Computing power resource network topology generation method and device of intelligent computing center

Abstract

The invention provides a computing power resource network topology generation method and device of an intelligent computing center, and relates to the technical field of intelligent computing centers, intelligent computing centers and computing power infrastructures. The method is applied to a computing power resource network, the computing power resource network comprises a first node and a plurality of second nodes, and the method comprises the following steps: step S1, in a target monitoring time period, based on the first node, receiving a plurality of pieces of original information reported by the plurality of second nodes respectively, the original information comprising node information and link information; and S2, in the target monitoring time period, based on the first node, topology calculation is carried out on the multiple pieces of original information, a network topological graph is generated, and the network scheduling operation of the computing power resource network is completed based on the network topological graph. According to the invention, the efficiency of obtaining the topological graph can be improved, the accuracy of the obtained topological graph can be improved, and the efficiency and effect of network scheduling operation based on the topological graph can be improved.

Description

Technical Field

[0001] The present invention relates to the technical fields of intelligent computing centers, intelligent computing centers, and computing power infrastructure technologies, and in particular, to a method and device for generating a network topology of computing power resources in an intelligent computing center. Background Art

[0002] With the rapid development of artificial intelligence technology, "intelligent computing centers" and "intelligent computing centers" have emerged as the times require.

[0003] An "intelligent computing center" refers to a facility that uses large-scale heterogeneous computing power resources, including general computing power and intelligent computing power, and mainly provides the required computing power, data, and algorithms for artificial intelligence applications (such as scenarios such as artificial intelligence deep learning model development, model training, and model inference). The intelligent computing center covers facilities, hardware, and software, and can provide full-stack capabilities from underlying computing power to top-level application enabling.

[0004] The "intelligent computing center" includes, but is not limited to, the "intelligent computing center".

[0005] An "intelligent computing center", that is, an artificial intelligence computing center, is a type of computing power infrastructure that provides computing power services, data services, and algorithm services required for artificial intelligence applications based on artificial intelligence theory and using an artificial intelligence computing architecture.

[0006] "Computing power" is the core of "intelligent computing centers" and "intelligent computing centers", and is the ability of computer nodes or computing / data centers to process information. It is the ability of computer hardware and software to cooperate to jointly execute a certain computing requirement. It is the computing ability to output a target result by processing information data. It is a new type of productive force that integrates information computing power, network carrying capacity, and data storage capacity, and mainly provides services to society through computing power infrastructure.

[0007] Currently, the existing technology is to manually collect node information (also known as host information) and link information in the computing power resource network of the intelligent computing center, and manually summarize it into a structured topology diagram. This manual operation method is greatly affected by human factors, not only making the acquisition efficiency of the topology diagram very low, but also resulting in very low accuracy of the obtained topology diagram, and seriously restricting the efficiency and effect of network scheduling operations (such as computing power scheduling, routing scheduling, disaster recovery, etc.) based on the topology diagram. Summary of the Invention

[0008] The purpose of the present invention is to provide a method and device for obtaining information based on the computing power of an intelligent computing center, which is used to solve the technical problems existing in the network topology perception of the computing power resource network in the prior art, such as very low efficiency in obtaining the topology diagram, very low accuracy of the obtained topology diagram, and serious restrictions on the efficiency and effect of network scheduling operations based on the topology diagram.

[0009] To solve the above technical problems, the present invention is implemented as follows:

[0010] In a first aspect, the present invention provides a method for generating a computing power resource network topology of an intelligent computing center, which is applied to a computing power resource network. The computing power resource network includes a first node and multiple second nodes. The method includes:

[0011] Step S1: During a target monitoring period, based on the first node, receive multiple pieces of original information respectively reported by the multiple second nodes. Among them, the multiple second nodes and the multiple pieces of original information are in one-to-one correspondence. The original information includes: node information of the corresponding second node, node information of the adjacent node of the corresponding second node, and link information of the link connected by the corresponding second node;

[0012] Step S2: During the target monitoring period, based on the first node, perform topology calculation on the multiple pieces of original information to generate a network topology diagram. The network scheduling operation of the computing power resource network is completed based on the network topology diagram. One piece of node information corresponds to one node in the network topology diagram, and one piece of link information corresponds to one edge in the network topology diagram.

[0013] In one embodiment, before the step S1, the method further includes:

[0014] Step S01: Based on a first target node, receive a link detection message broadcast by a second target node through a target link; where the first target node is one of the multiple second nodes, the second target node is an adjacent node of the first target node, and the target link is a link connecting the first target node and the second target node;

[0015] Step S02: Based on the first target node, parse the link detection message to obtain second node information of the second target node and link information of the target link;

[0016] Step S03: Based on the first target node, generate original information of the first target node according to the first node information of the first target node, the second node information, and the link information of the target link.

