Communication method, communication device and computer readable storage medium

Abstract

The invention provides a communication method, a communication device and a computer readable storage medium, relates to the field of communication, and can determine a service path according to acquired network data. The communication method comprises: obtaining network data, the network data comprising network topology information and a computing power demand of a service, the network topology information comprising device information of network elements, device information of a computing power center, link information of links between the network elements and the computing power center, and link information of links between the network elements; according to the network data, determining at least one path of the service reaching a first computing power center through at least one network element, the path satisfies one or more of the following conditions: the computing power cost of the service using the first computing power center determined according to the computing power demand of the service satisfies a first condition, and the network cost of at least one link and at least one network element in the service using path determined according to the network topology information satisfies a second condition. The embodiment of the invention is applied to the field of communication.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of communications, and in particular, to a communication method, a communication device, and a computer-readable storage medium. Background Art

[0002] Computing force network (CFN) is a new type of information infrastructure that flexibly schedules and allocates computing resources, storage resources, and network resources between cloud, edge, and end according to demand. Specifically, computing force refers to the computing power of the device, which is usually divided into: high-performance computing power with large computing volume (also known as cluster computing power) and general computing power with relatively small computing volume. Network refers to a digital telecommunications network that allows nodes to share resources, such as an optical transport network (OTN).

[0003] Unlike traditional networks, computing networks bring different types of user needs in different application scenarios. For example, in scenarios involving large amounts of data transmission, it is necessary to ensure the security of data transmission (i.e. lossless data transmission and no packet loss in the network) and provide users with deterministic network security for data transmission; in scenarios involving ultra-long-distance data transmission, it is necessary to ensure that the bandwidth of the transmission link is ultra-large to reduce costs for users.

[0004] Based on the above, this puts higher demands on the planning of computing power networks. Summary of the invention

[0005] The present application provides a communication method, a communication device, and a computer-readable storage medium, which can obtain network data and determine the path of the service based on the network data, thereby achieving better planning for the computing power network.

[0006] In a first aspect, a communication method is provided. The communication method can be executed by a computing power node device (e.g., a computing power center), including: obtaining network data, the network data including network topology information and computing power requirements of a service, wherein the network topology information includes equipment information of network elements, equipment information of the computing power center, link information of links between network elements and the computing power center, and link information of links between network elements; determining at least one path for a service to reach a first computing power center through at least one network element according to the network data, and the path satisfies one or more of the following conditions: the computing power cost of the first computing power center used by the service confirmed according to the computing power requirements of the service satisfies the first condition, and the network cost of at least one network element used by the service confirmed according to the network topology information satisfies the second condition.

[0007] Then, the above scheme can determine the path of the service based on the network data including network topology information and the computing power requirements of the service. Among them, the above scheme can obtain the relevant information of the network elements, computing power centers, links between network elements and computing power centers, and links between network elements (such as equipment information of network elements) included in the network through network topology information, and can obtain the relevant information of the computing power required by the service (such as computing power cost) through the computing power requirements of the service. Usually, the topology of the network is relatively complex, and may include multiple network elements and multiple computing power centers, which makes a certain service often have multiple optional entry points, computing power centers or paths from the entry points to the computing power center. Then, the above scheme can confirm whether the computing power cost of a computing power center used by the service meets the corresponding conditions based on the acquired computing power requirements, or confirm whether the links used by the service and the network costs of the network elements meet the corresponding conditions based on the network topology information, and determine one or more paths that meet the corresponding conditions, thereby determining a better path (for example, the path with the lowest computing power cost) among the multiple optional paths of the service. In one possible implementation, the above scheme can determine the path that meets the computing power requirements of the business and uses the lowest network cost of the links and network elements. Then, by obtaining network data, the above scheme can determine one or more paths of the business based on the computing power cost or network cost and the corresponding conditions (i.e., the first condition and the second condition), thereby effectively reducing the cost of computing power network planning. It is not difficult to understand that the above scheme can also determine the paths of multiple businesses at the same time, thereby effectively improving the efficiency of computing power network planning.

[0008] In a possible implementation manner, the condition satisfied by the path further includes: an extended bandwidth cost of at least one link in the service usage path confirmed according to the network topology information satisfies a third condition.

[0009] Then, the above scheme can confirm that the extended bandwidth cost of the link in the service usage path meets the corresponding conditions based on the network topology information, and determine one or more paths of the service. Generally, when the bandwidth of multiple links in the network cannot meet the computing power requirements of the service, or when the bandwidth of the same link used by multiple services exceeds the available bandwidth of the link, it is necessary to expand the bandwidth of the link. Although the bandwidth can be increased and bandwidth expansion can be achieved through methods such as link aggregation, bandwidth expansion will undoubtedly lead to a significant increase in costs. In this way, the above scheme can determine the links in the service usage path by judging whether the extended bandwidth cost of the link in the service usage path meets the corresponding conditions. For example, the corresponding condition is set to the extended bandwidth cost threshold. When the extended bandwidth cost exceeds the extended bandwidth cost threshold, the link in the service usage path is not feasible, that is, the path is not feasible. It is not difficult to understand that this scheme does not mean that bandwidth expansion must be performed when determining the service path. For example, when the available bandwidth of the link in the network can meet the bandwidth required by the service, there is no need to expand the bandwidth, and the extended bandwidth cost is 0, so the corresponding condition in the above scheme can be set to 0. Then, the above solution can determine the service path based on the cost of expanding bandwidth to achieve computing power network planning and reduce the cost of expanding bandwidth.

[0010] In a possible implementation, the link information includes link delay; the above communication method further includes: when it is determined that the link delay of the shortest path in at least one path exceeds a delay threshold, generating first prompt information.

[0011] Then, the above scheme can determine the delay of the determined path based on the acquired link information including the link delay, and generate corresponding prompts. Usually, the delay of the path for a service to pass through a certain entry point to reach a certain computing power center is relatively fixed, and the shorter the path, the smaller the delay. In one possible implementation, the delay of the shortest path of the determined service is the smallest, but still exceeds the delay threshold. Then, the above scheme can give corresponding prompts to the user by generating prompt information. For example, by generating prompt information to be confirmed by the user, the user can be prompted to consider whether to modify the delay threshold.

[0012] In a possible implementation manner, the communication method further includes: generating second prompt information when it is determined that a total bandwidth of a path used by at least one service is greater than a bandwidth threshold.

[0013] Then, the above scheme can determine the total bandwidth of the path of the determined service and generate corresponding prompts. In one possible implementation, when the path of a service through a certain entry point to a certain computing power center has been determined, it is limited by the network architecture (for example, the total bandwidth of the links between all network elements in the network) has been fixed, and the total bandwidth of the path used by the service is usually fixed. Then, when it is determined that the total bandwidth used by the service is greater than the bandwidth threshold, it means that the determined path will be difficult to guarantee the implementation of the service. Then, the above scheme can generate prompt information to give corresponding prompts to the user. For example, by generating a prompt message that the total bandwidth has exceeded the threshold, the user can be prompted to consider whether to make corresponding adjustments to the network architecture (for example, to expand the link bandwidth).

[0014] In one possible implementation, before determining at least one path for a service to pass through at least one network element to reach the first computing power center based on network data, it also includes: calculating the computing power cost of the service using the first computing power center based on the unit price of the dimension used by the service in the first computing power center and the usage of the dimension.

[0015] Then, the above scheme can calculate the computing power cost of a computing power center through the unit price of the dimension used by the business in a computing power center and the usage of the dimension, that is, determine the path of the business according to the computing power demand of the business. Among them, the computing power center (for example, the first computing power center) used by the business can be determined based on the computing power demand of the acquired business and the network topology. Since the network usually includes multiple computing power centers, any computing power center has multiple dimensions, and the unit prices of different dimensions using the same computing power center may not be the same, the computing power cost of the business is often determined by whether the business uses a certain computing power center. For example, when it is determined that a business uses a certain computing power center, the cost of the business using the computing power center can be calculated by summing the product of the unit price of any dimension used by the business in the first computing power center and the usage of the dimension. Then, the above scheme can calculate the computing power cost used by the business in different computing power centers and determine the path whose computing power cost meets the corresponding conditions.

