allocation technologies are difficult to cope with multiple challenges such as the surge in computing tasks, uneven loads on computing power nodes, and message latency in distributed network architectures, resulting in increased management complexity, unsuitability for large-scale deployment, insufficient scalability, and the need to use complex technologies such as cross-domain VPNs for interconnection, which gradually complicates business deployment.

[0003] In response to the above problems, no effective solutions have been proposed yet. Summary of the Invention

[0004] Embodiments of the present application provide a computing power allocation method, system, non-volatile storage medium, and electronic device to at least solve the technical problems of unbalanced computing power resource allocation and low utilization rate caused by low computing power resource scheduling efficiency and message latency of traditional computing power allocation technologies.

[0005] According to one aspect of the embodiments of the present application, a computing power allocation method is provided, including: a first router obtains network status information, where the first router is an edge router connected to a client, and the network status information includes network performance data between the first router and a second router, and the second router is an edge router connected to a computing power server, and the network performance includes at least one of the following: packet loss rate, latency, latency jitter; receives computing power resource status data sent by the second router, where the computing power resource status data includes the remaining computing power resource size of the computing power server connected to the second router; determines a computing power routing table based on the network status information and the computing power resource status data, where the computing power routing table includes a computing power service type identifier and the IPv4 address of the second router corresponding to the computing power service type identifier; forwards a computing power request message sent by the client to a target computing power server based on the computing power routing table, where the computing power request message includes a computing power service type identifier.

[0006] Optionally, determining the computing power routing table based on the computing power resource status data and the network status information includes: using the computing power resource status data and the network status information as evaluation metrics; determining the comparison order of the evaluation metrics corresponding to the computing power service type identifier; comparing the values of the evaluation metrics of each second router according to the evaluation metric comparison order until there are different metrics with unequal values; determining the second router with a superior different metric according to a preset rule as the second router corresponding to the computing power service type identifier; and writing the IPv4 address of the second router corresponding to the computing power service type identifier into the computing power routing table.

[0007] Optionally, sending the computing power request message to the target computing power server according to the computing power routing table includes: searching for the IPv4 address of the target second router in the computing power routing table, where the target second router is the second router corresponding to the computing power service type identifier of the computing power request message; searching for the target segment routing identifier corresponding to the IPv4 address of the target second router in the address mapping table, where the target segment routing identifier records the IPv6 address and the forwarding action of the target second router; encapsulating the computing power request message according to the target segment routing identifier; and sending the encapsulated computing power request message to the target computing power server through an intermediate node, where the intermediate node is used to forward the computing power request message to the target second router according to the IPv6 address in the target segment routing identifier, and the target second router is used to decapsulate the encapsulated computing power request message into the IPv4 format according to the forwarding action and send the decapsulated computing power request message to the target computing power server according to the computing power service type identifier.

[0008] Optionally, after sending the encapsulated computing power request message to the target computing power server through the intermediate node, the method further includes: receiving the first response message of the computing power request message sent by the target second router through the intermediate node, where the first response message is the second response message encapsulated by the target second router according to the segment routing identifier of the first router, and the second response message is the response message sent by the target computing power server to the target second router; decapsulating the first response message into the IPv4 format; and sending the first response message decapsulated into the IPv4 format to the client.

[0009] Optionally, obtaining the network status information includes: sending an Internet Control Message Protocol message to the second router according to a first preset period; receiving the response message of the second router to the Internet Control Message Protocol message; and obtaining the network status information according to the response message.

[0010] Optionally, before the first router obtains the network status information, the method further includes: configuring the neighbor relationship with the second router according to the Border Gateway Protocol.

[0011] Optionally, after determining the computing power routing table based on the network status information and the computing power resource status data, the method further includes: receiving the computing power resource status data sent again by the second router according to a second preset period; updating the computing power routing table based on the computing power resource status data sent again and the network status information.

[0012] According to another aspect of the embodiments of the present application, there is also provided a computing power allocation system, including: a first router, a second router, a client, an intermediate node, and a computing power server, wherein: the first router is configured to obtain network status information, wherein the first router is an edge router connected to the client, and the network status information includes network performance data between the first router and the second router, and the second router is an edge router connected to the computing power server, and the network performance includes at least one of the following: packet loss rate, latency, latency jitter; receiving the computing power resource status data sent by the second router, wherein the computing power resource status data includes the remaining computing power resource size of the computing power server connected to the second router; determining a computing power routing table based on the network status information and the computing power resource status data, wherein the computing power routing table includes a computing power service type identifier and the IPv4 address of the second router corresponding to the computing power service type identifier; forwarding the computing power request message sent by the client to the target computing power server according to the computing power routing table, wherein the computing power request message includes a computing power service type identifier; the second router is configured to collect computing power resource status data; sending the computing power resource status data to the first router according to a second preset period; receiving the Internet Control Message Protocol message sent by the first router according to a first preset period; sending a response message to the Internet Control Message Protocol message to the first router; obtaining the computing power request message sent by the first router through the intermediate node, and sending the computing power request message to the target computing power server according to the computing power service type identifier; the client is configured to send a computing power request message to the first router and receive a response message to the computing power request message; the intermediate node is configured to forward the computing power request message.

[0013] According to another aspect of the embodiments of the present application, there is also provided a non-volatile storage medium, in which a program is stored, and when the program runs, it controls the device where the non-volatile storage medium is located to execute the computing power allocation method.

[0014] According to another aspect of the embodiments of the present application, there is also provided an electronic device, including: a memory and a processor, and the processor is configured to run the program stored in the memory, and when the program runs, it executes the computing power allocation method.

[0015] According to another aspect of the embodiments of the present application, there is also provided a computer program product, including a computer program, and when the computer program is executed by a processor, it implements the computing power allocation method.