[0017] In one embodiment, the step S2 includes:

[0018] Step S21: Based on the first node, convert the multiple pieces of original information into multiple pieces of standard information respectively. The standard information is information in a standard data format, and the standard data format is a data format for forming an adjacency list;

[0019] Step S22: Based on the first node, splice the multiple standard information to obtain a topology information table, where the topology information table is the adjacency list, and the topology information table is used to represent the network topology relationship of all nodes included in the computing power resource network;

[0020] Step S23: Based on the first node, construct the network topology graph according to the topology information table.

[0021] In one embodiment, after the step S2, the method further includes:

[0022] Step S3: Based on the first node, receive the change information reported by the third node, where the third node is a node preparing to access the computing power resource network, or the third node is a node preparing to exit the computing power resource network, or the third node is the second node that needs to perform original information change indicated by the change information;

[0023] Step S4: Based on the first node, re - perform topology calculation according to the change information and the network topology graph to generate an updated network topology graph.

[0024] In one embodiment, the target monitoring period is the period indicated by one monitoring period among multiple consecutive monitoring periods. Among the multiple consecutive monitoring periods, the duration of the periods indicated by any two different monitoring periods is a preset threshold.

[0025] In one embodiment, the first node is:

[0026] Among all nodes included in the computing power resource network, the node with the minimum total time consumption for receiving the original information reported by other nodes; or,

[0027] Among all nodes included in the computing power resource network, the node with the minimum packet loss rate for receiving the original information reported by other nodes; or,

[0028] Among all nodes included in the computing power resource network, the node with the minimum total time consumption for receiving the original information reported by other nodes and the minimum packet loss rate for receiving the original information reported by other nodes.

[0029] In a second aspect, the present invention further provides a computing power resource network topology generation device for an intelligent computing center, which is applied to a computing power resource network. The computing power resource network includes a first node and multiple second nodes. The device includes:

[0030] An information receiving module, configured to receive, based on the first node, multiple pieces of original information respectively reported by the multiple second nodes during a target monitoring period, where the multiple second nodes and the multiple pieces of original information are in one-to-one correspondence, and the original information includes: node information of the corresponding second node, node information of the adjacent nodes of the corresponding second node, and link information of the link connected by the corresponding second node;

[0031] A topology calculation module, configured to perform topology calculation on the multiple pieces of original information based on the first node during the target monitoring period to generate a network topology diagram, and network scheduling operations of the computing power resource network are completed based on the network topology diagram. One piece of node information corresponds to one node in the network topology diagram, and one piece of link information corresponds to one edge in the network topology diagram.

[0032] In a third aspect, the present invention further provides a server, including: a processor, a memory, and a program stored on the memory and executable on the processor, where when the program is executed by the processor, the steps in the method for generating the topology of the computing power resource network of the intelligent computing center as described in the first aspect above are implemented.

[0033] In a fourth aspect, the present invention further provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps in the method for generating the topology of the computing power resource network of the intelligent computing center as described in the first aspect above are implemented.

[0034] In a fifth aspect, the present invention provides a computer program product, including computer instructions, where when the computer instructions are executed by a processor, the steps in the method for generating the topology of the computing power resource network of the intelligent computing center as described in the first aspect above are implemented.

[0035] In the present invention, by receiving multiple pieces of original information respectively reported by multiple second nodes in the computing power resource network, and performing topology calculation based on the node information and link information included in the multiple pieces of original information, a network topology diagram for supporting network scheduling operations of the computing power resource network is generated, so as to replace the manual operation method adopted in the prior art in an automated manner, avoid interference of human factors, improve the acquisition efficiency of the topology diagram, improve the accuracy of the obtained topology diagram, and improve the efficiency and effect of network scheduling operations based on the topology diagram. Description of the Drawings

[0036] By reading the detailed description of the preferred embodiments below, various other advantages and benefits will become clear to those of ordinary skill in the art. The drawings are only for the purpose of showing the preferred embodiments and are not considered to be a limitation of the present invention. Moreover, throughout the drawings, the same reference numerals are used to represent the same components. In the drawings:

[0037] Figure 1 It is a schematic flowchart of a method for generating a computing power resource network topology of an intelligent computing center provided by the present invention;

[0038] Figure 2 It is a schematic diagram of the process of constructing a topology diagram of a computing power resource network of an intelligent computing center provided by the present invention;

[0039] Figure 3 It is a schematic structural diagram of a device for generating a computing power resource network topology of an intelligent computing center provided by the present invention;

[0040] Figure 4 It is a schematic structural diagram of an electronic device provided by the present invention. Detailed implementation manners

[0041] Next, the technical solutions in the present invention will be clearly and completely described in conjunction with the accompanying drawings in the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present invention without creative efforts shall fall within the protection scope of the present invention.

[0042] First, the technical terms related to the present invention will be briefly described below.

[0043] The "computing power" described in the present invention is the ability of a computer device or a computing / data center to process information, which is the ability of computer hardware and software to cooperate to jointly execute a certain computing requirement. It is the computing ability to process information data and output a target result, and is a new type of productive force integrating information computing power, network carrying capacity, and data storage capacity, and mainly provides services to society through computing power infrastructure.