[0016] In one possible implementation, before determining at least one path through which the service passes through at least one network element to reach the first computing power center based on network data, it also includes: calculating the network cost based on the link cost of at least one link through which the path passes and the node cost of at least one network element through which the path passes.

[0017] Then, the above scheme can calculate the network cost used by the service through the link cost of the link passed by the path and the node cost of the network element passed by the path, that is, determine the path of the service according to the network topology information. Among them, the link and network element used by the service can be determined according to the obtained network topology information. Since the network usually includes multiple network elements, the network elements and links that the service passes through from a certain network element to the computing power center are not unique, so the network cost of the service is often determined by whether the service passes through the link cost of a certain link and the node cost of the network element passed. In one possible implementation method, the service passes through a certain network element and a certain link to reach the corresponding computing power center. By summing the node cost of the network element passed by the service and the link cost of the link, the network cost of the service can be calculated. Then, the above scheme can calculate the network costs corresponding to the use of different network elements and different links by the service, and determine the path whose network cost meets the corresponding conditions.

[0018] In a possible implementation, before determining at least one path for a service to pass through at least one network element to reach the first computing power center based on network data, it also includes: calculating the capacity expansion cost of the extended bandwidth of at least one link passed by the path, and calculating the extended bandwidth cost.

[0019] Then, the above scheme can calculate the extended bandwidth cost through the expansion cost of the extended bandwidth of at least one link through which the path passes, that is, determine the path of the service according to the extended bandwidth cost. Usually, when the link bandwidth cannot meet the computing power requirements of the service, the bandwidth can be increased by means of link aggregation and the like to achieve bandwidth expansion. Although link expansion can ensure the normal transmission of the service, it is accompanied by an increase in cost. In a possible implementation method, the extended bandwidth of the link through which the service passes can be calculated by summing the expansion cost of the extended bandwidth to calculate the extended bandwidth cost of the service. Then, the above scheme can determine the path of the service based on the cost of the extended bandwidth. It is not difficult to understand that this does not mean that bandwidth expansion must be performed when determining the service path. Then, the above scheme can calculate the extended bandwidth cost of the link used by the service and determine the path whose extended bandwidth cost meets the corresponding conditions.

[0020] In a possible implementation manner, the device information of the network element includes one or more of the following: a device model of the network element, a device capability of the network element, and a service level agreement of the network element.

[0021] Then, the above solution can determine the path of the service based on the equipment model of the network element, the equipment capability of the network element, and the network element service level agreement, thereby determining the path of the service in combination with different network elements.

[0022] In one possible implementation, the equipment information of the computing power center includes one or more of the following: processor computing power, graphics processor computing power, and unit price of dimension.

[0023] Then, the above solution can determine the path of the business based on the processor computing power, graphics processor computing power, and unit price of the dimension, thereby determining the path of the business in combination with different computing power centers.

[0024] In a possible implementation manner, the link information includes one or more of the following: link delay, link bandwidth, link availability, and link utilization.

[0025] Then, the above solution can determine the path of the service based on link delay, link bandwidth, link availability, and link utilization, thereby determining the path of the service in combination with link information of different links.

[0026] In one possible implementation, the computing power requirements of the service include: one or more of the calculated network elements, the network service level agreement requirements, and the computing power service level agreement requirements; the calculated network elements are the network elements of the service entry path; among them, the network service level agreement requirements include one or more of the following: network latency, network bandwidth, network availability, and network security; the computing power service level agreement requirements include one or more of the following: processor computing power, graphics processor computing power, and computing power cost.

[0027] Then, through the above scheme, the path of the service can be determined according to one or more of the computing power requirements of the acquired service, including the network elements to be calculated, the network service level agreement requirements, and the computing power service level agreement requirements, that is, the path can be determined comprehensively from different angles. For example, when it is necessary to consider the location and cost of the network elements to be calculated, it is necessary to determine the path of the service in combination with the network elements to be calculated. Then, the above scheme can determine the path of the service based on different computing power requirements, thereby determining the path of the service in combination with the computing power requirements of different services, thereby being compatible with more scenarios.

[0028] In a second aspect, a communication device is provided. The communication device may be a cloud server (e.g., an operator server), the communication device may also be a module or chip in a cloud server, the cloud server may also be a chip or a system on chip, and the communication device may also be a communication device as a control node. The communication device includes: an acquisition unit and a processing unit; the acquisition unit is used to acquire network data, the network data includes network topology information and computing power requirements of the service, wherein the network topology information includes equipment information of the network element, equipment information of the computing power center, link information of the link between the network element and the computing power center, and link information of the link between the network elements; the processing unit is used to determine at least one path for the service to reach the first computing power center through at least one network element according to the network data acquired by the acquisition unit, and the path satisfies one or more of the following conditions: the computing power cost of the service using the first computing power center confirmed according to the computing power requirements of the service satisfies the first condition, and the network cost of at least one network element and the service using the path confirmed according to the network topology information satisfies the second condition.

[0029] In a possible implementation manner, the processing unit is further configured to determine that the condition satisfied by the path further includes: an extended bandwidth cost of at least one link in the service usage path confirmed according to the network topology information satisfies a third condition.

[0030] In a possible implementation, the link information acquired by the acquisition unit includes link delay; the processing unit is further configured to generate first prompt information when determining that a link delay of a shortest path among at least one path exceeds a delay threshold.

[0031] In a possible implementation manner, the processing unit is further configured to generate second prompt information when it is determined that a total bandwidth of a path used by at least one service is greater than a bandwidth threshold.

[0032] In one possible implementation, the processing unit determines at least one path through which the service passes through at least one network element to reach the first computing power center based on the network data acquired by the acquisition unit, and is also used to calculate the computing power cost of the service using the first computing power center based on the unit price of the dimension used by the service in the first computing power center and the usage of the dimension.

[0033] In one possible implementation, the processing unit determines at least one path through which the service passes through at least one network element to reach the first computing power center based on the network data acquired by the acquisition unit, and is also used to calculate the network cost based on the link cost of at least one link through which the path passes and the node cost of at least one network element through which the path passes.

[0034] In one possible implementation, the processing unit determines at least one path through which the service passes through at least one network element to reach the first computing power center based on the network data acquired by the acquisition unit, and is also used to: calculate the capacity expansion cost of the extended bandwidth of at least one link passed by the path, and calculate the extended bandwidth cost.

[0035] In a possible implementation manner, the device information of the network element acquired by the acquisition unit includes one or more of the following: a device model of the network element, a device capability of the network element, and a service level agreement of the network element.

[0036] In a possible implementation, the device information of the computing power center acquired by the acquisition unit includes one or more of the following: processor computing power, graphics processor computing power, and unit price of dimension.

[0037] In a possible implementation manner, the link information acquired by the acquisition unit includes one or more of the following: link delay, link bandwidth, link availability, and link utilization.

[0038] In one possible implementation, the computing power requirements of the service acquired by the acquisition unit include: one or more of the calculated network elements, network service level agreement requirements, and computing power service level agreement requirements; the calculated network elements are network elements of the service entry path; wherein the network service level agreement requirements include one or more of the following: network latency, network bandwidth, network availability, and network security; the computing power service level agreement requirements include one or more of the following: processor computing power, graphics processor computing power, and computing power cost.

[0039] In a third aspect, a communication device is provided, comprising: a processor and an interface circuit, wherein the processor is coupled to the interface circuit; the processor is configured to execute a computer program or instruction stored in a memory, and control the interface circuit to execute the communication method described in any one of the first aspects.

[0040] In a fourth aspect, a computer-readable storage medium is provided. The computer-readable storage medium stores a computer program or instruction, and when a computer reads and executes the computer program or instruction, the computer executes any method described in the first aspect.

[0041] In a fifth aspect, a computer program product comprising instructions is provided, the computer program product comprising: a computer program code, when the computer program code is run on a computer, the computer can execute the method described in any one of the above aspects.

