A computing power resource management method, system, device, medium and program product

Abstract

The application discloses a kind of computing power resource management method, system, equipment, medium and program product, computing power resource management system includes resource demand platform, agent gateway and computing power resource pool;The resource demand platform is used to obtain the demand description text input by user;According to the demand description text, the business type and the computing power resource parameter of user demand are parsed;The agent gateway is used to obtain corresponding computing power resource node in the agent proxy computing power resource pool according to the business type and the computing power resource parameter, and the access address of the computing power resource node is generated by configuration.It can recommend and provide suitable computing power resource for user according to the technical means of the present application and business scenario and user demand, effectively meet the use demand of user, improve the use experience of user.

Description

Technical Field

[0001] This invention relates to the field of cloud computing technology, and in particular to a computing resource management method, system, device, medium, and program product. Background Technology

[0002] In existing methods of providing computing power services, users are offered fixed packages of computing power server parameters. After the user selects a package, the corresponding server IP address is queried in the relevant services to provide the service interface for computing power resources.

[0003] However, the inventors have found that the existing technology has at least the following problems: the existing method cannot recommend computing power resource parameters according to business scenarios. For users who do not understand the actual effect of resource parameters, they cannot accurately select server parameter packages according to their own business scenarios when faced with several packages with fixed parameters provided by computing power providers. In some cases, users only know the natural language expression of computing power requirements. In this case, users may choose the wrong computing power resources, resulting in computing power resources that do not meet the user's actual needs. Summary of the Invention

[0004] The purpose of this invention is to provide a computing resource management method, system, device, medium, and program product that can recommend and provide appropriate computing resources to users based on business scenarios and user needs, effectively meeting user needs and improving user experience.

[0005] To achieve the above objectives, embodiments of the present invention provide a computing power resource management system, including a resource demand platform, a proxy gateway, and a computing power resource pool;

[0006] The resource demand platform is used to obtain the demand description text input by the user; and to parse the demand description text to obtain the business type and computing resource parameters of the user's demand.

[0007] The proxy gateway is used to obtain the corresponding computing resource node from the computing resource pool it proxies, configure it, and generate the access address of the computing resource node according to the service type and the computing resource parameters.

[0008] As an improvement to the above solution, the resource demand platform includes a resource demand platform interface, a business type recommendation module, and a parameter parsing module;

[0009] The resource requirement platform interface is used to obtain the requirement description text input by the user; wherein, the requirement description text is either a business type description text or a resource parameter description text;

[0010] When the requirement description text is a business type description text, the business type recommendation module is used to parse the business type description text to obtain the business type; and,

[0011] Search a preset business knowledge base to obtain the computing power resource parameters corresponding to the business type; wherein, the preset business knowledge base records the correspondence between business types and computing power resource parameters;

[0012] When the demand description text is a resource parameter description text, the parameter parsing module is used to parse the resource parameter description text, obtain the computing power resource parameters, and send them to the business type recommendation module;

[0013] The business type recommendation module is also used to search the business knowledge base to obtain the business type corresponding to the computing power resource parameters.

[0014] As an improvement to the above solution, the resource demand platform further includes an address encoding module, and the proxy gateway includes a computing power resource configuration module;

[0015] The address encoding module is used to determine the address of the corresponding proxy gateway according to the service type, so as to establish a communication link between the resource demand platform and the proxy gateway;

[0016] The business type recommendation module is also used to send the computing power resource parameters to the computing power resource configuration module;

[0017] The computing power resource configuration module is used to match the corresponding computing power resource node in the proxy computing power resource pool according to the computing power resource parameters, and generate the access address of the computing power resource node.

[0018] As an improvement to the above solution, the resource demand platform interface is also used to display the demand description text, the business type, the computing power resource parameters, and the access address of the computing power resource node.

[0019] As an improvement to the above solution, the proxy gateway also includes a resource monitoring module and a computation offloading module;

[0020] The resource monitoring module is used to monitor the operating status information of each computing power resource node; wherein, the operating status information is either a normal operating status or an abnormal operating status;

[0021] The computing offloading module is used to acquire computing resource nodes that are not operating normally as nodes to be offloaded; and to allocate the configured services of the nodes to be offloaded to other available computing resource nodes.

[0022] As an improvement to the above solution, the calculation unloading module is specifically used for:

[0023] Identify computing resource nodes that are not operating normally and designate them as nodes to be unloaded;

[0024] Analyze the configured services of the node to be uninstalled and denote them as uninstallation tasks;

[0025] Based on the service type of the unloading task, other allocable computing power resource nodes are determined as allocable computing power resource nodes, and the addresses of the allocable computing power resource nodes are obtained.