[0044] The "computational power" (Computational Power, CP) described in the present invention is an ability of a data center server to process data and output a result, and is a comprehensive index for measuring the computing ability of a data center, including general computing ability, supercomputing ability, and intelligent computing ability. The commonly used measurement unit is the number of floating-point operations per second (FLOPS: Floating Point Operations Per Second, 1EFLOPS = 10^18 FLOPS). The larger the value, the stronger the comprehensive computing ability. It is estimated that 1EFLOPS is approximately the computing power output of 5 Tianhe 2A or 500,000 mainstream server CPUs or 2 million mainstream laptops. The calculation formula is: CP = CP 通用 +CP 智能 +CP 超级 .

[0045] The "Network Power (NP)" described in the present invention is an indication of the data transmission capacity of computing power facilities, and is a comprehensive ability including network architecture, network bandwidth, transmission delay, intelligent management and scheduling, etc. The network power involves network transmission within and between data centers, and is a comprehensive indicator for measuring network transmission scheduling ability. In the present invention, the network power uses video memory bandwidth.

[0046] The "Storage Power (SP)" described in the present invention is the comprehensive ability of a data center in four aspects: data storage capacity, performance, security and reliability, and green and low-carbon. It is a comprehensive indicator for measuring the data storage capacity of a data center, and includes external storage devices such as storage arrays and server internal storage devices. The common measurement unit for storage capacity is exabyte (EB, 1EB = 2^60 bytes), the common measurement unit for performance is the number of read and write operations per second per unit capacity (Input / Output Operations Per Second / TB, IOPS / TB), and the disaster recovery ratio is an important manifestation of security and reliability.

[0047] The "computing power infrastructure" described in the present invention is a new type of information infrastructure integrating information computing power, network power, and data storage power, and can realize centralized computing, storage, transmission, and application of information.

[0048] The "new type of information infrastructure" described in the present invention refers to: mainly including network infrastructures such as 5G networks, fiber broadband networks, backbone networks, international communication networks, and satellite Internet, computing power infrastructures such as data centers, general computing power centers, intelligent computing centers, and supercomputing centers, and new technology infrastructures such as artificial intelligence, blockchain, and quantum computing.

[0049] The "computing power" described in the present invention includes general computing power, intelligent computing power, and super computing power.

[0050] The "general computing power" described in the present invention is the computing power provided by servers based on central processing unit (CPU) chips, and is used to support basic general computing such as cloud computing and edge computing.

[0051] The "intelligent computing power" described in the present invention is a computing platform that is deployed on a large scale for various artificial intelligence innovation applications based on dedicated chips such as Graphics Processing Unit (GPU), Field Programmable Gate Array (FPGA), and Application Specific Integrated Circuit (ASIC), such as natural language processing, machine vision, etc.

[0052] The "super computing power" described in the present invention mainly refers to the computing power provided by high-performance computing clusters such as supercomputers. It utilizes the centralized computing resources of multiple computer systems working in parallel and processes extremely complex or data-intensive problems through a dedicated operating system. It is mainly used for computing in cutting-edge scientific fields, such as planetary simulation, drug molecule design, gene analysis, etc.

[0053] The "intelligent computing center" described in the present invention refers to a facility that provides the required computing power, data, and algorithms for artificial intelligence applications (such as scenarios like artificial intelligence deep learning model development, model training, and model inference) by using large-scale heterogeneous computing power resources, including general computing power (CPU) and intelligent computing power (GPU, FPGA, etc.). The intelligent computing center covers facilities, hardware, and software, and can provide full-stack capabilities from underlying computing power to top-level application enablement.

[0054] The "intelligent computing center" described in the present invention includes, but is not limited to, the "intelligent computing center".

[0055] The "intelligent computing center" described in the present invention, namely the artificial intelligence computing center, is a type of computing power infrastructure that provides computing power services, data services, and algorithm services required for artificial intelligence applications based on artificial intelligence theory and using an artificial intelligence computing architecture.

[0056] The "computing power center" described in the present invention refers to a facility mainly composed of infrastructure such as wind, fire, water, and electricity and IT software and hardware devices, and having computing power, carrying capacity, and storage capacity, including general data centers, intelligent computing centers, supercomputing centers, etc.

[0057] The "supercomputing center" described in the present invention refers to the supercomputing data center, which is a data center based on supercomputers or large-scale computing clusters and can provide functions such as large-scale computing, storage, and network services, and is widely used in application scenarios such as aerospace, national defense, oil exploration, climate modeling, and genome sequencing.

[0058] The "computing power resources" described in the present invention refer to the technologies and facilities with information computing, transmission, storage, and application capabilities required for the development of the digital society, including but not limited to computing resources such as CPUs and GPUs, network resources such as switches and routers, storage resources such as storage arrays and distributed storage, security resources such as firewalls and intrusion detection systems, and support and guarantee resources such as wind, fire, water, and electricity.

[0059] The "models" and "large models" described in the present invention include but are not limited to "large language models" and "multimodal large models".