[0016] In an embodiment of the present application, a first router is used to obtain network status information. Herein, the first router is an edge router connected to a client, and the network status information includes network performance data between the first router and a second router. The second router is an edge router connected to a computing power server, and the network performance includes at least one of the following: packet loss rate, latency, and latency jitter. Receive computing power resource status data sent by the second router, where the computing power resource status data includes the remaining computing power resource size of the computing power server connected to the second router. Determine a computing power routing table based on the network status information and the computing power resource status data, where the computing power routing table includes a computing power service type identifier and the IPv4 address of the second router corresponding to the computing power service type identifier. Forward a computing power request message sent by the client to a target computing power server according to the computing power routing table, where the computing power request message includes a computing power service type identifier. By receiving the computing power resource status data sent by an egress router and the ingress router obtaining network status information in a distributed computing power network, the purpose of dynamically perceiving the network status and real-time computing power notification is achieved, thereby realizing the technical effects of flexible allocation of computing tasks and balanced load of the computing power network, and further solving the technical problems of unbalanced computing power resource allocation and low utilization rate caused by low computing power resource scheduling efficiency and message lag in traditional computing power allocation technologies. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation to the present application. In the drawings:

[0018] Figure 1 is a schematic structural diagram of a computer terminal provided according to an embodiment of the present application;

[0019] Figure 2 is a schematic diagram of a computing power allocation system provided according to an embodiment of the present application;

[0020] Figure 3 is a schematic flowchart of a computing power allocation method provided according to an embodiment of the present application;

[0021] Figure 4 is a schematic flowchart of the generation process of a computing power routing table provided according to an embodiment of the present application;

[0022] Figure 5 is a schematic diagram of the field attributes of a segmented routing identifier provided according to an embodiment of the present application;

[0023] Figure 6 is a schematic flowchart of a computing power allocation method provided according to an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0024] To enable those skilled in the art to better understand the solution of this application, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of this application. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without making creative efforts shall fall within the protection scope of this application.

[0025] It should be noted that the terms "first", "second", etc. in the specification and claims of this application and the above-mentioned drawings are used to distinguish similar objects, and do not necessarily need to be used to describe a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances so that the embodiments of this application described here can be implemented in an order other than those illustrated or described here. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products, or devices.

[0026] To better understand the embodiments of this application, the technical terms involved in the embodiments of this application are explained as follows:

[0027] 1. SRv6 technology

[0028] The SRv6 technology is a segment routing technology based on IPv6. It combines the source routing advantages of segment routing and the rich features of IPv6 to achieve efficient routing control on the IPv6 data plane. Segment routing is a source routing technology. By assigning a Segment ID (SID, segment routing identifier) to each node or link in the network and combining these SIDs into a Segment sequence (Segment List), it guides the data packet to be forwarded according to this sequence. This technology greatly simplifies network deployment, reduces the state maintenance of intermediate nodes, and improves the scalability and flexibility of the network.

[0029] 2. Computing power routing

[0030] In a distributed computing power network, computing power routing refers to the process of dynamically selecting appropriate computing power resources (such as servers, computing clusters, etc.) and network resources (such as bandwidth, etc.) according to the requirements of computing tasks for task allocation and execution. The core of the SRv6 computing power routing method lies in using the SRH in the IPv6 extension header to carry segmented transmission information. The SRH contains a list of SIDs represented by IPv6 addresses, and these SIDs represent specific paths or network functions in the network. When a data packet enters the SRv6 network, the network node will update the destination address of the data packet segment by segment according to the SID list in the SRH and forward it along the path specified by the SID. This way of forwarding segment by segment not only improves the flexibility of routing but also simplifies network configuration and management. In the SRv6 computing power routing method, each SID can be assigned different functions and parameters, such as routing information (Locator), network function (Function), and parameters (Arguments).

[0031] In current research on computing power network routing and computing power allocation methods, most consider routing decisions under traditional IPv4 and MPLS networks, mainly considering the resource usage of computing power servers for computing power allocation. When the computing power state of the server changes, a new computing power server that can provide better computing services for users in the current situation is selected. However, with the expansion of the network scale and the update of services, due to the fact that such traditional network architectures rely on complex routing control and addressing methods, the management complexity increases, it is not suitable for large-scale deployment, has insufficient scalability, and requires the use of complex technologies such as cross-domain VPNs to interconnect, resulting in problems such as complex business deployment becoming prominent.

[0032] To solve the above problems, relevant solutions are provided in the embodiments of this application, which are described in detail below.

[0033] According to the embodiments of this application, a method embodiment of a computing power allocation method is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order than here.

[0034] The method embodiments provided by the embodiments of this application can be executed on a mobile terminal, a computer terminal, or a similar computing device. Figure 1 A hardware structure block diagram of a computer terminal for implementing the computing power allocation method is shown. As Figure 1As shown, the computer terminal 10 may include one or more processors 102 (shown as 102a, 102b, ……, 102n in the figure) (the processor 102 may include, but is not limited to, a processing device such as a microprocessor MCU or a programmable logic device FPGA), a memory 104 for storing data, and a transmission device 106 for communication functions. In addition, it may further include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of the BUS bus), a network interface, a power supply, and / or a camera. Those of ordinary skill in the art can understand that Figure 1 the structure shown is only schematic and does not limit the structure of the above electronic device. For example, the computer terminal 10 may further include more or fewer components than Figure 1 shown in, or have a different configuration from Figure 1 that shown.

[0035] It should be noted that the above one or more processors 102 and / or other data processing circuits are generally referred to as "data processing circuits" herein. The data processing circuit may be embodied in whole or in part as software, hardware, firmware, or any combination thereof. In addition, the data processing circuit may be a single independent processing module, or be incorporated in whole or in part into any one of the other elements in the computer terminal 10. As involved in the embodiments of the present application, the data processing circuit is a kind of processor control (such as the selection of a variable resistor terminal path connected to an interface).