[0026] An uninstallation notification data packet is generated and sent to each of the allocable computing power resource nodes according to the address, so that the allocable computing power resource nodes can obtain their assigned uninstallation tasks through the uninstallation notification data packet and return a response result.

[0027] As an improvement to the above scheme, the unloading notification data consists of data fields corresponding to each of the allocatable computing power resource nodes, and each of the data fields includes an encoding bit, a first interface bit, and a second interface bit.

[0028] The encoding bit consists of an encoding segmentation identifier and the encoding value of the unloading task. The first interface bit is used to provide a data interface for the allocable computing power resource node to obtain the unloading task assigned to it. The second interface bit is used to provide an interface for the allocable computing power resource node to return a response result after obtaining the unloading task.

[0029] This invention also provides a method for managing computing resources, including:

[0030] Obtain the user's input description of their requirements;

[0031] Based on the requirement description text, the user's required business type and computing resource parameters are parsed to obtain the data.

[0032] Configure the corresponding computing resource nodes according to the business type and the computing resource parameters, and generate the access address of the computing resource nodes.

[0033] As an improvement to the above solution, the requirement description text is either a business type description text or a resource parameter description text;

[0034] The step of parsing the user's service type and computing resource parameters based on the requirement description text includes:

[0035] When the requirement description text is a business type description text, the business type description text is parsed to obtain the business type;

[0036] Search a preset business knowledge base to obtain the computing power resource parameters corresponding to the business type; wherein, the preset business knowledge base records the correspondence between business types and computing power resource parameters;

[0037] When the demand description text is a resource parameter description text, the resource parameter description text is parsed to obtain the computing power resource parameters;

[0038] Search the business knowledge base to obtain the business type corresponding to the computing power resource parameters.

[0039] As an improvement to the above solution, the step of configuring corresponding computing resource nodes according to the service type and the computing resource parameters, and generating the access address of the computing resource nodes, includes:

[0040] Determine the corresponding computing resource pool based on the business type;

[0041] Match the corresponding computing resource nodes in the computing resource pool according to the computing resource parameters;

[0042] Generate the access address of the computing resource node.

[0043] As an improvement to the above solution, the method further includes:

[0044] Monitor the operational status information of each computing resource node; wherein the operational status information is either a normal operating status or an abnormal operating status;

[0045] Identify computing resource nodes that are not operating normally and designate them as nodes to be unloaded;

[0046] The configured services of the node to be unloaded are allocated to other available computing resource nodes.

[0047] This invention also provides a computing resource management device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the computing resource management method as described in any of the above embodiments.

[0048] This invention also provides a computer-readable storage medium, which includes a stored computer program, wherein the computer program, when running, controls the device where the computer-readable storage medium is located to execute the computing resource management method as described in any of the above embodiments.

[0049] This invention also provides a computer program product, which includes a computer program or computer instructions. When the computer program or computer instructions are executed by a processor, they implement the computing resource management method as described above.

[0050] Compared with existing technologies, the computing power resource management method, system, device, medium, and program products disclosed in this invention analyze computing power requirements based on natural language. This allows users to describe the required business types or resource parameters using natural language. By combining user needs with experience in professional computing power resource application scenarios, the invention makes bidirectional inferences between resource parameters and business types, recommending and providing users with suitable computing power resources. This achieves more comprehensive computing power requirement analysis and more efficient resource utilization. It effectively solves the problems of insufficient resource scheduling flexibility and scalability in the use of computing power resources due to the relatively fixed computing power provision methods in existing technologies. It can meet various application scenarios, effectively satisfy user needs, and improve user experience. Attached Figure Description

[0051] Figure 1 This is a first structural schematic diagram of the computing resource management system provided in an embodiment of the present invention;

[0052] Figure 2 This is a schematic diagram of the second structure of the computing resource management system in an embodiment of the present invention;

[0053] Figure 3 This is a first schematic diagram of the resource demand platform interface in an embodiment of the present invention;

[0054] Figure 4 This is a schematic diagram illustrating the working principle of the address encoding module in an embodiment of the present invention;

[0055] Figure 5 This is a schematic diagram of the first workflow of the computing resource management system in an embodiment of the present invention;

[0056] Figure 6 This is a second schematic diagram of the resource demand platform interface in an embodiment of the present invention;

[0057] Figure 7 This is a third schematic diagram of the resource demand platform interface in an embodiment of the present invention;

[0058] Figure 8 This is a schematic diagram of the third structure of the computing resource management system in an embodiment of the present invention;

[0059] Figure 9 This is a first schematic diagram of the unloading notification data packet in an embodiment of the present invention;

[0060] Figure 10 This is a second schematic diagram of the unloading notification data packet in an embodiment of the present invention;

[0061] Figure 11 This is a schematic diagram of the first process of the computing resource management method provided in the embodiment of the present invention;

[0062] Figure 12 This is a schematic diagram of the second process of the computing resource management method in an embodiment of the present invention. Detailed Implementation

[0063] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0064] In the description of this application, it should be understood that the terms "center", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application.