[0060] The "large language model" described in the present invention refers to a large language model (LLM), which is a language model with a relatively large number of parameters, aiming to understand and generate human language. It is trained with a large amount of text data and can perform a wide range of tasks including text summarization, translation, sentiment analysis, etc.

[0061] The "multimodal large models" described in the present invention refer to models that jointly train multimodal information such as text, images, videos, and audio, including but not limited to multimodal large language models.

[0062] The "computing power resource network" described in the present invention refers to a network composed of multiple computing power resources managed by an intelligent computing center. Among them, the computing power resource network includes multiple nodes, and different nodes among the multiple nodes are connected by links. For example, a node can be a server or a network device (such as a switch). Among them, the server is at least used to provide computing power support (such as processing computing power tasks), and the network device is at least used to support data transmission between different servers.

[0063] The "topology of the computing power resource network of the intelligent computing center" described in the present invention refers to: the node information of all nodes constituting the computing power resource network, the link information of all links, and the corresponding relationship between nodes and links; logically, it is manifested as a network topology diagram. Based on the network topology diagram, network scheduling operations of the computing power resource network can be completed, such as computing power scheduling, routing scheduling, disaster recovery, etc.

[0064] The "node information" described in the present invention refers to: information used to describe the capabilities of the corresponding node in terms of computing power support and / or data transmission capabilities, such as: node name, ports included in the node, maximum computing power supported by the node, maximum data transmission bandwidth supported by the node, data transmission protocols supported by the node, IP address of the node, media access control (MAC) address of the node, maximum transmission unit (MTU) of the node, etc.

[0065] The "link information" described in the present invention refers to information used to describe the corresponding link and the two nodes connected by the link. For example: link number / name, data transmission protocol adopted by the link (such as ipv4, ipv6), maximum data transmission bandwidth supported by the link, packet loss rate of the link, latency of the link, jitter of the link, coverage range of the link, node name and port of one of the two nodes connected by the link, and node name and port of the other of the two nodes connected by the link.

[0066] Please refer to Figure 1 , Figure 1 which is a method for generating a computing power resource network topology of an intelligent computing center provided by the present invention, applied to a computing power resource network. The computing power resource network includes a first node and multiple second nodes. As Figure 1 shown, it includes the following steps:

[0067] Step S1: During a target monitoring period, based on the first node, receive multiple pieces of original information respectively reported by the multiple second nodes.

[0068] Among them, the multiple second nodes and the multiple pieces of original information are in one-to-one correspondence. The original information includes: node information of the corresponding second node, node information of the adjacent nodes of the corresponding second node, and link information of the link connected by the corresponding second node.

[0069] It should be understood that the above-mentioned target monitoring period can be a period during the operation of the computing power resource network. For example: the period corresponding to the first startup process of the computing power resource network, the period manually indicated by the user, the period automatically indicated based on a set program (such as a periodic monitoring program, which triggers a topology awareness operation every fixed time to form a network topology map of the computing power resource network), etc.

[0070] The above-mentioned first node can be understood as a control node among all the nodes included in the computing power resource network. This first node can also be called a controller; the first node is used to perform the relevant operations of steps S1 to S2 to obtain a network topology map of the computing power resource network.

[0071] The above-mentioned second node can be understood as: among all the nodes included in the computing power resource network, the nodes other than the first node.

[0072] The adjacent node of the second node mentioned above can be understood as: a node in the computing resource network that is directly connected to the second node through a link. For example, assume that there are five nodes A, B, C, D and E in the computing resource network, where node A is connected to nodes B and C, node B is connected to node D, node C is also connected to node D, and finally node D is connected to node E. Then, in this example, the adjacent nodes of node A are B and C, the adjacent nodes of node B are A and D, the adjacent nodes of node C are A and D, and the adjacent nodes of node D include B, C and E, and finally the adjacent node of node E is D.

[0073] In the application, since the switch records the node information and corresponding link information of each server connected to the switch, when the controller (that is, the first node) collects multiple node information and multiple link information, multiple switches in the computing resource network can report the multiple node information and multiple link information, thereby eliminating the resource overhead caused by the server reporting information.

[0074] That is, the method further includes step S1 ′, receiving, based on the first node, a plurality of to-be-processed information respectively reported by the plurality of switches within the target monitoring period.

[0075] The information to be processed includes: node information of the corresponding switch, node information of the server connected to the corresponding switch, node information of the adjacent node of the server connected to the corresponding switch, and link information of the link to which the corresponding switch is connected.

[0076] It should be understood that the information set formed by the multiple pieces of information to be processed is completely consistent with the information set formed by the multiple pieces of original information.

[0077] Step S2: During the target monitoring period, based on the first node, perform topology calculation on the multiple original information to generate a network topology map.

[0078] The network scheduling operation of the computing resource network is completed based on the network topology graph, one node information corresponds to a node in the network topology graph, and one link information corresponds to an edge in the network topology graph.