[0042] Among them, the technical effects brought about by the technical features of the second to fifth aspects and their possible implementation methods can refer to the corresponding effects described in the first aspect and possible implementation methods, and will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 A schematic diagram of a computing power network provided for an embodiment of the present application;

[0044] Figure 2 A schematic diagram of an application scenario provided for an embodiment of the present application;

[0045] Figure 3 A schematic diagram of a communication method provided in an embodiment of the present application;

[0046] Figure 4 A flowchart for implementing computing power network planning provided in an embodiment of the present application;

[0047] Figure 5 A schematic diagram of an integer programming solver provided for an embodiment of the present application;

[0048] Figure 6 An architectural diagram of a computing power network provided for an embodiment of the present application;

[0049] Figure 7 A schematic diagram of a prompt information provided in an embodiment of the present application;

[0050] Figure 8 A schematic diagram of prompt information provided by another embodiment of the present application;

[0051] Fig. 9 A schematic diagram of a communication method provided for another embodiment of the present application;

[0052] Fig.10 A schematic diagram of a computing power network planning result provided in an embodiment of the present application;

[0053] Fig.11 A schematic diagram of a communication device provided in an embodiment of the present application;

[0054] Fig.12 A schematic diagram of a communication device provided for another embodiment of the present application. DETAILED DESCRIPTION

[0055] The technical solutions in the embodiments of the present application will be described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments.

[0056] Unless otherwise defined, all scientific and technological terms used herein have the same meanings as those known to those of ordinary skill in the art. In the embodiments of the present application, the words "first", "second", etc. do not limit the quantity and order. In the embodiments of the present application, "and / or" describes the association relationship of the associated objects, indicating that there may be three relationships, for example, A and / or B, which may represent: the existence of A alone, the existence of A and B at the same time, and the existence of B alone, wherein A and B may be singular or plural.

[0057] It should be noted that, in this application, words such as "exemplary" or "for example" are used to indicate examples, illustrations or descriptions. Any embodiment or design described as "exemplary" or "for example" in this application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "exemplary" or "for example" is intended to present related concepts in a specific way.

[0058] Computing force network (CFN) is a new type of information infrastructure that flexibly schedules and allocates computing resources, storage resources and network resources between cloud, edge and end according to demand. Among them, cloud, also known as cloud server, is a simple, efficient, safe and reliable computing service with elastically scalable processing power, which can provide public Internet infrastructure services to users. Specifically, computing power, that is, the computing power of the device, is usually divided into: general computing power that consumes a small amount of computing power, and high-performance computing power (also known as cluster computing power) with a large amount of computing power. Among them, general computing power is usually used for conventional applications, with a small amount of computing power and only consumes a small amount of computing power. High-performance computing power usually requires a large amount of computing resources to complete a task, and its computing power is large. At the same time, computing power also has multiple measurement dimensions, and different applications usually require different types of computing power. For example, central processing unit (CPU) computing power, graphics processing unit (GPU) computing power, etc. Network refers to a digital telecommunications network that allows nodes to share resources, such as an optical transport network (OTN).

[0059] For example, refer to Figure 1 As shown in Figure 1, a schematic diagram of a computing network is provided. Figure 1 As shown, the computing power network includes: computing power node devices, namely three computing power centers (i.e., computing power center 1, computing power center 2, and computing power center 3), five entry points (i.e., entry point 1, entry point 2, entry point 3, entry point 4, and entry point 5); and the corresponding computing power of the computing power node devices, namely, CPU computing power (refer to Figure 1 The thick solid line in the figure) and GPU computing power (refer to Figure 1 It is not difficult to understand that a computing network usually includes multiple computing points and multiple computing centers. Figure 1 The computing network shown in the figure is taken as an example, and the architecture of the computing network, including the number of computing points and computing centers, should not be limited by this.

[0060] Typically, computing power node devices include: computing power center, which refers to a high-performance computer equipment cluster and network resources, which can provide efficient and stable computing services (such as CPU computing power, GPU computing power), and mainly solve problems such as large-scale data processing and complex computing tasks. In addition to the computing power center, the computing power node device can also optionally include other devices in the computing power network, which can be called network elements (or network element devices). Among them, the network element includes the computing network element, that is, the computing point (also called the edge node or edge computing node), which is located between the user and the cloud and is usually deployed close to the user. In one possible implementation, the network element also includes a single board.

[0061] Among them, the computing power demand can be transmitted through the link between two computing power node devices (for example, between network elements and network elements, between network elements and computing power centers). In one possible implementation, the computing power demand is first sent to the computing entry point, and the computing power demand is directly transmitted by the computing entry point to the computing power center that can meet the computing power demand. Optionally, after the computing power demand is transmitted to the computing entry point, it can be transmitted to the computing power center that can meet the computing power demand through multiple network elements and corresponding multiple links in sequence. Among them, the computing power demand is transmitted from the computing entry point to the computing power center that can meet the computing power demand through multiple links in sequence, forming the path of the computing power demand.

[0062] Unlike traditional networks, computing networks bring different types of user needs in different application scenarios. For example: elasticity and agility, that is, to achieve elastic supply of computing power and provide corresponding computing power according to user needs (including bandwidth, duration, etc.); lossless network, that is, in scenarios involving large amounts of data transmission, to ensure the security of data transmission (that is, lossless data transmission, no network packet loss), and provide users with deterministic network security for data transmission; cloud-network collaboration, that is, the computing network can perceive the computing power requirements of applications and guide the computing power used by the computing network; cross-domain collaboration, that is, to provide unified scheduling for computing power requirements based on the computing power distribution in the computing network; user cost sensitivity, that is, in scenarios involving ultra-long-distance data transmission, to ensure that the bandwidth of the transmission link reaches ultra-large, reducing costs for users; deterministic latency, that is, the computing network has deterministic latency to ensure user experience, etc.

[0063] For example, refer to Figure 2 As shown in the figure, different application scenarios of computing power network are provided. Figure 2 As shown, Figure 2 (1) shows that the computing network is applied to financial enterprises and large enterprises; Figure 2 (2) shows that computing power networks are applied to industries such as astronomy, meteorology, and medicine; Figure 2 (3) in the figure shows the application of computing power network to artificial intelligence (AI) computing; Figure 2 (4) in FIG. 4 shows the application of computing power network to high-performance video.

[0064] Specifically, Figure 2 In (1), the cloud service devices in the upper dotted box represent the enterprise platforms of various levels corresponding to financial enterprises and large enterprises (for example, the national cloud service devices and provincial cloud service devices of a large enterprise). The two local institutions at the bottom upload corresponding data to the provincial cloud service devices respectively, and then transmit them to the national cloud service devices through the provincial cloud service devices. Figure 2 In (2), the cloud service platform corresponding to the astronomy, meteorology, medical and other industries in the upper dotted box (for example, the cloud service platform of a hospital) includes multiple cloud service devices, which collect data through the data collection side (corresponding devices) below and upload it to the cloud service platform. The data is processed by the cloud service platform and output to the data analysis side (corresponding devices). Figure 2 In (3), the upper dotted box includes multiple cloud service devices for remote training, which can provide distributed AI computing, and each independent device below is used for real-time reasoning. Figure 2 In (4), the top cloud service device is used for central cloud model learning and model distribution, and the three cloud service devices below it provide VP / view computing, VP / view computing and VR respectively. The corresponding device at the bottom is used to determine the network (expanded memory, xms) expanded memory.

[0065] In order to meet the above-mentioned user needs, it is necessary to plan the computing network accordingly, which is also called computing network planning. Usually, the resources that need to be used in the optical transmission network and computing node equipment are extracted first, managed independently, and fully pre-configured so that the real computing requests can be used at the optimal cost when they are used. In other words, computing power is truly infrastructure.

[0066] However, since the users of computing power networks are usually commercial customers, their main application scenarios include: high-performance simulation computing, artificial intelligence (AI) training, and large-scale data centers. Therefore, for commercial users, in addition to basic computing power requirements and network requirements, their user requirements can also be divided into several types of complex communication requirements, such as:

[0067] Complex communication requirements: Compared with the traditional model of dedicated network lines (an organization deploys an independent network line, that is, an independent local area network), the planning of computing power networks (such as the paths of computing power networks) and computing power resources requires considering more complex constraints and performing comprehensive calculations.

[0068] Complex demand matching: When planning the computing network, it is necessary to match the computing demand with the appropriate supply (computing center, computing entry point, etc.). For example, high-performance simulation usually requires high-performance CPU computing power, AI training requires high-performance GPU computing power or neural processing unit (NPU) computing power, and large-scale data centers consider storage input / output (I / O) performance.