[0036] The memory 104 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the computing power allocation method in the embodiments of the present application. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, that is, implements the above-mentioned computing power allocation method. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memories, or other non-volatile solid-state memories. In some instances, the memory 104 may further include a memory remotely provided with respect to the processor 102, and these remote memories may be connected to the computer terminal 10 through a network. Examples of the above network include, but are not limited to, the Internet, an enterprise intranet, a local area network, a mobile communication network, and combinations thereof.

[0037] The transmission device 106 is used to receive or send data via a network. Specific examples of the above-mentioned network may include a wireless network provided by the communication provider of the computer terminal 10. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 can be a Radio Frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0038] The display can be, for example, a touch-screen liquid crystal display (LCD), which enables the user to interact with the user interface of the computer terminal 10.

[0039] The embodiments of the present application provide a computing power allocation system. Figure 2 A schematic diagram of a computing power allocation system is shown, as Figure 2 shown, the computing power allocation system includes: a first router (i.e., Figure 2 the ingress node in), a second router (i.e., Figure 2 the egress node in), a client (i.e., Figure 2 the client node in), an intermediate node, and a computing power server (i.e., Figure 2The computing service nodes), where: The first router is used to obtain network status information. The first router is an edge router connected to the client. The network status information includes network performance data between the first router and the second router. The second router is an edge router connected to the computing power server. The network performance includes at least one of the following: packet loss rate, latency, latency jitter; Receive the computing power resource status data sent by the second router. The computing power resource status data includes the remaining computing power resource size of the computing power server connected to the second router; Determine the computing power routing table based on the network status information and the computing power resource status data. The computing power routing table includes a computing power service type identifier and the IPv4 address of the second router corresponding to the computing power service type identifier; Forward the computing power request message sent by the client to the target computing power server according to the computing power routing table. The computing power request message includes a computing power service type identifier; The second router is used to collect computing power resource status data; Send the computing power resource status data to the first router according to the second preset period; Receive the Internet Control Message Protocol message sent by the first router according to the first preset period; Send a response message to the Internet Control Message Protocol message to the first router; Obtain the computing power request message sent by the first router through the intermediate node, and send the computing power request message to the target computing power server according to the computing power service type identifier; The client is used to send a computing power request message to the first router and receive a response message to the computing power request message; The intermediate node is used to forward the computing power request message.

[0040] As an optional implementation manner, the ingress node and the egress node are configured with an SRv6 Locator and a VRF (Virtual Routing Forwarding) table is established, so that service requests from the client can pass through the ingress node and be forwarded to the egress node based on the SRv6 technology. The intermediate node device supports IPv6, and all nodes are configured with the IS-IS (Intermediate System to Intermediate System) protocol. Among them, the ingress node and the egress node advertise routes to each other through IS-IS and publish SRv6 Locator information. The Locator information can help the ingress node in the network locate the egress node corresponding to the SRV6 VPN SID. The backbone network between the ingress and the egress uses the IS-IS protocol to publish the routes between all nodes.

[0041] SRv6 computing power routing provides strong support for the efficient operation and resource optimization of computing power networks with its unique flexibility and intelligence. By embedding segment routing information in network data packets, SRv6 can achieve precise control and dynamic adjustment of the computing power resource path, effectively avoiding congestion in computing power nodes and ensuring the smooth transmission of computing tasks. At the same time, SRv6 computing power routing can also intelligently allocate and schedule computing power resources according to network status and the requirements of computing tasks, improve resource utilization, reduce unnecessary resource waste, and can be seamlessly integrated with existing IPv6 networks, support a gradual evolution network upgrade strategy, and have the potential for large-scale deployment. Therefore, in the process of building an efficient and reliable computing power network, SRv6 computing power routing can better realize the intelligent and dynamic management of computing power networks. The SRv6 technology itself has strong flexibility and programmability, enabling the network to be flexibly programmed and configured according to business requirements. At the same time, SRv6 simplifies the control plane by extending IGP (Interior Gateway Protocol) / BGP (Border Gateway Protocol) and removing MPLS (Multi-Protocol Label Switching) tunneling technologies such as LDP (Label Distribution Protocol) and RSVP-TE (Resource Reservation Protocol-Traffic Engineering). On the data plane, SRv6 directly uses IPv6 addresses as forwarding labels, realizing the unified bearer of the control plane and the data plane. This not only reduces the complexity of network protocols, but also reduces the configuration workload of network devices and the operation and maintenance costs. And SRv6 can smoothly evolve on the existing IPv6 network, greatly reducing the time and cost of upgrading and being more convenient for actual deployment. Computing power routing can achieve intelligent scheduling and on-demand allocation of computing power resources, avoiding resource waste and idleness. By combining with SRv6 technology, resource utilization can be further optimized, and the efficiency and reliability of the network can be improved.

[0042] Optionally, use the method of statically adding routes to add the IPv4 routes of the client and the server to each other, so that the packets sent to each other can be correctly sent to the ingress and egress nodes.

[0043] Under the above operating environment, the embodiment of the present application provides a computing power allocation method, as Figure 3 shown, the method includes the following steps:

[0044] Step S302: The first router obtains network status information. Here, the first router is an edge router connected to the client, and the network status information includes network performance data between the first router and the second router. The second router is an edge router connected to the computing power server. The network performance includes at least one of the following: packet loss rate, latency, and latency jitter.

[0045] In the embodiment provided in step S302, obtaining network status information includes: sending Internet Control Message Protocol (ICMP) packets to the second router according to a first preset period; receiving the response packets of the second router to the ICMP packets; and obtaining network status information based on the response packets.