[0065] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0066] See Figure 1 This is a first structural diagram of a computing resource management system provided in an embodiment of the present invention. The embodiment of the present invention provides a computing resource management system 10, including a resource demand platform 11, a proxy gateway 12, and a computing resource pool 13; wherein,

[0067] The resource demand platform 11 is used to obtain the demand description text input by the user; and to parse the demand description text to obtain the business type and computing resource parameters of the user's demand.

[0068] The proxy gateway 12 is used to obtain the corresponding computing resource node from the computing resource pool it proxies according to the service type and the computing resource parameters, configure it, and generate the access address of the computing resource node.

[0069] In this embodiment of the invention, the computing resource management system 10 consists of three parts. The resource demand platform 11 provides an interactive entry point based on user-submitted business scenarios or resource parameter descriptions, parses the user's requested business type and computing resource parameters from the descriptions, and provides a platform for subsequent feedback on computing resource interfaces. The resource demand platform 11 exists in the form of a server device.

[0070] The proxy gateway 12 is used to proxy several computing resource pools, perform uplink and downlink data forwarding control, and manage the computing resource pools. It can obtain and configure the corresponding computing resource nodes in the proxied computing resource pools according to the service type and computing resource parameters, and generate the access addresses of the computing resource nodes. The proxy gateway 12 can be a server or router, etc., and is a collection of devices supporting gateway and computing functions.

[0071] The computing resource pool 13 is a resource cluster formed according to different business types. The unit providing computing resources in the computing resource pool is a computing resource node, and each computing resource pool includes several computing resource nodes. The computing resource pool 13 is a server cluster and also includes devices with routing capabilities for communication.

[0072] The computing resource nodes, as nodes providing various capabilities, can exist in the form of computers, servers, virtual machines, etc., and support IPv6 (Internet Protocol Version 6) and SRv6 (Segment Routing IPv6). The computing resource parameters are parameters used to describe the resource status of the computing node, including but not limited to memory, disk space, operating system type, number of cores, etc.

[0073] By employing the technical means of this invention, computing power demand analysis is performed based on natural language. This allows users to describe the required business types or resource parameters using natural language. By combining user needs with experience in professional computing power resource application scenarios, bidirectional inference is made between resource parameters and business types. This enables the recommendation and provision of suitable computing power resources to users, achieving more comprehensive computing power demand analysis and more efficient resource utilization. It effectively solves the problem in existing technologies where the computing power provision method is relatively fixed, resulting in insufficient flexibility and scalability in resource scheduling during computing power resource use. It can meet various application scenarios, effectively satisfy user needs, and improve user experience.

[0074] As a preferred embodiment, this invention further optimizes the structure of the computing resource management system 10 based on the above embodiments. The resource demand platform 11 includes a resource demand platform interface, a business type recommendation module, and a parameter parsing module.

[0075] The resource requirement platform interface is used to obtain the requirement description text input by the user; wherein, the requirement description text is either a business type description text or a resource parameter description text;

[0076] When the requirement description text is a business type description text, the business type recommendation module is used to parse the business type description text to obtain the business type; and,

[0077] Search a preset business knowledge base to obtain the computing power resource parameters corresponding to the business type; wherein, the preset business knowledge base records the correspondence between business types and computing power resource parameters;

[0078] When the demand description text is a resource parameter description text, the parameter parsing module is used to parse the resource parameter description text, obtain the computing power resource parameters, and send them to the business type recommendation module;

[0079] The business type recommendation module is also used to search the business knowledge base to obtain the business type corresponding to the computing power resource parameters.

[0080] See Figure 2 This is a schematic diagram of the second structure of the computing resource management system in this embodiment of the invention. The resource demand platform includes a resource demand platform interface, a parameter parsing module, and a business type recommendation module.

[0081] The resource requirement platform interface is used to obtain the requirement description text input by the user. It supports two types of description input by the user: one is the text content describing the business type, that is, the business type description text, and the other is the text content describing the resource parameters, that is, the resource parameter description text.