[0069] Users are cost-sensitive: The unit price of computing power and networks (such as optical networks) is not expensive, but when used in large quantities, costs that cannot be ignored will be incurred. Since commercial users are usually more sensitive to costs, the costs need to be considered in combination with the corresponding costs of computing power and networks, and cannot be calculated separately (considering only a single node).

[0070] Network security requirements: In the application scenarios of computing networks, a large amount of data usually needs to be transmitted, and the interruption loss is large (for example, the interruption of streaming data in AI training may lead to the need to retrain the model). However, the cost of independently protecting the security of transmission through redundancy is very expensive.

[0071] It is not difficult to understand that the above different types of user needs are not completely independent, but coupled with each other. For example, when planning a computing network, it is often necessary to comprehensively consider the economic cost and network security needs. Therefore, this requires comprehensive consideration of the above different types of user needs when planning the computing network, so as to achieve comprehensive planning of the computing network.

[0072] Therefore, when planning the computing network, it is necessary to comprehensively consider the needs of different types of users, which puts higher requirements on the planning of the computing network. In order to meet the needs of different types of customers, the planned computing network needs to support the network service level agreement (SLA), computing power SLA, and joint decision-making of network and computing power. At the same time, the security of transmission needs to be guaranteed.

[0073] It should be noted that the service level agreement (SLA) is a formal commitment defined between the service provider and the user, which stipulates the service quality and performance requirements that the service provider should provide. Generally, SLA refers to the performance indicators of the service, including: bandwidth, that is, the amount of data that can be transmitted per unit time, which is one of the important indicators for measuring the speed of network connection. Latency, that is, the time required for data to travel from the source to the destination. The smaller the delay, the faster the network connection speed. Packet loss rate, that is, the proportion of data packets lost during data transmission. The lower the packet loss rate, the higher the stability and reliability of the network connection. Availability, that is, the availability ratio of network connection within a certain period of time. For example, 99.99% availability means that the network connection will only be interrupted for no more than 4.39 minutes per year. Fault recovery time, that is, the time required to restore to normal operation after a network failure. Then, network SLA includes: network bandwidth, network delay, network packet loss rate, network availability, and network fault recovery time. Computing power SLA includes: computing power bandwidth, computing power delay, computing power packet loss rate, computing power availability, and computing power fault recovery time.

[0074] Based on the above problems, computing network planning usually includes the following methods:

[0075] Method 1: The control node allocates computing power centers according to computing power requirements. The control node calculates the latency and energy consumption of the allocated computing power centers through the pre-trained reinforcement learning model, and updates the reinforcement learning model based on latency and energy consumption. Although latency and energy consumption are taken into account, this solution does not support computing power SLA, cannot guarantee the security of transmission, and has a high cost.

[0076] Method 2: Obtain and store relevant information of each computing node device (including computing center, entry point, etc.) in the computing network, formulate the information issued by the entry point based on the relevant information, and notify each entry point; the entry point can make routing decisions for the access computing demand based on the issued information. Although this solution combines relevant information to plan the computing network so that the computing network can support computing SLA, it does not support network SLA, cannot guarantee the security of transmission, and has a high cost.

[0077] Method 3: According to the computing power demand, determine the allocation route that meets the preset latency, and determine the computing power node device that executes the computing power demand based on the allocation route. This solution only considers latency, does not support computing power SLA, cannot guarantee the security of transmission, and has a high cost.

[0078] Method 4: Through the computing power control module in the computing power network system, based on the acquired computing power routing table, combined with the computing power and routing capabilities of the computing power node devices (such as the computing power center), the computing power node devices that can better execute the computing power requirements are determined. Although this solution takes latency into account and the planned computing power network can support network SLA, this solution does not support computing power SLA, and the security and cost of transmission are difficult to guarantee.

[0079] Based on the above problems, the embodiments of the present application provide a communication method that can obtain network data and determine the path of the service according to the network data, so as to achieve better planning for the computing network. Figure 3 As shown, a schematic diagram of a communication method is provided. Figure 3 , the communication method provided in the embodiment of the present application is described in detail. Figure 1 The communication method provided in the embodiment of the present application is described by taking the architecture shown as an example, but the communication method provided in the embodiment of the present application is not limited thereto. The communication method includes steps 301 to 303, and the communication method is applied to a communication device, which may be a control node, a cloud server, an operator server, etc. The following specifically describes the application of the communication method to a cloud server as an example, and the specific description is as follows:

[0080] Step 301: Obtain network data.

[0081] Combination Figure 3 As shown, the communication device obtains network data. Specifically, the cloud server obtains network data. The network data includes network topology information and computing power requirements of the business, wherein the network topology information includes equipment information of the network element, equipment information of the computing power center, link information of the link between the network element and the computing power center, and link information of the link between the network elements.

[0082] Optionally, the above network topology information can be used to represent the internal structure and connection relationship of the computing network. In one possible implementation, the network topology information of the computing network can be obtained based on the number of computing node devices in the computing network, the location of the computing node devices, and multiple links between the computing node devices. Figure 1 The computing power network shown can obtain the network topology information of the computing power network.

[0083] Optionally, the equipment information of the network element includes one or more of the following: the equipment model of the network element, the equipment capability of the network element, and the service level agreement of the network element; the equipment information of the computing power center includes one or more of the following: processor computing power, graphics processor computing power, and unit price of dimension; the link information between the network element and the computing power center includes one or more of the following: link delay, link bandwidth, link availability, and link utilization.

[0084] In one possible implementation, the computing power requirements of the service include: one or more of the computing network element (i.e., computing point), network service level agreement requirements, and computing power service level agreement requirements, wherein the computing network element is the network element of the service entry path; the network SLA requirements include one or more of the following: network latency, network bandwidth, network availability, and network security; the computing power SLA requirements include one or more of the following: processor computing power, graphics processor computing power, and computing power cost.

[0085] Step 302: Calculate the cost of the service path.

[0086] Combination Figure 3 As shown, the communication device calculates the cost of the path through which the service passes. Specifically, the cloud server calculates the cost of the path through which the service passes.

[0087] Optionally, the cost of the path that the service passes through includes one or more of the following: the computing power cost of the computing center used by the service, the network cost of the link and network element used by the service, and the extended bandwidth cost of the link used by the service. Exemplarily, the specific calculation of the cost of the path that the service passes through includes:

[0088] (1) For example, the computing power cost of a computing center used by a business can be calculated by a mathematical expression, as follows:

[0089]

[0090] Among them, the business (i.e. the computing power demand of the business) is expressed as: i, and the computing power cost of business i is expressed as: computeCost i , the computing center is represented by: j, and the computing center j used by business i is represented by: The unit price of the kth dimension of computing center j (unit: yuan / flops) is expressed as: The usage of business i in the kth dimension is expressed as:

[0091] It is not difficult to understand that, based on Formula 1, the computing power cost of a business is determined by whether the business uses a certain computing center. Specifically, the computing power cost of a business can be calculated based on Formula 1 based on the unit price of the dimension used by the business in the computing center and the usage of the dimension.

[0092] (2) For example, the network cost of a service passing through a link and a network element can be calculated by a mathematical expression, as follows:

[0093]

[0094] The network cost of service i is expressed as: networkCost i , the link that service i passes through is represented as: m, and service i uses link m as: The network element that service i passes through is represented by: c, and the network element that service i passes through is represented by: The network element used by service i is represented by: p, and the network element used by service i is represented by: The link cost (unit: yuan) of link m is expressed as: linkCost m , the node cost of network element c (unit: yuan) is expressed as: nodeCost c .

[0095] It is not difficult to understand that, based on Formula 2, the network cost of a service passing through a link and a network element is determined by whether the service uses a certain link and a certain network element. Specifically, Formula 2 can be used to understand the composition of the network cost. The network cost of a service passing through a link and a network element is calculated based on Formula 2 by the link cost of at least one link that the path passes through and the node cost of at least one network element that the path passes through.

[0096] (3) For example, the extended bandwidth cost of the link used by the service can be calculated by a mathematical expression, as follows:

[0097]

[0098] The expansion bandwidth cost used by service i is expressed as: expandcosts i , the expansion cost of service i using the extended bandwidth of link m is expressed as: expand m , the coefficient of link m used by service i is expressed as: M.

[0099] It is not difficult to understand that, based on Formula 3, the extended bandwidth cost of the link used by the service is determined by whether the service uses a certain link and a certain network element. Specifically, the extended bandwidth cost of the link used by the service is calculated based on Formula 3 by the expansion cost of the extended bandwidth of at least one link that the path passes through.