[0046] Optionally, the first router actively obtains the network communication status and generates a corresponding network status table. In the prior art, network resource awareness is mainly based on relevant protocols in TCP / IP, such as ICMP, TCP, and UDP. In addition, there are also methods based on the SNMP protocol. Compared with other methods, ICMP has the advantages of being lightweight, having good real-time performance, being widely supported, and being simple and easy to use. It can quickly detect network connectivity, latency, and packet loss conditions, with low deployment costs, and is supported by default in almost all network devices and operating systems. Therefore, the method based on ICMP is used to realize network resource awareness.

[0047] As an alternative implementation, the ingress node (i.e., the first router) periodically (according to the first preset period) sends ICMP packets to all egress nodes (i.e., the second routers) to detect the network status between the ingress node and the egress nodes (such as the latency, jitter, and packet loss rate of the backbone network), and uploads the relevant network information to the database to form local network status information.

[0048] As an alternative implementation, before the first router obtains network status information, the method further includes: configuring the neighbor relationship with the second router according to the Border Gateway Protocol.

[0049] Optionally, a BGP neighbor relationship and BGP VRF routes are configured between the ingress node (i.e., the first router) and the egress node (i.e., the second router). The IPv6 routing tables on the ingress node and the egress node generate a local SID table according to the BGP-related configurations. There is a route in the table that defines the SID actions of the corresponding node. The ingress and egress nodes unseal the SRv6 packets of the forward and return trips into the original IPv4 packets according to this route and then send them to the corresponding client and server. After the neighbor relationship is established, the RD and RT attributes carried in the BGP Update message match the corresponding neighbor. At this time, the VPNv4 routes corresponding to each egress node are generated in the VRF table of the ingress node, that is, the SRV6 VPN SID, which has the attributes of a specified path and forwarding behavior. The VPNv4 routes corresponding to each ingress node are generated in the VRF table of the egress node. The ingress node and the egress node support encapsulating a task sent to the peer with an SRH header (Segment Routing Header, an extended packet header for segment routing) containing the SID information of the corresponding destination node and forwarding it in the form of the SRv6 technology. The encapsulated task with SRv6 attributes is sent to the corresponding destination node along the shortest path according to the inter-node routes published by IS-IS.

[0050] Step S304, receive the computing power resource status data sent by the second router, where the computing power resource status data includes the remaining computing power resource size of the computing power server connected to the second router.

[0051] As an optional implementation, the computing power resource awareness technology provides support for task scheduling and resource allocation by collecting the resource status of computing nodes in real time. Common computing power resource awareness methods include resource collection based on system monitoring tools and frameworks, dynamic awareness of distributed scheduling systems, resource advertisement of network protocol extensions (such as BGP and SRv6), lightweight monitoring of proxy node deployment, intelligent optimization of machine learning prediction, and resource interfaces of operating systems or virtualization platforms. The BGP extension technology has the advantages of seamless integration with the existing network, adaptation to large-scale distributed environments, support for computing-network collaborative optimization, strong real-time performance, and strong cross-domain deployment ability in computing power resource awareness. It realizes the dynamic propagation and optimization of computing power status by embedding computing power information into the extended attributes of BGP, and at the same time combines routing decisions to complete computing-network collaboration. Compared with other technologies, the BGP extension has low deployment costs and high flexibility. Therefore, the method based on the BGP protocol extension is adopted to realize computing power resource awareness.

[0052] Optionally, in the distributed computing power network scenario, the computing power information of each computing service node (i.e., the computing power server) is collected by the edge router (i.e., the second router) directly connected to it and accessing the core network. Each edge router has an independent database for storing and managing the real-time computing power status information of local computing service nodes. Subsequently, every computing power announcement period (the second preset period), each egress node will announce its local computing power information to the ingress node through BGP update messages. After summarization, the ingress node (i.e., the first router) obtains the computing power status information of the entire network. At the same time, each ingress node also maintains an independent database for storing and managing network status information and the computing power status information of the entire network.

[0053] Through network resource awareness and computing power resource awareness, the ingress node can obtain local network status information and the computing power resource status data of the entire network for subsequent computing power routing calculations. In addition, the embodiments of this application adopt a distributed method. The computing power information of each computing service node (including the remaining computing power resource size of the CPU, the remaining computing power resource size of the memory, etc.) and the network status information between each ingress node and egress node are collected, managed, and announced separately, omitting the intermediate step of sending information to the controller compared with the centralized method, improving the real-time nature of the information.

[0054] Step S306: Determine the computing power routing table based on the network status information and the computing power resource status data. The computing power routing table includes a computing power service type identifier and the IPv4 address of the second router corresponding to the computing power service type identifier.

[0055] In the embodiment provided in step S306, determining the computing power routing table based on the computing power resource status data and the network status information includes: using the computing power resource status data and the network status information as evaluation metrics; determining the comparison order of the evaluation metrics corresponding to the computing power service type identifier; comparing the values of the evaluation metrics of each second router according to the evaluation metric comparison order until there are different metrics with unequal values; determining the second router with a superior different metric according to the preset rule as the second router corresponding to the computing power service type identifier; writing the IPv4 address of the second router corresponding to the computing power service type identifier into the computing power routing table.

[0056] Optionally, using the network status information and the computing power resource status data, calculate the optimal computing power routing egress node corresponding to each service through a computing power routing algorithm for load balancing to generate the computing power routing table. Figure 4 Shows the generation process of the computing power routing table, as Figure 4 shown, the process includes:

[0057] Step S401, obtain the computing power status table and network status table from the network computing resource collaborative perception module: that is, obtain the stored network status information and computing power resource status data from the database.

[0058] Step S402, perform calculations through the computing power routing method for load balancing.

[0059] Step S403, determine whether the routing table needs to be updated according to the calculation result. If the node recorded in the current computing power routing table is no longer the optimal node, update the computing power routing table; if it is still the optimal node, end the generation process.