[0082] The parameter parsing module is a module that parses computing power resource parameters from resource parameter description text.

[0083] The business type recommendation module is a module that parses the business type from the business type description text; and determines the corresponding business type and characteristics based on the computing power resource parameters, or provides computing power resource parameter recommendations based on the business type.

[0084] In practical application, when the resource demand platform interface receives a business type description text, it sends it to the business type recommendation module. The business type recommendation module matches the description text with existing business types, including but not limited to "image recognition model inference business," "speech recognition model training business," "video encoding and decoding business," "software development management business," "enterprise employee office business," and "web service deployment business." Each business type corresponds to a computing resource pool, which is an independent SRv6 domain. This determines the business type required by the user. Then, based on the business type, corresponding computing resource parameters are recommended. Specifically, this can be achieved by determining the numerical range of the required parameters based on data restriction terms. For business types and corresponding data restriction terms, the business type recommendation module queries a preset business knowledge base. This business knowledge base contains data on business type application scenarios and the computing resource parameters used. By comparing the computing resource parameter requirements of a certain business type under specific data restriction terms using this business knowledge base, a set of computing resource parameters is recommended.

[0085] When the resource demand platform interface receives the business type description text, it sends it to the parameter parsing module. After the resource parameter analysis range in the parameter parsing module is analyzed, various computing power resource parameters required by the user are parsed and sent to the business type recommendation module. The business type recommendation module also searches for the computing power resource parameter information in the business knowledge base and analyzes the business type to determine the business scenario type corresponding to the user description.

[0086] More preferably, see Figure 3 This is a first schematic diagram of the resource demand platform interface in an embodiment of the present invention. The resource demand platform interface is also used to display the demand description text, the business type, the computing power resource parameters, and the access address of the computing power resource node.

[0087] Specifically, the resource demand platform interface exists in the form of a web page, including page elements such as resource description, resource parameters, business type, resource node address, and resource entry point. The resource description section is used to input text describing the user's demand for computing resources, that is, the business type description text or resource parameter description text.

[0088] The resource parameter section is used to display the names of the computing power resource parameters (i.e., "parameter 1", "parameter 2", etc. in the figure) and their corresponding resource parameter values ​​as reflected in the resource parameter description text; or it is used to display the names of the computing power resource parameters and their corresponding resource parameter values ​​recommended after the business type recommendation module analyzes the business type description text.

[0089] The business type section is used to display the business type reflected in the business type description text, or the business type obtained by the parameter parsing module after parsing the resource parameters from the resource parameter description text input by the user, and then by the business type recommendation analysis.

[0090] The resource node address is the IP address of the computing power resource node returned by the proxy gateway after configuring the resources; the resource entry is the entry point for users to access resources, making it easy for users to click and enter the computing power resource usage page.

[0091] In a preferred embodiment, the resource demand platform 11 further includes an address encoding module, and the proxy gateway 12 includes a computing power resource configuration module.

[0092] The address encoding module is used to determine the address of the corresponding proxy gateway according to the service type, so as to establish a communication link between the resource demand platform and the proxy gateway;

[0093] The business type recommendation module is also used to send the computing power resource parameters to the computing power resource configuration module;

[0094] The computing power resource configuration module is used to match the corresponding computing power resource node in the proxy computing power resource pool according to the computing power resource parameters, and generate the access address of the computing power resource node.

[0095] See Figure 2 In this embodiment of the invention, the address encoding module is used to calculate the corresponding IPv6 address allocation of the proxy gateway based on the service type required by the user by running an address encoding conversion method, thereby establishing a communication link between the resource demand platform 11 and the proxy gateway 12. The computing power resource configuration module in the proxy gateway 12 is used to configure computing power resource nodes corresponding to the computing power resource parameters required by the user in the computing power resource pool, forming computing power resource nodes and configuring IP addresses and networks.

[0096] Preferably, the address encoding module is specifically used to: encode the text of the business type into binary data, and then convert the binary data into a hexadecimal address to obtain the address of the corresponding proxy gateway, thereby establishing a communication link between the resource demand platform and the proxy gateway.

[0097] Specifically, the address encoding module receives the service type text from the service type recommendation module, runs the address encoding conversion method, and obtains the IPv6 address of the proxy gateway corresponding to the service type after processing by the address encoding conversion method, so as to establish a communication link between the resource demand platform and the specific gateway.

[0098] The address encoding conversion method includes a set of encoding methods, divided into two types: text encoding method and address encoding method. The text encoding method is used to encode business type text into binary form, while the address encoding method is used to convert binary data into IPv6 addresses (hexadecimal). Its types cover various encoding formats, including but not limited to text, images, and audio. For multimedia encoding methods such as images and audio, the input is the main data (e.g., RGB data and grayscale data in an image), rather than metadata describing file formats, parameters, etc.