[0079] Exemplarily, the network scheduling operations of the computing power resource network may include: load balancing scheduling (making the computing power load of each node in the computing power resource network tend to be consistent or lower than the set load threshold), priority scheduling (preferentially allocating nodes for processing high-priority computing tasks), routing scheduling (dynamically selecting data transmission paths to optimize network resource utilization of the computing power resource network, improve data transmission efficiency and ensure network reliability, and avoid network congestion and bottleneck problems as much as possible), and disaster recovery (such as data backup, system recovery, redundancy mechanisms and emergency response strategies, etc.).

[0080] It should be understood that the first node performs topological calculations on its own node information and multiple pieces of original information to generate a network topology map.

[0081] In the present invention, by receiving multiple pieces of original information respectively reported by multiple second nodes in the computing power resource network, and performing topological calculations based on the node information and link information included in the multiple pieces of original information, a network topology map for supporting network scheduling operations of the computing power resource network is generated, so as to replace the manual operation method adopted in the prior art in an automated manner, avoid interference from human factors, improve the acquisition efficiency of the topology map, enhance the accuracy of the obtained topology map, and enhance the efficiency and effect of network scheduling operations based on the topology map.

[0082] In one embodiment, before the step S1, the method further includes:

[0083] Step S01: Based on a first target node, receive a link detection message broadcast by a second target node through a target link; wherein, the first target node is one of the multiple second nodes, the second target node is an adjacent node of the first target node, and the target link is a link connecting the first target node and the second target node;

[0084] Step S02: Based on the first target node, parse the link detection message to obtain the second node information of the second target node and the link information of the target link;

[0085] Step S03: Based on the first target node, generate the original information of the first target node according to the first node information of the first target node, the second node information, and the link information of the target link.

[0086] Among them, the broadcast operation of the link detection message is completed through the link detection service of the second target node, and the link detection service is used to detect whether the address stored in the node (i.e., the address of the adjacent node of the node) is reachable, so as to judge the link state of the corresponding link (such as whether the link is available and the performance of the link). For example, the link detection service can be a service corresponding to the Link Layer Discovery Protocol (LLDP).

[0087] In this embodiment, for a second node, by receiving the link detection packets broadcast by its adjacent nodes, parsing to obtain the node information of its adjacent nodes and the link information of the links connecting its adjacent nodes, and then combining the node information of the second node itself, the original information of the second node is summarized and formed for subsequent reporting to the first node, thereby completing the announcement of the node information of the second node, the node information of the adjacent nodes of the second node, and the link information of the links connected by the second node.

[0088] It should be noted that before the step S1, the method further includes:

[0089] Step S01': Based on the first target node, broadcast a first link detection packet, so that the second target node generates the original information of the second target node based on the first link detection packet, where the first link detection packet includes: the first node information of the first target node and the link information of the target link.

[0090] The broadcast operation of the first link detection packet is completed through the link detection service of the first target node.

[0091] That is to say, for any one of the foregoing multiple second nodes, the link detection service can be started to broadcast the link detection broadcast packet corresponding to the second node; and receive the link detection broadcast packet broadcast by the adjacent node corresponding to the second node. By means of receiving and sending the packets corresponding to the link detection service, each of the multiple second nodes can record the link information of the links it connects and the node information of the adjacent nodes connected through the links locally at the node.

[0092] In one embodiment, the step S2 includes:

[0093] Step S21: Based on the first node, convert the multiple original information into multiple standard information respectively, where the standard information is information in a standard data format, and the standard data format is a data format for forming an adjacency list;

[0094] Step S22: Based on the first node, splice the multiple standard information to obtain a topology information table, where the topology information table is the adjacency list, and the topology information table is used to represent the network topology relationship of all nodes included in the computing power resource network;

[0095] Step S23: Based on the first node, construct the network topology graph according to the topology information table.

[0096] In the present invention, the format of the adjacency list is as follows:

[0097] Index 1 Index 2 Index 3 … Index N … … … … … … … … … … … … … … …

[0098] In the above style, one line of table data corresponds to one standard information, and the indicators 1-N (N is a positive integer) indicate multiple parameters included in the original information. For example: node name, adjacent node name, link name, maximum computing power of the node, ports of the links connected to the node, etc.

[0099] In this embodiment, the multiple original information are respectively converted into multiple standard information to achieve the format unification of the multiple original information, facilitating the subsequent construction of the topological relationship. Among them, first, a topological information table is constructed based on the multiple standard information, and then a network topology graph is constructed according to the topological information table. During the process of constructing the network topology graph, a visual table (i.e., the topological information table) for representing the topological relationship of the computing power resource network is generated to facilitate the user to maintain the information of the network topology graph.