[0100] Step 303: Determine and determine at least one path of the service.

[0101] Combination Figure 3 As shown, the communication device will judge and determine at least one path of the service. In a possible implementation, the cloud server can judge and determine at least one path of the service based on corresponding preset conditions.

[0102] In one possible implementation, the cloud server can judge and determine at least one path of the service based on whether the cost of the service path, including the computing power cost of the computing power center used by the service, meets the preset computing power cost conditions.

[0103] Optionally, the cloud server can judge and determine at least one path of the service based on whether the cost of the service path, including the link used by the service and the network cost of the network element, meets a preset network cost condition.

[0104] For another example, the cloud server determines and determines at least one path of the service according to whether the extended bandwidth cost of the link used by the service, included in the cost of the path of the service, meets the extended bandwidth cost preset condition.

[0105] It is not difficult to understand that this is only used as an example to illustrate a possible implementation method of judging and determining the path of a service based on the cost of the path of the service, and should not be used to limit the communication method provided in the embodiment of the present application. Of course, those skilled in the art can also set the above-mentioned computing power cost preset conditions, network cost preset conditions, and extended bandwidth cost preset conditions according to actual conditions.

[0106] Based on the above, the path of the service can also be judged according to the preset constraints of the path of the service. Exemplarily, the preset constraints of the path of the service can be expressed by mathematical inequalities, including:

[0107] (1) For example, the path of the service may be determined by using a preset mathematical inequality corresponding to the constraint of the correspondence between the network element and the link, as follows:

[0108]

[0109]

[0110] Among them, the two network elements are represented as c1 and c2 respectively, c2 is the reference network element, The value of includes 0 (indicating that service i does not use link m) and 1 (indicating that service i uses link m). The value of includes 0 (indicating that service i does not use network element c) and 1 (indicating that service i uses network element c). m =(c1,c2)∈E means that the direction of service i using link m is from c1 to c2, e m =(c2,c1)∈E means that the direction in which service i uses link m is from c2 to c1. For example, based on the above formula 4, if The value is 1. According to the inequality relationship, The value is 1, and the direction of link m used by service i is from c1 to c2, that is, c1 is not the computing power center, and the service will continue to be transmitted through the link until it is transmitted to the computing power center. Then, through the above formulas 4 and 5, the network elements used by the service path can be judged, and thus the service path can be judged.

[0111] (2) Exemplarily, the path of the service may be determined by a preset mathematical inequality corresponding to the protection level of the path (eg, primary path, backup path, redundant path), as follows:

[0112] Optionally, the path composition type may be different according to the protection level of the preset path. For example, the path composition type may include: a single main path, a main path and a backup path, a main path, a backup path and a redundant path, a main path, a backup path and a redundant path.

[0113] It is not difficult to understand that, for the sake of ease of explanation, only the four path composition types shown here are taken as examples, and the types of paths and the communication methods provided in the embodiments of the present application should not be limited thereto.

[0114]

[0115] The l path of service i is represented by link m as follows: The values ​​of include 0 (indicating that the l path of service i does not use link m) and 1 (indicating that the l path of service i uses link m). m =(c1,c2)∈E means that the direction of path l of service i using link m is from c1 to c2, e m =(c2,c1)∈E means that the direction of path l of service i using link m is from c2 to c1. Service i uses network element c1 as the entry point and is expressed as: The value of includes 0 (indicating that service i does not use network element c1 as the computing point) and 1 (indicating that service i uses network element c1 as the computing point). Service i uses network element c1 as the computing power center as follows: The values ​​include 0 (indicating that service i does not use network element c1 as the computing power center) and 1 (indicating that service i uses network element c1 as the computing power center).

[0116] For example, based on the above formula 6, if In e m =(c1,c2)∈E direction takes value 1, in e m =(c2,c1)∈E direction takes value 0, then The value of is 1, that is, service i uses network element c1 as the entry point, and the service will continue to be transmitted through the link until it is transmitted to the computing center. Then, through the above formula 6, it is possible to judge whether a network element used by the service is an entry point or a computing center, thereby judging the path of the service.

[0117]

[0118]

[0119]

[0120]

[0121] When the l path of service i uses link m as the main path, l takes the value of 1 (i.e. ); When the l path of service i uses link m as the backup path, the value of l is 2 (i.e. ); When the l path of service i uses link m as a redundant path, the value of l is 3 (i.e. ); When the l path of service i uses link m as the redundant second path, the value of l is 4 (i.e. ),and The value of still includes 0 (link m is not used) or 1 (link m is used). For example, based on the above formula 7, if The value is 1. The value is 1. According to the inequality relationship, If the value is 0, the l path of service i uses link m as the main path, and the l path of service i does not use link m as the backup path, that is, the main path and the backup path do not overlap (for link m).

[0122] Then, formula seven can be used to determine that the main path and the backup path do not overlap, formula eight can be used to determine that the main path and the redundant second path do not overlap, formula nine can be used to determine that the backup path and the redundant first path do not overlap, and formula ten can be used to determine that the redundant first path and the redundant second path do not overlap, thereby determining the path of the service.

[0123]

[0124]

[0125]

[0126]

[0127] Among them, the path composition type n of service i reuses link m and is expressed as: The limit on the number of times service i reuses link m is expressed as: Specifically, n takes a value of 1, indicating that the path composition type is a primary path and a redundant path, that is, the primary path and the redundant path reuse link m as follows: Maximum overlap times; n is 2, indicating that the path composition type is a backup path and a redundant second path, that is, the backup path and the redundant second path reuse link m as follows: Maximum overlap Second-rate.

[0128] Then, based on Formula 11 to Formula 14, by The value of can limit the number of path overlaps, and then judge the path of the service.

[0129] (3) For example, the path of the service may be determined by a preset mathematical inequality corresponding to the computing power constraint of the computing power center, as follows:

[0130]

[0131] Among them, the usage of business i in the kth dimension of computing center j is expressed as: The computing power of computing center j in the kth dimension is expressed as: Then, based on Formula 15, the path of the business can be judged by ensuring that the usage of the business in a certain dimension of the computing center does not exceed the computing power of the computing center j in a certain dimension.

[0132] (4) For example, the path of the service may be determined by a preset mathematical inequality corresponding to the link bandwidth constraint, as follows:

[0133]

[0134] Among them, the path composition type n of service i uses the bandwidth of link m to be expressed as: The available bandwidth of link m is expressed as: capacity m , the expanded bandwidth of link m is expressed as: expand m Optional, n takes a value of 1, indicating that the path composition type is a primary path and a redundant first path, and n takes a value of 2, indicating that the path composition type is a backup path and a redundant second path.

[0135] In one possible implementation, expand m The value of can be set to 0, which means that when the l path of service i uses link m, the bandwidth of link m is not expanded. When the bandwidth of link m used by the path composition type n of service i exceeds the available bandwidth of link m, the bandwidth of link m needs to be expanded, that is, expand m The value of is not 0.

[0136] Then, based on Formula 16, in actual application, it is possible to choose whether to expand the bandwidth of the link according to the scenario, and calculate the corresponding cost of the expanded bandwidth, so as to judge the path of the service.

[0137] Based on the above, when performing computing network planning, multiple paths of the service can be judged and determined according to the preset constraints corresponding to the corresponding relationship between network elements and links. Then, the embodiments of the present application can judge the path of the service according to the preset constraints of the path of the service, so as to comprehensively consider the needs of different types of users.

[0138] Combination Figure 3 The communication method shown, exemplary, refers to Figure 4 As shown in Figure 1, a flowchart for realizing computing network planning is provided. Figure 4 As shown, by analyzing the network topology (refer to Figure 4 (1) in the figure) to identify the planning information and obtain the network topology information, including: the equipment information of the network element, the computing center (refer to Figure 4 The device information of the circular area in (1), the link information of the link between the network element and the computing center, and the link information of the link between the network elements; then, based on the obtained network topology information, the preset constraints of the service path are obtained through constraint integration digitization.