[0099] Alternatively, the set of encoding methods is represented as follows:

[0100]

[0101] Choosing one of the text encoding methods and one of the address encoding methods to form the address encoding conversion function, it can be expressed as: f (Business type, text encoding method, address encoding method);

[0102] See Figure 4 This is a schematic diagram illustrating the working principle of the address encoding module in this embodiment of the invention. The address encoding conversion function incorporates the metadata required by the address encoding method to provide necessary parameters during the encoding process. Furthermore, the address encoding conversion function includes an address encoding method adapter component to convert binary data into the input format of the address encoding method and extract the output data of the address encoding method at appropriate stages. After base conversion and length adjustment steps, an IPv6 address is formed. For example, the input data for the JPEG encoding method is in 8×8 grayscale decimal data units, including discrete cosine transform, matrix quantization, Z-coding, DC / AC separation, and entropy encoding. The adapter component selects the decimal form output result of the Z-coding step and converts it into hexadecimal data. Based on the format of a 128-bit IPv6 address, any excessive length is truncated, and insufficient length is padded with zeros.

[0103] Further, see Figure 5 This is a schematic diagram of the first workflow of the computing resource management system in this embodiment of the invention. This embodiment further explains the computing resource provision process. When a user enters the resource demand platform interface to select computing resources, the resource demand platform, proxy gateway, and resource pool will enter the computing resource provision process.

[0104] Specifically, when a user clearly knows the computing power resource parameters required for the business, the user enters a resource parameter description text in the resource description section of the resource demand platform interface. When a user only knows the business type scenario, the user enters a business type description text in the resource description section of the resource demand platform interface.

[0105] For the business type description text, the business type recommendation module matches it with existing business types in the business knowledge base and determines the required computing resource parameter value range based on data-limiting words such as "low pixel" and "high definition." For the resource parameter description text, the parameter parsing and analysis obtains the specific computing resource parameters, and the business type recommendation module matches the corresponding business type. The resource demand platform interface displays the business type and the computing resource parameters respectively.

[0106] The service type recommendation module passes the service type to the address encoding module, which calculates the IP address of the proxy gateway. The resource demand platform sends the computing power resource parameters and the computing power resource address to the computing power resource configuration module in the proxy gateway. The computing power resource configuration module configures computing power resource nodes that match the computing power resource parameters in the corresponding computing power resource pool, and configures their IP addresses and networks. After successful configuration, the computing power resource configuration module returns the result to the resource demand platform. The resource demand platform interface displays the computing power resource address at the resource node address and provides a web computing power resource entry point for users to click and use the resources.

[0107] The embodiments of the present invention explain the process of providing computing resources using specific application scenarios as follows:

[0108] When a user selects computing resources on the resource demand platform interface, the resource demand platform, proxy gateway, and resource pool will enter the computing resource provision process. The specific steps are as follows:

[0109] See Figure 6 This is a second schematic diagram of the resource demand platform interface in this embodiment of the invention. The user enters the business type description text "WAV format speech recognition model training business" on the resource demand platform interface. The business type recommendation module extracts the business type description from the text, matches the business type "speech recognition model training business", and queries the business knowledge base according to the data restriction word "WAV format" to obtain the common computing power resource parameters in this scenario as "GPU, 8-core processor, 64G memory, 1T disk, Linux system". The web page displays the corresponding business type and computing power resource parameters.

[0110] See Figure 7This is a third schematic diagram of the resource demand platform interface in this embodiment of the invention. The service type recommendation module passes the user's service type "speech recognition model training service" to the address encoding module. The address encoding module calculates the IP address "2001:DB8::1" of its proxy gateway based on the service type "speech recognition model training service" using an address encoding conversion method. The resource demand platform sends the computing power resource parameters and computing power resource address to the computing power resource configuration module in the proxy gateway. The computing power resource configuration module configures virtual machines that conform to the computing power resource parameters in the computing power resource pool to form computing power resource nodes, and configures the IP address "2001:DB8::45" and network. After successful configuration, the computing power resource configuration module returns the result to the resource demand platform. The resource demand platform interface displays the computing power resource address and provides a web entry point for using the resources.