[0060] As an optional implementation manner, different evaluation index comparison orders are set according to different service types. For example, for tasks with large computing amounts, the remaining computing power resources are compared first; for tasks with requirements for response speed, the network status is compared first. Determining the second router with better different indexes according to the preset rules includes determining the better index values according to different comparison rules for each index. For example, when comparing computing power resources, the computing power resource index with larger remaining computing power resources is better; when comparing the packet loss rate, the smaller packet loss rate is better.

[0061] Optionally, in order to calculate the optimal egress node for a certain type of service, let A be the current egress node and B be the node to be compared. The specific rules of this algorithm are as follows:

[0062] 1. If the remaining computing power resources of node B are greater than those of node A, immediately update node B as the new egress node.

[0063] 2. If the remaining computing power resources of node B are less than those of node A, the egress node remains node A.

[0064] 3. If the remaining computing power resources of nodes A and B are equal, further compare their network performances. The comparison priority is: packet loss rate > latency > latency jitter.

[0065] 4. Only when the network performance of node B is better will it be updated as the new egress node, and only when the attributes with higher priority are equal will the attributes with lower priority be compared. Write the new egress node into the computing power routing table and update the computing power routing table.

[0066] As an optional implementation manner, after determining the computing power routing table based on the network status information and computing power resource status data, the method further includes: receiving the computing power resource status data sent again by the second router according to the second preset period; updating the computing power routing table based on the computing power resource status data sent again and the network status information.

[0067] Step S308, forward the computing power request message sent by the client to the target computing power server according to the computing power routing table, where the computing power request message includes a computing power service type identifier.

[0068] In the embodiment provided in step S308, sending the computing power request message to the target computing power server according to the computing power routing table includes: looking up the IPv4 address of the target second router in the computing power routing table, where the target second router is the second router corresponding to the computing power service type identifier of the computing power request message; looking up the target segment routing identifier corresponding to the IPv4 address of the target second router in the address mapping table, where the target segment routing identifier records the IPv6 address and forwarding action of the target second router; encapsulating the computing power request message according to the target segment routing identifier; sending the encapsulated computing power request message to the target computing power server through an intermediate node, where the intermediate node is used to forward the computing power request message to the target second router according to the IPv6 address in the target segment routing identifier, and the target second router is used to decapsulate the encapsulated computing power request message into the IPv4 format according to the forwarding action and send the decapsulated computing power request message to the target computing power server according to the computing power service type identifier.

[0069] Optionally, the ingress node calculates the computing power routing table through the network status table and the computing power status table. The computing power routing table reflects the mapping relationship between the service ID (computing power service type identifier) and the IPv4 address of the optimal egress node (i.e., the target second router). An SRv6 VPN SID mapping table is generated at the ingress node, which reflects the IPv4 addresses of all egress nodes and the SRv6 VPN SIDs of these nodes. Combining the computing power routing table and the SRv6 VPN SID mapping table, each service ID can be mapped to the SRv6 VPN SID of the corresponding optimal egress node.

[0070] The client initiates a service request according to the computing service requirement through the serviceID (i.e., the computing power service type identifier, identified in the IPv4 address format), and this service request will first be forwarded to the ingress node to which the client is connected.

[0071] After receiving the computing service request sent by the client, the ingress node finds the SRv6 VPN SID (i.e., the target segment routing identifier) corresponding to the service ID through the mapping of the computing power routing and the SRv6 VPN SID. This SID comes from the egress node connecting to the optimal computing service node. The ingress node encapsulates the IPv4 message sent by the client into the SRv6 form according to this SID and forwards it according to the IPv6 routing table of the ingress node.

[0072] After receiving the message in the SRv6 form, the intermediate node forwards it to the previously mapped egress node according to the route advertised by IS-IS following the longest matching principle.

[0073] After receiving the encapsulated packet, the egress node searches the local SID table generated by the BGP protocol for the SRv6 VPN SID, and according to the forwarding action defined by the SID, removes the encapsulated SRH header, restores it to an IPv4 packet, and then matches the VPN instance according to the SRv6 SID, and searches the VPN instance routing table for forwarding. At this time, the packet is restored to the IPv4 packet form sent by the client, and then the IPv4 packet is forwarded to the computing service node bound to the VPN. After receiving the IPv4 packet, the computing service node forwards it to the corresponding port according to the IPv4 address in the packet (i.e., the computing power service type identifier) to provide computing services.

[0074] Optionally, when creating an SRv6 routing forwarding path, it is necessary to define the SRv6 VPN SID (i.e., the segment routing identifier). In the distributed computing power network scenario, the attributes of the custom 128-bit SRv6 VPN SID field are as Figure 5 shown. In the configured 128-bit SID, the first 32 bits (2001:db8) are used as the common prefix, X:X is used as the 32-bit representing the specific node, and these 64 bits together form the Locator part as the network identifier (and as an IPv6 address). The 16 bits (0100) after the Locator are defined as the specific forwarding action. At this time, the entire SID form is 2001:db8:X:X:100::, which is defined as the Endpoint SID of the PE type, i.e., End.DT4, and the corresponding forwarding action is to decapsulate the IPv6 packet and search the IPv4 VPN instance routing table for forwarding. By customizing the attributes of each field of the SRv6 VPN SID, different edge nodes can be distinguished by the SRv6 VPN SID in the distributed computing power network for subsequent computing power routing forwarding based on SRv6.

[0075] Optionally, after the ingress node and the egress node have configured the local SRv6 SID, the local SRv6 SID is sent to the BGP neighbor node through the BGP protocol, and the ingress node and the egress node can now learn the SRv6 SID of each other. Since the intermediate node only supports IPv6 but not SRv6, it is necessary to use the IS-IS protocol to let the intermediate node know the next forwarding action when receiving the message containing the SRv6 SID. The specific method is to set the IPv6 address of the lo interface (local loopback interface) of each ingress node and egress node to the form of locator (for example, if the locator of the PE1 node is 2001:db8:1:1:: / 64, then set the IPv6 address of the lo interface of PE1 to 2001:db8:1:1:: / 64), and then announce the IPv6 address of the lo interface to each node in the domain through the IS-IS protocol, so that all nodes in the domain can learn the route to each PE node. When the intermediate node receives the SRv6 message, it will forward it based on the IPv6 rules and match the SRv6 SID according to the longest match principle. The locator part of the SRv6 SID will match the IPv6 address of the lo interface of a PE node, and then forward it according to the corresponding route.