[0139] Optional, combined Figure 4 As shown in (2) of the figure, by the computing power requirements of the business (refer to Figure 4 The computing power requirement table shown in (2) in the figure includes the computing network elements, network SLA requirements, and computing power SLA requirements. Related parameters (such as business type, computing power center, entry point, bandwidth, latency, and GPU computing power) are used to convert business development intentions. The relevant parameters can be converted from numerical values ​​to corresponding requirements for planning the computing power network (such as constraints on the computing power center used for the business path), and then the preset constraints on the business path are obtained through constraint integration digitization.

[0140] In this way, based on network data (including network topology information and business computing power requirements), the preset constraints of the business path can be obtained through constraint integration digitization. Furthermore, based on the preset constraints of network data and business paths, a mathematical programming algorithm (refer to Figure 5 As shown in the figure, the path of the business can be judged and determined to realize the computing network planning. According to the computing network planning, resource division can be performed (such as the selection of computing center). Then, through the above process, it is possible to Figure 4Alternatively, based on the preset constraints of network data and service paths, a mathematical programming algorithm (see Figure 5 If the network planning cannot be implemented, that is, the computing power requirements of the business cannot be met, an unsatisfactory recommendation will be generated.

[0141] For example, refer to Figure 5 As shown, the embodiment of the present application provides a schematic diagram of an integer programming solver implementing the above mathematical programming algorithm. Figure 4 (3) in the figure) can be implemented by the integer programming solver, and the optimal solution can be found through the integer programming solver (i.e., the path of the business is judged and determined to find the optimal path of the business); of course, the embodiments of the present application are not limited to the implementation of the above-mentioned mathematical programming algorithm only through the integer programming solver.

[0142] Specific, combined Figure 5 As shown, the integer programming solver includes the following steps:

[0143] Step 501: Reduce branches by using a cutting plane algorithm.

[0144] The branches in the algorithm are used to represent the links used by the paths of the services; for example, Figure 6 As shown, an embodiment of the present application provides an architecture diagram of a computing network. Among them, the service use link m can be expressed as: (For example, m=1 represents link 1, m=2 represents link 2, and m=3 represents link 3), The value of includes 0 and 1. For example, Indicates that the service does not use link 1. Indicates that the service uses link 1. It is not difficult to understand that when actually planning the computing network, the number of links used by the service path is usually multiple, that is, the service will be transmitted to the computing center through multiple links in sequence to meet the computing power requirements of the service. In a possible implementation, combined with Figure 6 As shown, the service uses link 1 Includes the corresponding two branches (0 and 1), using link 2 Includes the corresponding two branches (0 and 1), using link 3 Including two corresponding branches (0 and 1), if the computing power demand of the business needs to be realized by computing center 1. In this way, the business can be transmitted to computing center 2 using links 1 and 2, without using link 3, that is, the reduced branch is the branch corresponding to link 3.

[0145] Step 502: Update the upper bound of the solution using a heuristic algorithm.

[0146] Step 503: Update the lower bound of the solution through a linear relaxation algorithm.

[0147] Step 504: If the difference between the upper bound and the updated lower bound is less than the threshold, the solution is already the optimal solution, and the integer programming solver will no longer continue to seek the optimal solution and will end directly. Of course, the threshold in step 504 can be set to 10 -6 , depending on the accuracy of the algorithm, the threshold can also be set to other values; in some examples, the threshold may also be 10 -5 , 10 -7 wait.

[0148] Step 505: If the difference between the upper bound and the updated lower bound is greater than or equal to the threshold, the integer programming solver will calculate the next part of the relaxation branch (the branch corresponding to the range of the solution expanded by the linear relaxation algorithm).

[0149] Step 506: If it is determined that the branch is not feasible, delete the branch.

[0150] Step 507: If it is determined that the branch is feasible, continue to determine whether the solution corresponding to the branch is an integer and is less than the upper bound.

[0151] Step 508: If it is determined that the solution corresponding to the branch is an integer and is smaller than the upper bound, update the upper bound of the solution.

[0152] Step 509: If it is determined that the solution corresponding to the branch is not an integer and is smaller than the upper bound, continue to determine whether the branch set is empty.

[0153] Step 510: If it is determined that the branch set is not empty, jump to step 505 (calculate the next part of the relaxed branch) and continue to loop and execute the corresponding steps.

[0154] Step 511: If the branch set is determined to be empty, the lower bound and the corresponding solution are output.

[0155] In combination with the above steps 501 to 511, if the computing power network planning cannot be achieved based on the preset constraints of the network data and the path of the service (for example, the delay of the shortest link between network elements cannot meet the delay included in the computing power demand of the service), a prompt will be provided by providing unsatisfactory suggestions so that the user can make corresponding adjustments accordingly. For example, the user can choose to adjust the delay included in the computing power demand of the service.

[0156] Based on the above problems, for example, refer to Figure 7 As shown, an embodiment of the present application provides a schematic diagram of prompt information. Figure 3 As shown, through the communication method provided in the embodiment of the present application, the link delay can also be judged and corresponding prompts can be generated.

[0157] Optional, combined Figure 3 As shown, based on step 301 to step 303, the communication method further includes:

[0158] Step 304: Determine the link delay of the shortest path of the service and generate prompt information.

[0159] Specifically, the network data acquired by the communication device includes link information of links between network elements, and the link information includes link delay. Then, when the communication device determines that the link delay of the shortest path among one or more paths of the service exceeds the delay threshold, a prompt message is generated. Among them, those skilled in the art can set the delay threshold according to actual conditions.

[0160] Based on the above step 304, combined with Figure 7 As shown, in a possible implementation, the prompt information includes the number of links whose delays do not meet the delay threshold, the problem location, and the reason for not meeting the delay threshold. Figure 7 It shows that the number of links whose latency does not meet the latency threshold is 1, the problem location is Shanghai, and the reason for not meeting the latency threshold is that a computing power center that meets the latency threshold cannot be found.

[0161] In one possible implementation, the communication device may also use the Dijkstra algorithm to pre-calculate each service based on the delay and protection settings set by the user, analyze whether the shortest path can meet the delay threshold, and if not, prompt the customer in a prompt message.

[0162] It should be noted that the above delay determination process can be performed before step 301, and the user can choose to continue computing network planning through prompt information, or adjust the delay threshold before computing network planning. For another example, the above delay determination process can be performed after step 303 (refer to Figure 3 As shown), the user can make corresponding adjustments based on the prompt information, such as adjusting the constraints of the path.

[0163] For example, refer to Figure 8 As shown, the embodiment of the present application also provides a schematic diagram of prompt information. Figure 3 As shown, through the communication method provided in the embodiment of the present application, the link bandwidth can also be judged and corresponding prompts can be generated.

[0164] Optional, combined Figure 3 As shown, based on step 301 to step 303, the communication method further includes:

[0165] Step 305: Determine the total bandwidth of the path used by the service and generate prompt information.

[0166] Specifically, when the communication device determines that the total bandwidth of the path used by the service is greater than the bandwidth threshold, it generates a prompt message. Among them, those skilled in the art can set the bandwidth threshold according to actual conditions.

[0167] Based on the above step 305, combined with Figure 8 As shown, in a possible implementation, the prompt information includes the number of links whose bandwidth does not meet the bandwidth threshold, the problem location, and the reason for not meeting the threshold. Figure 8 It is shown that the number of links whose bandwidth does not meet the delay threshold is 1, the problem location is Xinyang to Xiangyang, and the reason for not meeting the threshold is the bandwidth difference of 50000.0 Mbps (bitper second, ie, bit rate).

[0168] It should be noted that the above bandwidth determination process can be performed before step 301, and the user can choose to continue computing network planning through prompt information, or adjust the bandwidth threshold before computing network planning. For another example, the above bandwidth determination process can be performed after step 303 (refer to Figure 3 As shown), the user can make corresponding adjustments based on the prompt information, such as adjusting the constraints on the extended bandwidth.

[0169] It is not difficult to understand that there is no strict sequence between the above steps 304 and 305. The embodiments of the present application are only taken as an example, and the specific sequence of executing the above steps (step 304, step 305) should not be limited thereto.

[0170] In a possible implementation, the calculation algorithm part included in the method in the above embodiment can also be implemented by designing a self-developed algorithm based on the Dantzig-Wolfe decomposition algorithm, supporting other single-source single-destination path calculations in the controller and the top-level planning algorithm. For example, in combination with the above formulas 6 to 14, in actual calculations, the Dantzig-Wolfe decomposition algorithm can be used, other graph theory algorithms can be used as atomic algorithms, and the top level uses a mathematical programming algorithm for calculation.