[0111] The technical means employed in this invention involve parsing computing power requirements based on natural language, supporting both business type descriptions and resource parameter descriptions. This facilitates users in describing their required business types or resource parameters using natural language. Through a business knowledge base and combined with experience in professional computing power resource application scenarios, bidirectional inference is performed between resource parameters and business types. For users who clearly know their computing power resource parameters, matching computing power resources can be provided. For users who can only provide a description of their business type scenario, more scenario-related information can be provided to assist their selection. This leads to the recommendation and provision of suitable computing power resources, achieving more comprehensive computing power requirement analysis and more efficient resource utilization.

[0112] As a preferred embodiment, the present invention is further implemented based on any of the above embodiments. It should be noted that in the prior art, users select a purchase period when choosing computing power resources, which is generally several months or even several years. For the computing power resources that have been purchased, users may experience insufficient computing, storage, and professional processing capabilities due to business adjustments and changes in application resource usage. When the user's actual computing power demand exceeds the purchased computing power resources, the application cannot run normally, or even cause service crashes.

[0113] To address this issue, in this embodiment of the invention, the proxy gateway 12 further includes a resource monitoring module and a computation offloading module.

[0114] The resource monitoring module is used to monitor the operating status information of each computing power resource node; wherein, the operating status information is either a normal operating status or an abnormal operating status;

[0115] The computing offloading module is used to acquire computing resource nodes that are not operating normally as nodes to be offloaded; and to allocate the configured services of the nodes to be offloaded to other available computing resource nodes.

[0116] See Figure 8 This is a schematic diagram of the third structure of the computing resource management system in this embodiment of the invention. The proxy gateway 12 further includes a resource monitoring module and a computation unloading module. The resource monitoring module is used to monitor the application operation status in the computing resource nodes, and the computation unloading module is used to obtain the operation status information of the computing resource nodes and manage the computation unloading process of the computing resource nodes. For example, when the computing resources of a certain computing resource node are insufficient, and it is detected that the computing resource node has entered a bottleneck state or cannot provide services in some resource parameters or professional applications, then it is considered that the computing resource node is in an abnormal operation state, this resource node is designated as a node to be unloaded, and computation unloading is started.

[0117] In a preferred embodiment, the calculation unloading module is specifically used for:

[0118] Identify computing resource nodes that are not operating normally and designate them as nodes to be unloaded;

[0119] Analyze the configured services of the node to be uninstalled and denote them as uninstallation tasks;

[0120] Based on the service type of the unloading task, other allocable computing power resource nodes are determined as allocable computing power resource nodes, and the addresses of the allocable computing power resource nodes are obtained.

[0121] An uninstallation notification data packet is generated and sent to each of the allocable computing power resource nodes according to the address, so that the allocable computing power resource nodes can obtain their assigned uninstallation tasks through the uninstallation notification data packet and return a response result.

[0122] Preferably, the unloading notification data consists of data fields corresponding to each of the allocatable computing power resource nodes, and each data field includes an encoding bit, a first interface bit, and a second interface bit.

[0123] The encoding bit consists of an encoding segmentation identifier and the encoding value of the unloading task. The first interface bit is used to provide a data interface for the allocable computing power resource node to obtain the unloading task assigned to it. The second interface bit is used to provide an interface for the allocable computing power resource node to return a response result after obtaining the unloading task.

[0124] In this embodiment of the invention, the resource monitoring module has mastered the SID (Segment Identifier) ​​and availability status of each computing resource node in the SRv6 domain, i.e. whether it can act as a resource provider to support computing offload tasks. It also has mastered the offload task types of different resource nodes, including but not limited to intensive computing, data storage, high concurrency, GPU acceleration, image processing, and audio processing. These offload task types are the basic units of computing offload tasks, and each type has a corresponding type code.

[0125] When the resource monitoring module detects that the computing power of a computing power resource node is insufficient, it will calculate the type of uninstallation task to be uninstalled and the resource parameter conditions that this application type needs to meet based on the program running status. Then, it will compare and select the available resource nodes and resource parameter configurations in the current resource pool. The SID of the available resource node that meets the conditions will be transmitted to the computing uninstallation module along with its uninstallation task type code.

[0126] The unloading notification data packet is sent by the node to be unloaded. The Segment list in the Segment Routing Header (SRH) of the data packet encapsulates the SIDs of several allocable resource nodes provided by the resource monitoring module, which serve as the forwarding path for this data packet. During the data packet forwarding process, each allocable resource node reads the computation unloading related information carried in the data packet, including the computation task data interface (i.e., the first interface) and the computation result receiving interface (i.e., the second interface). Both interfaces are open on the node to be unloaded when the computation unloading task is generated and closed when the computation unloading task ends. The computation task data interface is used by the resource node to obtain the unloading task data of the node to be unloaded as the input data source for the unloading task. The computation result receiving interface is used by the resource node to send the response result back to the node to be unloaded after completing the acquisition of the unloading task.