[0076] As an optional implementation, after sending the encapsulated computing power request message to the target computing power server through the intermediate node, the method also includes: receiving a first response message to the computing power request message sent by the target second router through the intermediate node, wherein the first response message is a second response message encapsulated by the target second router according to the segment routing identifier of the first router, and the second response message is a response message sent by the target computing power server to the target second router; decapsulating the first response message into IPv4 format; and sending the first response message decapsulated into IPv4 format to the client.

[0077] Optionally, after completing the computing task, the target computing power server generates and sends a response message (i.e., the second response message) to the connected egress router (i.e., the target second router). After receiving the response message, the egress router re - encapsulates the response message based on the SRv6 segment routing identifier (SID) information received from the ingress router before, forming the first response message. This first response message contains a specific SID sequence, indicating how the intermediate nodes should send the message back to the ingress router. Subsequently, the intermediate nodes forward the first response message segment by segment according to the locator information in the SID, based on the IPv6 rules and the routing information advertised by the IS - IS protocol, following the longest matching principle until it reaches the ingress router. After receiving the first response message, the ingress router identifies and performs the corresponding de - encapsulation action according to the local SID table, removes the SRv6 extension header, and restores the message to the original IPv4 format. Finally, the restored IPv4 response message is sent back to the client, completing the closed - loop of task request and response in the SRv6 - based distributed computing power network, significantly improving the efficiency and reliability of computing power scheduling and network communication, solving the problems of low resource scheduling efficiency and message lag in traditional computing power allocation technologies, and achieving dynamic balanced allocation and high utilization of computing power resources.

[0078] Under the above - mentioned operating environment, an embodiment of this application provides a computing power allocation method. According to the current network state and the resource occupancy of computing servers, combined with the computing power routing policy, it finally realizes computing power routing addressing and forwarding, as Figure 6 shown. The method includes the following steps:

[0079] The network - computing resource collaborative perception module actively obtains the network communication status (i.e., network state information) and the resource situation of the computing power server (i.e., computing power resource status), and generates the corresponding network state table and computing power state table.

[0080] The computing power routing calculation module uses the network state table and the computing power state table, and calculates the optimal computing power routing egress node corresponding to each service through the computing power routing algorithm for load balancing, generating a computing power routing table.

[0081] The SRv6 - based computing power routing creation module performs SRv6 - related configurations, establishes the SRv6 forwarding path between the ingress node and the egress node, and constructs the SRv6 computing power routing table based on the computing power routing table and the SRv6 SID mapping table.

[0082] The SRv6 - based computing power routing forwarding module is used to encapsulate the ordinary IPv4 message sent by the client into the SRv6 forwarding form, and then forward it through the SRv6 routing path established by the SRv6 - based computing power routing creation module.

[0083] Specifically, the remaining resource sizes of the CPU and memory of the computing server are used as the computing state, the latency, jitter, and packet loss rate of the backbone network are used as the network state. The edge routers are divided into ingress routers and egress routers, and the functions of BGP protocol, IS-IS protocol, and SRv6 encapsulation and decapsulation are deployed on the edge routers. At the egress router connected to the computing server, the computing power notification function is implemented by extending the Update message in the BGP protocol. At the ingress router connected to the client, the computing power awareness is performed. First, the egress router notifies the computing resource status of the server connected to it to the ingress router through the Update message in the BGP protocol. The ingress router generates a computing power routing table based on the computing power situation notified by the egress router and the network state collected by itself, and combines the current routing policy to decide to allocate the task to a specific computing service node. The implementation method is that the ingress router queries the computing power routing table generated by the ingress router according to the result of the finally selected service node, maps the task from the egress router IP to the SRv6 SID corresponding to the egress router, and the ingress router performs encapsulation according to the mapping result to match the egress router connected to the finally selected server. The intermediate path is determined according to the shortest path algorithm in the IS-IS protocol. After reaching the corresponding egress router, the egress router decapsulates the encapsulated task into the original form according to the SRv6 protocol and allocates it to the adjacent server. Finally, while meeting the computing task requirements, the resource utilization rate and overall performance of the distributed computing power network are improved by combining the SRv6 technology. There is still room for further adjustment in the embodiments of the present application, specifically including adopting a better routing policy on the ingress router and adding parameters to the computing server as the computing power state to provide to the ingress router to complete decision optimization, etc.

[0084] Through the above steps, in the distributed computing power network scenario, for different service requirements, combined with the real-time network state and the computing power state of the computing service nodes, the service requirements can be reasonably forwarded to different computing service nodes for processing, realizing the intelligent collaborative scheduling of computing resources and network resources, ensuring the effective execution of computing tasks and improving the network resource utilization rate and user comprehensive satisfaction.

[0085] In the distributed computing power network based on SRv6, when a computing task request arrives, the ingress router will start the computing power routing function and allocate the computing task to a server with higher task execution efficiency according to the current computing power state and network state, combined with the adopted routing policy. This not only avoids the possible resource tension situation of the servers in the computing power network, enables the computing tasks to be executed efficiently, but also reduces the possibility of network congestion, improves the utilization rate of network resources and computing resources, and improves the overall performance of the network, solving the problems such as complex configuration and difficult deployment existing in traditional networks such as IPv4 and MPLS. Specifically, the present application has the following advantages:

[0086] 1. A distributed state acquisition method is adopted. The computing power information of each computing service node and the network state information between each ingress node and egress node are collected, managed, and announced separately. Compared with the centralized method, the intermediate step of sending information to the controller is omitted, improving the real-time performance of the information.