[0171] Based on the above, exemplary, reference Fig. 9 As shown, the embodiment of the present application also provides a schematic diagram of a communication method. Specifically, the communication method includes steps 801 to 803, which are specifically described as follows:

[0172] Step 801: Obtain network data.

[0173] Combination Fig. 9As shown in Figure 1, obtaining network data includes: graph data preparation, business planning processing, and reachability abstraction. Fig. 9 As shown in (1) in the figure, graph data preparation refers to obtaining the network topology structure from the database. For example, the network topology structure can be a large network topology graph (refer to Fig. 9 As shown in (1) in FIG. 1 ). Of course, network topology information can be obtained based on the network topology structure. Fig. 9 As shown in (2) in the figure, business planning processing refers to tailoring the obtained network topology structure based on the computing power requirements of the business, so as to obtain the simplest network topology structure. For example, discarding network elements and links that are not related to the business in the network topology structure (refer to Fig. 9 The dotted part of the network topology in (3)). Combined with Fig. 9 As shown in (4) in the figure, reachability abstraction refers to converting the reachability of services based on logical judgment into the reachability of network topology. For example, even if two services use NE 1, NE 2, and NE 4 (usually the paths of the two services do not overlap logically), and NE 2 is an independent NE, reachability abstraction can separate the paths of the two services in the network topology, thereby distinguishing the paths of the two services.

[0174] Step 802: Planning algorithm.

[0175] Combination Fig. 9 As shown, the planning algorithm includes: source-destination matching, one or more service path calculations, and historical obstacle SLA planning. Specifically, source-destination matching refers to matching the entry points and computing centers of the service paths. One or more service path calculations refer to determining the path of each service and determining one or more paths for each service. Historical obstacle SLA planning refers to judging and planning the path of each service based on the parameters corresponding to the historical fault SLA of the network, reducing the number of failures in the service path, and making the failure rate of computing network planning lower.

[0176] In a possible implementation, step 802 may also use mathematical programming logic to establish an overall model, and decompose the design module (such as the source-destination matching module described above) based on the Dantzig-Wolfe decomposition algorithm. In this way, the algorithm model of the above-mentioned planning algorithm (except the source-destination matching module) can be calculated independently, that is, it meets the structure required for decomposition of the Dantzig-Wolfe decomposition algorithm.

[0177] Step 803: Implement computing power network planning.

[0178] Computing network planning is achieved based on the paths planned by the planning algorithm for one or more services.

[0179] Then, based on Figure 3 The communication method shown, the embodiment of the present application can be compatible with other basic path calculation algorithms as the underlying algorithm.

[0180] Based on the communication method described above, for example, refer to Fig.10 As shown, the embodiment of the present application provides a schematic diagram of the result of computing power network planning. That is, through the communication method provided in the embodiment of the present application, computing power network planning is performed, and the result of the computing power network planning can be directly observed through the generated corresponding interface, that is, the visualization of computing power network planning is realized.

[0181] Specific, combined Fig.10 As shown, the results of computing network planning include: graphical interface (refer to Fig.10 (1) in the table above) and a detailed list of resource usage (including network data) (see Fig.10 (as shown in (2) in the figure).

[0182] Combination Fig.10 As shown in (1), the graphical interface includes a topology diagram of the computing network planning and a prompt bar. Optionally, the prompt bar includes a bandwidth unsatisfied quantity, a latency unsatisfied quantity, and a planning button. Fig.10 (1) in the figure shows that the number of bandwidth non-satisfactions and the number of latency non-satisfactions of the computing power network planning are 0.

[0183] Combination Fig.10 As shown in (2), the detailed list of resource usage includes a search bar, city, network element name, board, port, and reserved bandwidth. Fig.10 (2) in FIG. 1 shows that the network element name used by a service in Xi'an is Xi'an-NE90601, the board is 1-TNS2N504, the port is Port1, and the reserved bandwidth is 210G.

[0184] For example, refer to Fig.11As shown, an embodiment of the present application provides a schematic diagram of a communication device. The communication device may also be a module or chip in a cloud server, the cloud server may also be a chip or a system on chip, and the communication device may also be a communication device as a control node. The communication device includes: an acquisition unit 901 and a processing unit 902; the acquisition unit 901 is used to acquire network data, the network data includes network topology information and computing power requirements of the service, wherein the network topology information includes equipment information of the network element, equipment information of the computing power center, link information of the link between the network element and the computing power center, and link information of the link between the network elements; the processing unit 902 is used to determine at least one path for the service to reach the first computing power center through at least one network element according to the network data acquired by the acquisition unit 901, and the path satisfies one or more of the following conditions: the computing power cost of the service using the first computing power center confirmed according to the computing power requirements of the service satisfies the first condition, and the network cost of at least one network element and the service using the path confirmed according to the network topology information satisfies the second condition.

[0185] The acquisition unit 901 is further used to execute the communication method described in step 301; the processing unit 902 is further used to execute the communication methods described in steps 302, 303, 304 and 305. It can be understood that the communication device can directly refer to the above Figure 3 The description of the various functions and effects in the communication method shown will not be repeated here.

[0186] In a possible implementation, an embodiment of the present application further provides a communication device, which may be a chip or a chip system. The communication device includes: a processor and an interface circuit, and the processor is used to read instructions to execute the method in any of the above method embodiments. In a possible design, the communication device also includes a memory. The memory is used to store necessary program instructions and data, and the processor can call the program code stored in the memory to instruct the communication device to execute the method in any of the above method embodiments. Of course, the memory may not be in the communication device. When the communication device is a chip system, it may be composed of chips, or it may include chips and other discrete devices, which is not specifically limited in the embodiments of the present application.

[0187] For example, refer to Fig.12 As shown, an embodiment of the present application provides a schematic diagram of a communication device. Specifically, the communication device includes one or more transceivers 1001 and a processor 1002. Exemplarily, the transceiver 1001 is used to perform the following Fig.11 The functions or steps implemented by the acquisition unit 901 shown in FIG. 1 are as follows: the processor 1002 is used to execute Fig.11 The functions or steps implemented by the processing unit 902 shown in FIG. 1001 and the processor 1002 can be referred to in detail. Figure 3 Or the method embodiments shown above will not be described in detail here.

[0188] In the above-mentioned embodiments, for the specific description of each term or step, please refer to the introduction in the above method embodiment, which will not be described in detail here.

[0189] exist Fig.12 In various implementations of the communication device shown, the transceiver may include a receiver, which is used to perform the acquisition function mentioned above, and the acquisition function may specifically include receiving, for example, the receiver is used to perform the function (or operation) of receiving. And the transceiver is used to communicate with other devices / devices through a transmission medium.

[0190] Optionally, the communication device may also include one or more memories 1003 for storing program instructions and / or data. The memory 1003 is coupled to the processor 1002. The coupling in the embodiment of the present application is an indirect coupling or communication connection between devices, units or modules, which may be electrical, mechanical or other forms, and is used for information exchange between devices, units or modules. The processor 1002 may operate in conjunction with the memory 1003. The processor 1002 may execute program instructions stored in the memory 1003. Optionally, at least one of the above one or more memories may be included in the processor.

[0191] The specific connection medium between the transceiver 1001, the processor 1002 and the memory 1003 is not limited in the embodiment of the present application. Fig.12 In the embodiment, the memory 1003, the processor 1002 and the transceiver 1001 are connected via a bus 1004. Fig.12 The connections between the other components are shown in bold lines, which are only for illustration and are not intended to be limiting. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Fig.12 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0192] In the embodiments of the present application, the processor may be a general-purpose processor, a digital signal processor, an application-specific integrated circuit, a field programmable gate array or other programmable logic device, a discrete gate or transistor logic device, a discrete hardware component, etc., and may implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of the present application. The general-purpose processor may be a microprocessor or any conventional processor, etc. The steps of the method disclosed in the embodiments of the present application may be directly embodied as being executed by a hardware processor, or may be executed by a combination of hardware and software modules in the processor, etc.