[0127] See Figure 9This is a first schematic diagram of the unloading notification data packet in an embodiment of the present invention. During the forwarding process of the unloading notification data packet, each intermediate computing resource node that passes through it needs to read the computation unloading related information in the data packet, store the information in the TLV (Type Length Value) structure in the SRH, and implement hop-by-hop parsing through an extended form. The data field of each allocable resource node consists of "encoding", "first interface x-1", and "second interface x-2". "Encoding" consists of two parts: "encoding segmentation identifier" and "encoding value". The encoding segmentation identifier is used to separate the information of different computing resource nodes, so as to inform each computing resource node that the next computing resource node's computation unloading related information follows, avoiding parsing errors. "Encoding value" represents the encoding of the unloading task. In the TLV, the computation unloading related information also exists in an "other" structure, including the Type and Length fields in the TLV. Since the total length of the TLV may have an 8-bit integer multiple requirement, a "padding" part may be needed at the end. The computation task data interface x-1 and the computation result receiving interface x-2 occupy 2 to 5 bits, and their order is not required. The figure shows one implementation. The order of interface identification during the parsing of computing resource nodes is agreed upon with the nodes to be unloaded, and is not required here.

[0128] This invention further explains the process of expanding computing resources. During the use of computing resources by a user, if the resource node providing the service encounters bottlenecks or is unable to provide service in certain resource parameters or professional applications, the proxy gateway will assist the computing resource pool in providing computation offloading services. The specific steps are as follows:

[0129] The resource monitoring module in the proxy gateway 12 detects insufficient computing resources at a certain address. The computation unloading module identifies this as a node to be unloaded, analyzes the service types within the node to be unloaded, and determines the services requiring computation unloading as unloading tasks. The unloading tasks are analyzed to obtain the allocable computing resource nodes and their addresses corresponding to different service types. The allocable computing resource nodes and their unloading task types are communicated to the node to be unloaded. The node to be unloaded encapsulates a computation unloading notification data packet carrying an SRv6 extension header. The segment list in the extension header of the unloading notification data packet contains the IPv6 address of the computing resource node, serving as the forwarding path for this data packet. The TLV of the unloading notification data packet carries the unloading task type, computation task data interface, and computation result receiving interface. The unloading notification data packet is sent by the computation unloading module to each computing resource node according to the forwarding path. The computing resource node retrieves the unloading task data from the node to be unloaded based on the computation task data interface provided in the unloading notification data packet, completes the computation unloading task, and sends a response result to the computation result receiving interface. Once the node to be unloaded receives all the response results, it informs the computing unloading module, indicating that the computing resource expansion is complete.

[0130] The following embodiments of the present invention explain the process of expanding computing resources using specific application scenarios:

[0131] When the resource monitoring module in the proxy gateway detects that the high-concurrency application and image processing application on computing resource node N1 are in an abnormal operating state due to performance bottlenecks, it marks node N1 as a node to be unloaded. The resource monitoring module analyzes the service type in the virtual machine and determines that the service types that need to be unloaded are "high concurrency" and "image processing," which are then designated as unloading tasks. The computing unloading module in the proxy gateway analyzes the unloading tasks, obtains the resource node P1 corresponding to the "high concurrency" unloading task type and the resource node P2 corresponding to the "image processing" unloading task type, and informs the unloading node N1 of the resource nodes and their computing task types.

[0132] See Figure 10 This is a second schematic diagram of the unloading notification data packet in an embodiment of the present invention. The node to be unloaded, N1, learns that the End SIDs of computing resource nodes P1 and P2 are "2001:DB8:4::4" and "2001:DB8:6::6" respectively, and determines that the SegmentList is represented as:

[0133]

[0134] Simultaneously, a TLV option is added. The encoding value for the "High Concurrency" offload task type is "02", with "A" added as an encoding segmentation identifier. "8001" is the computation task data interface, and "8002" is the computation result receiving interface. The encoding value for the "Image Processing" offload task type is "03", with "A" added as an encoding segmentation identifier. "8088" is the computation task data interface, and "8080" is the computation result receiving interface. Node N1 encapsulates an SRv6 extension header containing the above data to obtain an offload notification data packet, which is then sent to the computation offload module. The offload notification data packet is sent by the computation offload module to computing resource nodes P1 and P2 sequentially according to the forwarding path. Node P1 parses the TLV content in the computation offload notification data packet to obtain the computation task data interface "8001", retrieves the corresponding offload task data from node N1, completes the computation offload task, and sends the response result to the computation result receiving interface "8002". Node N1 then closes the "8001" and "8002" interfaces. Then, node P1 forwards the data packet to node P2. Node P2 parses the TLV content in the computation unloading notification data packet to obtain the computation task data interface "8088". It retrieves the unloading task data from node N1, completes the computation unloading task, and sends the response result to the computation result receiving interface "8080". Node N1 closes the "8088" and "8080" interfaces. The node N1 to be unloaded receives all the response results and informs the computation unloading module that the computational resource expansion is complete.