[0100] Furthermore, after constructing the network topology graph based on the topological information table based on the first node, the network topology graph can be projected onto a virtual map to support the user to more intuitively perceive the states of each node and each link in the computing power resource network through the virtual map. For example: in the virtual map, nodes are represented by node icons (such as circles with a certain area), links between nodes are represented by line segments in the virtual map, the relative position relationship between the node icons in the virtual map is consistent with the actual position relationship between the nodes in the computing power resource network, and the longer the distance between two nodes, the longer the length of the line segment corresponding to the link between the two nodes in the virtual map, and vice versa; and when the computing power load of a node exceeds the set load threshold, the node icon corresponding to the node can be highlighted in the virtual map; similarly, when the data transmission load of a link exceeds the set transmission threshold, the line segment corresponding to the link can be highlighted in the virtual map, so that the user can timely discover the nodes and / or links with too high load through the virtual map and take corresponding load balancing measures to reduce the risk of overload of some nodes or links in the computing power resource network.

[0101] In one embodiment, after the step S2, the method further includes:

[0102] Step S3: Based on the first node, receive the change information reported by the third node, where the third node is a node preparing to access the computing power resource network, or the third node is a node preparing to exit the computing power resource network, or the third node is the second node that needs to perform original information change indicated by the change information;

[0103] Step S4: Based on the first node, re-perform topological calculation according to the change information and the network topology graph to generate an updated network topology graph.

[0104] Based on the above settings, it is supported to dynamically update the topological information of the computing power resource network, so that the network topology map maintains high real-time performance and accuracy, thereby improving the effect of network scheduling operations performed based on the network topology map.

[0105] Furthermore, it can also be set to count the number of third nodes or change information, and only after the counted number exceeds the set number threshold, re-perform topology calculation based on the counted multiple third nodes or multiple table change information and the most recently generated network topology map, and generate an updated network topology map, so as to reduce the processing frequency of the first node in terms of topology calculation, and thereby reduce the overall energy consumption of the first node.

[0106] In one embodiment, the target monitoring period is the period indicated by one monitoring period among multiple consecutive monitoring periods. Among the multiple consecutive monitoring periods, the duration of the periods indicated by any two different monitoring periods is a preset threshold.

[0107] By setting the first node to periodically receive multiple pieces of original information and form a corresponding network topology map accordingly, the periodic update of the network topology map of the computing power resource network is realized, so as to adapt to the continuously changing topological relationship of the computing power resource network in actual applications and ensure the data reliability of the generated network topology map.

[0108] It should be understood that the above preset threshold can be adaptively set according to the actual application requirements of users. For example: one hour, one day, one week, etc. The specific value of the preset threshold of the present invention is not limited.

[0109] In one embodiment, the first node is:

[0110] Among all the nodes included in the computing power resource network, the node with the minimum total time consumption for receiving the original information reported by other nodes; or,

[0111] Among all the nodes included in the computing power resource network, the node with the minimum packet loss rate for receiving the original information reported by other nodes; or,

[0112] Among all the nodes included in the computing power resource network, the node with the minimum total time consumption for receiving the original information reported by other nodes and the minimum packet loss rate for receiving the original information reported by other nodes.

[0113] Among them, the total time consumption for receiving the original information reported by other nodes can be: the sum of the multiple time consumptions for this node to receive the original information reported by multiple other nodes.

[0114] The packet loss rate of the original information reported by other nodes received by a node may be: the average value of multiple sampled packet loss rates corresponding to the node, or, the number of data whose average value of multiple sampled packet loss rates corresponding to the node exceeds the set sampled packet loss rate;

[0115] Among them, multiple sampled packet loss rates corresponding to the node correspond one-to-one with multiple other nodes, and the sampled packet loss rate is used to represent the packet loss rate when the node receives the original information of the corresponding other node during the sampling period.

[0116] In this embodiment, based on the total time consumed by a node to receive the original information reported by other nodes, and / or, the packet loss rate of the original information reported by a node to receive other nodes, the first node is determined among all the nodes included in the computing power resource network, so as to reduce the transmission overhead when the computing power resource network executes the topology information generation method described in the present invention.

[0117] Specifically, reference can be made to Figure 2 , Figure 2 The part of the image numbered 1 can be understood as the aforementioned multiple original information (in the part of the image numbered 1, spine, node1, node2, leaf1, etc. can be understood as the original information corresponding to a node in the computing power resource network), Figure 2 The part of the image numbered 2 can be understood as the aforementioned topology information table, Figure 2 The part of the image numbered 3 can be understood as the aforementioned network topology graph.

[0118] Please refer to Figure 3 , Figure 3 which is a schematic structural diagram of a device for generating a topology of a computing power resource network of an intelligent computing center provided by the present invention. As Figure 3 shown, the device 300 for generating a topology of a computing power resource network of an intelligent computing center includes:

[0119] An information receiving module 301, configured to receive multiple pieces of original information respectively reported by the multiple second nodes based on the first node during the target monitoring period, where the multiple second nodes correspond one-to-one with the multiple pieces of original information, and the original information includes: node information of the corresponding second node, node information of the adjacent node of the corresponding second node, and link information of the link connected by the corresponding second node;

[0120] A topology calculation module 302, configured to perform topology calculation on the multiple pieces of original information based on the first node during the target monitoring period to generate a network topology graph, and the network scheduling operation of the computing power resource network is completed based on the network topology graph. One piece of node information corresponds to one node in the network topology graph, and one piece of link information corresponds to one edge in the network topology graph.