[0087] Through the distributed acquisition and management strategy, the edge router and the ingress node independently collect and store the computing power and network state information, avoiding the dependence on the controller in the traditional centralized method, improving the real-time performance and response speed, and adapting to the dynamic computing power network requirements. Innovatively combining the double-layer perception mechanism of computing power and network state, the edge router announces the computing power information through BGP update messages, and the ingress node obtains the network state through ICMP detection to achieve dynamic perception and provide accurate data support. The ingress node optimizes resource scheduling by fusing the network and computing power state information in the database. Compared with the traditional separate management method, the scheduling accuracy and performance are improved. Finally, through the extension of the BGP protocol and ICMP detection, the complexity of the dedicated perception protocol is avoided, ensuring compatibility with the existing network and providing an efficient, flexible, and easy-to-deploy resource perception and scheduling solution.

[0088] 2. The existing SRv6 technology mainly focuses on the implementation of network functions. In the computing power network scenario, there is a lack of clear definition and support for how to combine computing tasks and computing power resources. The method embodiment of this application flexibly configures the Locator field and uses the IS-IS protocol to achieve compatibility of non-SRv6 devices, ensuring that data packets can still be efficiently forwarded according to IPv6 rules in intermediate nodes that do not support SRv6. At the same time, combined with the network state table, it supports dynamic adjustment of the path by the Locator to meet the stability and adaptability of computing power routing in a dynamic network environment. This method not only makes up for the definition gap of the existing SRv6 technology in the computing power network, but also improves the resource scheduling efficiency, enhances the actual deployment ability of the SRv6 technology, and provides strong technical support for the efficient operation of the distributed computing power network.

[0089] The existing SRv6 technology has problems of inflexible mapping between tasks and paths and insufficient encapsulation adaptability in computing power routing and forwarding, making it difficult to meet the requirements of dynamic task allocation in computing power networks. The computing power routing and forwarding based on SRv6 realizes the efficient matching of task types and network paths through a three-level mapping mechanism of "service ID - egress node IPv4 address - SRv6 VPN SID", improving the scheduling ability of computing power resources. At the same time, by expanding the SID action field, the encapsulation and decapsulation processes of tasks are optimized, enabling tasks to be encapsulated as SRv6 packets at the ingress node and restored to the original IPv4 packets at the egress node, thus achieving seamless compatibility between SRv6 and the IPv4 environment. This design improves the flexibility and efficiency of SRv6 computing power routing, provides strong technical support for dynamic computing power networks, can distinguish each PE node with SRv6 SID in a distributed computing power network, and supports SRv6 computing power routing and forwarding even when the intermediate nodes do not support SRv6, realizing the full network connectivity of SRv6-based computing power routing.

[0090] An embodiment of this application provides a non-volatile storage medium. A program is stored in the non-volatile storage medium. When the program runs, it controls the device where the non-volatile storage medium is located to execute the following computing power allocation method: The first router obtains network status information, where the first router is an edge router connected to the client, and the network status information includes network performance data between the first router and the second router. The second router is an edge router connected to the computing power server, and the network performance includes at least one of the following: packet loss rate, latency, latency jitter; receives the computing power resource status data sent by the second router, where the computing power resource status data includes the remaining computing power resource size of the computing power server connected to the second router; determines a computing power routing table based on the network status information and the computing power resource status data, where the computing power routing table includes a computing power service type identifier and the IPv4 address of the second router corresponding to the computing power service type identifier; forwards the computing power request packet sent by the client to the target computing power server according to the computing power routing table, where the computing power request packet includes a computing power service type identifier.

[0091] An embodiment of the present application provides an electronic device, including: a memory and a processor, where the processor is configured to run a program stored in the memory. When the program runs, it executes the following computing power allocation method: The first router obtains network status information, where the first router is an edge router connected to the client, and the network status information includes network performance data between the first router and the second router. The second router is an edge router connected to the computing power server, and the network performance includes at least one of the following: packet loss rate, latency, latency jitter; receives the computing power resource status data sent by the second router, where the computing power resource status data includes the remaining computing power resource size of the computing power server connected to the second router; determines a computing power routing table according to the network status information and the computing power resource status data, where the computing power routing table includes a computing power service type identifier and the IPv4 address of the second router corresponding to the computing power service type identifier; forwards the computing power request message sent by the client to the target computing power server according to the computing power routing table, where the computing power request message includes a computing power service type identifier.

[0092] An embodiment of the present application provides a computer program product, including a computer program, which when executed by a processor, implements the following computing power allocation method: The first router obtains network status information, where the first router is an edge router connected to the client, and the network status information includes network performance data between the first router and the second router. The second router is an edge router connected to the computing power server, and the network performance includes at least one of the following: packet loss rate, latency, latency jitter; receives the computing power resource status data sent by the second router, where the computing power resource status data includes the remaining computing power resource size of the computing power server connected to the second router; determines a computing power routing table according to the network status information and the computing power resource status data, where the computing power routing table includes a computing power service type identifier and the IPv4 address of the second router corresponding to the computing power service type identifier; forwards the computing power request message sent by the client to the target computing power server according to the computing power routing table, where the computing power request message includes a computing power service type identifier.

[0093] In the above embodiments of the present application, the descriptions of the respective embodiments have their own emphases. For parts not detailed in a certain embodiment, reference may be made to the relevant descriptions of other embodiments.

[0094] In several embodiments provided in the present application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are merely illustrative. For example, the division of the units can be a logical function division. In actual implementation, there can be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the couplings or direct couplings or communication connections shown or discussed with each other can be through some interfaces. The indirect couplings or communication connections of units or modules can be in electrical or other forms.