[0193] In the embodiment of the present application, the memory may include, but is not limited to, non-volatile memories such as a hard disk (HDD) or a solid-state drive (SSD), random access memory (RAM), erasable programmable read-only memory (EPROM), read-only memory (ROM) or portable read-only memory (CD-ROM), etc. The memory is any storage medium that can be used to carry or store program codes in the form of instructions or data structures and can be read and / or written by a computer (such as the communication device shown in the present application), but is not limited thereto. The memory in the embodiment of the present application can also be a circuit or any other device that can realize a storage function, for storing program instructions and / or data.

[0194] The processor 1002 is mainly used to process the communication protocol and communication data, and to control the entire communication device, execute the software program, and process the data of the software program. The memory 1003 is mainly used to store the software program and data. The transceiver 1001 may include a control circuit and an antenna. The control circuit is mainly used to convert the baseband signal and the radio frequency signal and process the radio frequency signal. The antenna is mainly used to send and receive radio frequency signals in the form of electromagnetic waves. The input and output devices, such as a touch screen, a display screen, a keyboard, etc., are mainly used to receive data input by the user and output data to the user.

[0195] When the communication device is turned on, the processor 1002 can read the software program in the memory 1003, interpret and execute the instructions of the software program, and process the data of the software program. When data needs to be sent wirelessly, the processor 1002 performs baseband processing on the data to be sent, and outputs the baseband signal to the radio frequency circuit. The radio frequency circuit performs radio frequency processing on the baseband signal and then sends the radio frequency signal outward in the form of electromagnetic waves through the antenna. When data is sent to the communication device, the radio frequency circuit receives the radio frequency signal through the antenna, converts the radio frequency signal into a baseband signal, and outputs the baseband signal to the processor 1002. The processor 1002 converts the baseband signal into data and processes the data.

[0196] In another implementation, the RF circuit and antenna may be arranged independently of the processor performing baseband processing. For example, in a distributed scenario, the RF circuit and antenna may be arranged remotely from the communication device.

[0197] The communication device shown in the embodiment of the present application may also have Fig.12The embodiments of the present application do not limit the number of components and the like. The methods executed by the processor and the transceiver shown above are only examples, and the specific steps executed by the processor and the transceiver can refer to the methods described above.

[0198] In one possible implementation, an embodiment of the present application provides a communication device. The communication device includes: a processor and an interface circuit, wherein the processor is coupled to the interface circuit; the processor is used to execute a computer program or instruction stored in a memory, and control the interface circuit to execute the communication method in any of the above method embodiments. In a possible design, the communication device also includes a memory. The memory is used to store necessary program instructions and data, and the processor can call the program code stored in the memory to instruct the communication device to execute the method in any of the above method embodiments. Of course, the memory may not be in the communication device. When the communication device is a chip system, it can be composed of a chip, or it can include a chip and other discrete devices, and the embodiments of the present application do not specifically limit this.

[0199] In a possible implementation, an embodiment of the present application provides a computer-readable storage medium. The computer-readable storage medium stores a computer program or instruction, and when a computer reads and executes the computer program or instruction, the computer executes the communication method in any of the above method embodiments.

[0200] In a possible implementation, an embodiment of the present application further provides a computer program product comprising instructions, the computer program product comprising: a computer program code, when the computer program code is run on a computer, enables the computer to execute the communication method in any of the above method embodiments.

[0201] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using a software program, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions may be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions may be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium may be any available medium that can be accessed by a computer or a data storage device such as a server or data center that includes one or more available media integrated therein. The available medium may be a magnetic medium (e.g., a floppy disk, a hard disk, a magnetic tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid state disk (SSD)), etc. In an embodiment of the present application, the computer may include the device described above.

[0202] Although the present application is described herein in conjunction with various embodiments, in the process of implementing the claimed application, those skilled in the art may understand and implement other variations of the disclosed embodiments by viewing the drawings, the disclosure, and the appended claims. In the claims, the word "comprising" does not exclude other components or steps, and "one" or "an" does not exclude multiple situations. A single processor or other unit may implement several functions listed in a claim. Certain measures are recorded in different dependent claims, but this does not mean that these measures cannot be combined to produce good results.

[0203] Although the present application has been described in conjunction with specific features and embodiments thereof, it is obvious that various modifications and combinations may be made thereto without departing from the spirit and scope of the present application. Accordingly, this specification and the drawings are merely exemplary illustrations of the present application as defined by the appended claims, and are deemed to have covered any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, those skilled in the art may make various modifications and variations to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A communication method, characterized in that: include: Acquire network data, the network data including network topology information and computing power requirements of the business, wherein the network topology information includes equipment information of the network element, equipment information of the computing power center, link information of the link between the network element and the computing power center, and link information of the link between the network elements; Determine, based on the network data, at least one path for the service to reach the first computing power center through at least one of the network elements, and the path satisfies one or more of the following conditions: the computing power cost of the service using the first computing power center confirmed based on the computing power demand of the service satisfies the first condition, and the network cost of at least one of the links in the path used by the service confirmed based on the network topology information and at least one of the network elements satisfies the second condition.

2. The communication method according to claim 1, characterized in that: The condition satisfied by the path further includes: an extended bandwidth cost of the service using at least one of the links in the path confirmed according to the network topology information satisfies a third condition.

3. The communication method according to claim 1 or 2, characterized in that: The link information includes link delay; the communication method further includes: When it is determined that the link delay of the shortest path among the at least one path exceeds the delay threshold, first prompt information is generated.

4. The communication method according to claim 1 or 2, characterized in that: The communication method further comprises: When it is determined that the total bandwidth of the path used by at least one service is greater than the bandwidth threshold, second prompt information is generated.

5. The communication method according to claim 1, characterized in that: Before determining, according to the network data, at least one path for the service to reach the first computing center through at least one of the network elements, the method further includes: The computing power cost of the business using the first computing center is calculated based on the unit price of the dimension used by the business in the first computing center and the usage of the dimension.

6. The communication method according to claim 1, characterized in that: Before determining, according to the network data, at least one path for the service to reach the first computing center through at least one of the network elements, the method further includes: The network cost is calculated based on the link cost of at least one link through which the path passes and the node cost of at least one network element through which the path passes.

7. The communication method according to claim 2, characterized in that: Before determining, according to the network data, at least one path for the service to reach the first computing center through at least one of the network elements, the method further includes: The extended bandwidth cost is calculated based on the capacity expansion cost of the extended bandwidth of at least one link through which the path passes.

8. The communication method according to any one of claims 1 to 7, characterized in that: The equipment information of the network element includes one or more of the following: equipment model of the network element, equipment capability of the network element, and service level agreement of the network element.

9. The communication method according to any one of claims 1 to 8, characterized in that: The equipment information of the computing power center includes one or more of the following: processor computing power, graphics processor computing power, and unit price of dimension.

10. The communication method according to any one of claims 1 to 9, characterized in that: The link information includes one or more of the following: link delay, link bandwidth, link availability, and link utilization.

11. The communication method according to any one of claims 1 to 7, characterized in that: The computing power requirement of the service includes: one or more of: a computing network element, a network service level agreement requirement, and a computing power service level agreement requirement; the computing network element is a network element through which the service enters the path; The network service level agreement requirements include one or more of the following: network latency, network bandwidth, network availability, and network security; The computing power service level agreement requirements include one or more of the following: processor computing power, graphics processor computing power, and computing power cost.

12. A communication device, characterized in that: include: An acquisition unit, configured to acquire network data, wherein the network data includes network topology information and computing power requirements of a service, wherein the network topology information includes equipment information of a network element, equipment information of a computing power center, link information of a link between the network element and the computing power center, and link information of a link between the network elements; A processing unit is used to determine at least one path for the service to reach the first computing power center through at least one of the network elements based on the network data acquired by the acquisition unit, and the path satisfies one or more of the following conditions: the cost of the service using the computing power of the first computing power center confirmed according to the computing power demand of the service satisfies the first condition, and the cost of the service using at least one of the links in the path and at least one of the network elements confirmed according to the network topology information satisfies the second condition.

13. A communication device, characterized in that: The method comprises a processor and an interface circuit, wherein the processor is coupled to the interface circuit; The processor is used to execute a computer program or instruction stored in the memory, and control the interface circuit to execute the communication method according to any one of claims 1 to 11.

14. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program or instructions. When a computer reads and executes the computer program or instructions, the computer executes the method according to any one of claims 1 to 11.

Images