[0121] In one embodiment, the computing power resource network topology generation device 300 of the intelligent computing center further includes:

[0122] A message receiving unit, configured to receive a link detection message broadcast by a second target node through a target link based on a first target node; wherein, the first target node is one of the multiple second nodes, the second target node is an adjacent node of the first target node, and the target link is a link connecting the first target node and the second target node;

[0123] A message parsing unit, configured to parse the link detection message based on the first target node to obtain second node information of the second target node and link information of the target link;

[0124] An information generating unit, configured to generate original information of the first target node based on the first target node according to the first node information of the first target node, the second node information, and the link information of the target link.

[0125] In one embodiment, the topology calculation module 302 includes:

[0126] An information conversion unit, configured to convert the multiple pieces of original information into multiple pieces of standard information respectively based on the first node, wherein the standard information is information in a standard data format, and the standard data format is a data format for forming an adjacency list;

[0127] An information splicing unit, configured to splice the multiple pieces of standard information based on the first node to obtain a topology information table, wherein the topology information table is the adjacency list, and the topology information table is used to represent the network topology relationship of all nodes included in the computing power resource network;

[0128] A topology graph construction unit, configured to construct the network topology graph based on the first node according to the topology information table.

[0129] In one embodiment, the computing power resource network topology generation device 300 of the intelligent computing center further includes:

[0130] A change receiving module, configured to receive change information reported by a third node based on the first node, wherein the third node is a node ready to access the computing power resource network, or the third node is a node ready to exit the computing power resource network, or the third node is the second node that needs to perform original information change indicated by the change information;

[0131] A re-topology module, configured to perform re-topology calculation based on the first node according to the change information and the network topology graph to generate an updated network topology graph.

[0132] In one embodiment, the target monitoring period is the period indicated by one monitoring cycle among a plurality of consecutive monitoring cycles. Among the plurality of consecutive monitoring cycles, the duration of the periods indicated by any two different monitoring cycles is a preset threshold value.

[0133] In one embodiment, the first node is:

[0134] Among all the nodes included in the computing power resource network, the node with the minimum total time consumption for receiving the original information reported by other nodes; or,

[0135] Among all the nodes included in the computing power resource network, the node with the minimum packet loss rate for receiving the original information reported by other nodes; or,

[0136] Among all the nodes included in the computing power resource network, the node with the minimum total time consumption for receiving the original information reported by other nodes and the minimum packet loss rate for receiving the original information reported by other nodes.

[0137] The device for generating the topology of the computing power resource network of the intelligent computing center provided by the present invention can implement each process of the above-mentioned embodiments of the method for generating the topology of the computing power resource network of the intelligent computing center. The technical features correspond one by one and can achieve the same technical effects. To avoid repetition, they will not be elaborated here.

[0138] It should be noted that the device for generating the topology of the computing power resource network of the intelligent computing center in the present invention can be a device, or a component, an integrated circuit, or a chip in an electronic device.

[0139] The present invention also provides an electronic device. Refer to Figure 4 , Figure 4 which is a schematic structural diagram of an electronic device provided by an embodiment of the present invention. The electronic device includes a memory 401, a processor 402, and a program or instruction running on the memory 401. When the program or instruction is executed by the processor 402, it can implement Figure 1 any step in the corresponding embodiment of the method for generating the topology of the computing power resource network of the intelligent computing center and achieve the same beneficial effects, which will not be elaborated here.

[0140] Among them, the processor 402 can be a CPU, an ASIC, an FPGA, or a GPU.

[0141] Those of ordinary skill in the art can understand that all or part of the steps for implementing the above-mentioned embodiments of the method for generating the topology of the computing power resource network of the intelligent computing center can be completed by hardware related to program instructions. The program can be stored in a readable medium.

[0142] The present invention also provides a readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it can implement any of the steps in the foregoing embodiments of the method for generating a computing power resource network topology of the corresponding intelligent computing center, and can achieve the same technical effects. To avoid repetition, details are not described herein again. The storage medium may be, for example, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disc, etc. Figure 1

[0143] Terms such as "first" and "second" in the present invention are used to distinguish similar objects, and do not necessarily describe a specific order or sequence. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or node comprising a series of steps or units does not necessarily limit to those clearly listed steps or units, but may include other steps or units not clearly listed or inherent to these processes, methods, products, or nodes. In addition, the use of "and / or" in the present invention represents at least one of the connected objects. For example, A and / or B and / or C represents seven cases including A alone, B alone, C alone, A and B existing together, B and C existing together, A and C existing together, and A, B, and C existing together.

[0144] It should be noted that in this article, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "comprising an..." does not exclude the existence of additional identical elements in the process, method, article or device comprising the element.

[0145] Through the description of the above embodiments, those skilled in the art can clearly understand that the above embodiment methods can be implemented by means of software plus a necessary general hardware platform. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. The computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disc) and includes several instructions for causing a terminal (which may be a mobile phone, a computer, a server, an air conditioner, or a second terminal node, etc.) to execute the methods of the various embodiments of the present invention.