[0095] The units described as separate components may or may not be physically separated. The components shown as units may or may not be physical units, that is, they can be located in one place or distributed to multiple units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0096] In addition, in each embodiment of the present application, the functional units can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above-mentioned integrated units can be implemented in the form of hardware or in the form of software functional units.

[0097] If the above-mentioned integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, in essence, or the part that contributes to the related technology, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in each embodiment of the present application. The foregoing storage medium includes: various media such as USB flash drives, read-only memories (ROMs), random access memories (RAMs), mobile hard disks, magnetic disks, or optical discs that can store program codes.

[0098] The above are only the preferred embodiments of the present application. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present application, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of the present application.

Claims

1. A computing power allocation method, characterized in that: include: The first router acquires network status information, wherein the first router is an edge router connected to the client, the network status information includes network performance data between the first router and a second router, the second router is an edge router connected to a computing power server, and the network performance includes at least one of the following: packet loss rate, delay, and delay jitter; Receiving computing power resource status data sent by the second router, wherein the computing power resource status data includes the remaining computing power resource size of the computing power server connected to the second router; Determine a computing power routing table according to the network status information and the computing power resource status data, wherein the computing power routing table includes a computing power service type identifier and an IPv4 address of a second router corresponding to the computing power service type identifier; The computing power request message sent by the client is forwarded to the target computing power server according to the computing power routing table, wherein the computing power request message includes a computing power service type identifier.

2. The computing power allocation method according to claim 1, characterized in that: Determining a computing power routing table according to the computing power resource status data and the network status information includes: Using the computing power resource status data and the network status information as evaluation indicators; Determine a comparison order of evaluation indicators corresponding to the computing power service type identifier; According to the evaluation index comparison order, the values ​​of the evaluation indexes of the second routers are compared until there are different indexes with unequal values; Determine, according to a preset rule, a second router having a better different indicator as the second router corresponding to the computing power service type identifier; The IPv4 address of the second router corresponding to the computing power service type identifier is written into the computing power routing table.

3. The computing power allocation method according to claim 1, characterized in that: Sending the computing power request message to the target computing power server according to the computing power routing table includes: Searching for the IPv4 address of the target second router in the computing power routing table, wherein the target second router is the second router corresponding to the computing power service type identifier of the computing power request message; Searching for a target segment routing identifier corresponding to the IPv4 address of the target second router in the address mapping table, wherein the target segment routing identifier records the IPv6 address and forwarding action of the target second router; Encapsulating the computing power request message according to the target segment routing identifier; The encapsulated computing power request message is sent to the target computing power server through an intermediate node, wherein the intermediate node is used to forward the computing power request message to the target second router according to the IPv6 address in the target segment routing identifier, and the target second router is used to decapsulate the encapsulated computing power request message into IPv4 format according to the forwarding action, and send the decapsulated computing power request message to the target computing power server according to the computing power service type identifier.

4. The computing power allocation method according to claim 3, characterized in that: After sending the encapsulated computing power request message to the target computing power server through the intermediate node, the method further includes: Receive a first response message to the computing power request message sent by the target second router through the intermediate node, wherein the first response message is a second response message encapsulated by the target second router according to the segment routing identifier of the first router, and the second response message is a response message sent by the target computing power server to the target second router; Decapsulating the first response message into IPv4 format; The first response message decapsulated into IPv4 format is sent to the client.

5. The computing power allocation method according to claim 1, characterized in that: Obtaining network status information includes: Sending an Internet Control Message Protocol message to the second router according to a first preset period; receiving a response message from the second router to the Internet Control Message Protocol message; The network status information is obtained according to the response message.

6. The computing power allocation method according to claim 1, characterized in that: Before the first router acquires the network status information, the method further includes: A neighbor relationship with the second router is configured according to the Border Gateway Protocol.

7. The computing power allocation method according to claim 1, characterized in that: After determining the computing power routing table according to the network status information and the computing power resource status data, the method further includes: Receiving computing resource status data sent again by the second router according to a second preset period; The computing power routing table is updated based on the computing power resource status data and the network status information sent again.

8. A computing power allocation system, characterized in that: It includes a first router, a second router, a client, an intermediate node and a computing server, wherein: The first router is used to obtain network status information, wherein the first router is an edge router connected to the client, the network status information includes network performance data between the first router and the second router, the second router is an edge router connected to the computing power server, and the network performance includes at least one of the following: packet loss rate, delay, and delay jitter; receiving computing power resource status data sent by the second router, wherein the computing power resource status data includes the remaining computing power resource size of the computing power server connected to the second router; determining a computing power routing table based on the network status information and the computing power resource status data, wherein the computing power routing table includes a computing power service type identifier and an IPv4 address of the second router corresponding to the computing power service type identifier; forwarding a computing power request message sent by the client to a target computing power server based on the computing power routing table, wherein the computing power request message includes a computing power service type identifier; The second router is used to collect the computing power resource status data; send the computing power resource status data to the first router according to the second preset period; receive the Internet Control Message Protocol message sent by the first router according to the first preset period; send a response message to the Internet Control Message Protocol message to the first router; obtain the computing power request message sent by the first router through the intermediate node, and send the computing power request message to the target computing power server according to the computing power service type identifier; The client is configured to send the computing power request message to the first router, and receive a response message to the computing power request message; The intermediate node is used to forward the computing power request message.

9. A non-volatile storage medium, characterized in that: The non-volatile storage medium stores a program, wherein when the program is running, the device where the non-volatile storage medium is located is controlled to execute the computing power allocation method described in any one of claims 1 to 7.

10. An electronic device, characterized in that: include: A memory and a processor, wherein the processor is used to run a program stored in the memory, wherein the computing power allocation method described in any one of claims 1 to 7 is executed when the program is run.

11. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the computing power allocation method according to any one of claims 1 to 7 is implemented.

Images