[0146] The embodiments of the present invention have been described above in conjunction with the accompanying drawings. However, the present invention is not limited to the above specific embodiments. The above specific embodiments are merely illustrative and not restrictive. Under the inspiration of the present invention, those of ordinary skill in the art can also make many forms without departing from the spirit of the present invention and the scope protected by the claims, and all of them fall within the protection scope of the present invention.

Claims

1. A method for generating a computing power resource network topology of an intelligent computing center, characterized in that, Applied to a computing power resource network, the computing power resource network including a first node and a plurality of second nodes, the method comprising: Step S1, within a target monitoring period, based on the first node, receiving a plurality of original information respectively reported by the plurality of second nodes, wherein the plurality of second nodes and the plurality of original information are in one-to-one correspondence, and the original information includes: node information of the corresponding second node, node information of the adjacent nodes of the corresponding second node, and link information of the link connected by the corresponding second node; Step S2, within the target monitoring period, based on the first node, performing topology calculation on the plurality of original information to generate a network topology graph, and the network scheduling operation of the computing power resource network is completed based on the network topology graph, one piece of the node information corresponding to one node in the network topology graph, and one piece of the link information corresponding to one edge in the network topology graph.

2. The method according to claim 1, wherein Before the step S1, the method further comprises: Step S01, based on a first target node, receiving a link detection message broadcast by a second target node through a target link; wherein the first target node is one of the plurality of second nodes, the second target node is an adjacent node of the first target node, and the target link is a link connecting the first target node and the second target node; Step S02, based on the first target node, parsing the link detection message to obtain the second node information of the second target node and the link information of the target link; Step S03, based on the first target node, generating the original information of the first target node according to the first node information of the first target node, the second node information, and the link information of the target link.

3. The method according to claim 1, wherein The step S2 includes: Step S21, based on the first node, respectively converting the plurality of original information into a plurality of standard information, wherein the standard information is information in a standard data format, and the standard data format is a data format for forming an adjacency list; Step S22, based on the first node, splicing the plurality of standard information to obtain a topology information table, wherein the topology information table is the adjacency list, and the topology information table is used to represent the network topology relationship of all nodes included in the computing power resource network; Step S23, based on the first node, constructing the network topology graph according to the topology information table.

4. The method according to claim 1, wherein After the step S2, the method further comprises: Step S3, based on the first node, receiving change information reported by a third node, wherein the third node is a node preparing to access the computing power resource network, or the third node is a node preparing to withdraw from the computing power resource network, or the third node is the second node that needs to perform original information change indicated by the change information; Step S4, based on the first node, re-performing topology calculation according to the change information and the network topology graph to generate an updated network topology graph.

5. The method according to claim 1, characterized in that The target monitoring period is the period indicated by one monitoring period among a plurality of consecutive monitoring periods. Among the plurality of consecutive monitoring periods, the duration of the periods indicated by any two different monitoring periods is a preset threshold.

6. The method according to claim 1, wherein The first node is: Among all the nodes included in the computing power resource network, the node with the minimum total time consumption for receiving the original information reported by other nodes; or, Among all the nodes included in the computing power resource network, the node with the minimum packet loss rate for receiving the original information reported by other nodes; or, Among all the nodes included in the computing power resource network, the node with the minimum total time consumption for receiving the original information reported by other nodes and the minimum packet loss rate for receiving the original information reported by other nodes.

7. An apparatus for generating a computing power resource network topology of an intelligent computing center, characterized in that, Applied to a computing power resource network, the computing power resource network includes a first node and a plurality of second nodes, and the device includes: An information receiving module, configured to receive, based on the first node, a plurality of original information respectively reported by the plurality of second nodes within the target monitoring period, where the plurality of second nodes and the plurality of original information are in one-to-one correspondence, and the original information includes: the node information of the corresponding second node, the node information of the adjacent nodes of the corresponding second node, and the link information of the link connected by the corresponding second node; A topology calculation module, configured to perform topology calculation on the plurality of original information based on the first node within the target monitoring period to generate a network topology diagram, and the network scheduling operation of the computing power resource network is completed based on the network topology diagram. One piece of the node information corresponds to one node in the network topology diagram, and one piece of the link information corresponds to one edge in the network topology diagram.

8. A server, characterized in that, Including: A processor, a memory, and a program stored on the memory and executable on the processor. When the program is executed by the processor, the steps of the method for generating the topology of the computing power resource network of the intelligent computing center as described in any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium, characterized in that, A computer program is stored on the computer-readable storage medium. When the computer program is executed by a processor, the steps of the method for generating the topology of the computing power resource network of the intelligent computing center as described in any one of claims 1 to 6 are implemented.

10. A computer program product, characterized in that, Including computer instructions. When the computer instructions are executed by a processor, the steps of the method for generating the topology of the computing power resource network of the intelligent computing center as described in any one of claims 1 to 6 are implemented.

Images