[0135] By employing the technical means of this invention, if a user's purchased computing resources do not meet their actual computing needs during the use of computing resources, resources can be expanded according to the user's requirements. Professional expansion strategies are provided to ensure the normal operation of the user's business applications without the need to purchase additional computing resources and migrate data and applications. Specifically, computing resource expansion is based on SRv6. During the use of computing resources, the running status of node applications can be analyzed in real time. For scalable unloading business types, SRv6 computing unloading notification data packets are encapsulated to notify the unloading task, enabling other nodes to cooperate with the node to be unloaded to complete the current task. Communication between multiple nodes is completed through a single data packet forwarding path, which is convenient and efficient. Simultaneously, tasks are allocated for different unloading task types to ensure the effective completion of unloading tasks, avoiding situations where computing resource bottlenecks and reasons are not visible to users, which can hinder effective resource expansion.

[0136] See Figure 11 This is a first flowchart of the computing resource management method provided in the embodiments of the present invention. The embodiments of the present invention also provide a computing resource management method, which is executed by the computing resource management system 10 in the above embodiments, and includes steps S11 to S13:

[0137] S11. Obtain the user's input description of the requirements;

[0138] S12. Based on the requirement description text, parse out the user's required business type and computing resource parameters;

[0139] S13. Configure the corresponding computing resource nodes according to the service type and the computing resource parameters, and generate the access address of the computing resource nodes.

[0140] In a preferred embodiment, the requirement description text is either a business type description text or a resource parameter description text. Therefore, step S12, which involves parsing the requirement description text to obtain the user's required business type and computing resource parameters, includes steps S121 to S124:

[0141] S121. When the requirement description text is a business type description text, parse the business type description text to obtain the business type.

[0142] S122. Search a preset business knowledge base to obtain the computing power resource parameters corresponding to the business type; wherein, the preset business knowledge base records the correspondence between business types and computing power resource parameters;

[0143] S123. When the demand description text is a resource parameter description text, parse the resource parameter description text to obtain the computing power resource parameters;

[0144] S124. Search the business knowledge base to obtain the business type corresponding to the computing power resource parameters.

[0145] In a preferred embodiment, step S13, namely configuring the corresponding computing resource nodes according to the service type and the computing resource parameters, and generating the access address of the computing resource nodes, includes steps S131 to S133:

[0146] S131. Determine the corresponding computing resource pool according to the business type;

[0147] S132. Match the corresponding computing resource nodes in the computing resource pool according to the computing resource parameters;

[0148] S133. Generate the access address of the computing power resource node.

[0149] The technical means employed in this invention involve parsing computing power requirements based on natural language, supporting both business type descriptions and resource parameter descriptions. This facilitates users in describing their required business types or resource parameters using natural language. Through a business knowledge base and combined with experience in professional computing power resource application scenarios, bidirectional inference is performed between resource parameters and business types. For users who clearly know their computing power resource parameters, matching computing power resources can be provided. For users who can only provide a description of their business type scenario, more scenario-related information can be provided to assist their selection. This leads to the recommendation and provision of suitable computing power resources, achieving more comprehensive computing power requirement analysis and more efficient resource utilization.

[0150] See Figure 12 This is a schematic diagram of the second process of the computing resource management method in this embodiment of the invention. This embodiment further implements the method based on the above embodiments, and the method further includes steps S21 to S23:

[0151] S21. Monitor the operating status information of each computing resource node; wherein, the operating status information is either a normal operating status or an abnormal operating status;

[0152] S22. Obtain computing resource nodes that are not in normal operating condition and designate them as nodes to be unloaded;

[0153] S23. Allocate the configured services of the node to be unloaded to other allocable computing resource nodes.

[0154] Step S23, namely, allocating the configured services of the node to be unloaded to other allocable computing resource nodes, includes steps S231 to S234:

[0155] S231. Obtain computing resource nodes that are not in normal operating condition and designate them as nodes to be unloaded;

[0156] S232. Analyze the configured services of the node to be uninstalled and denote them as uninstallation tasks;

[0157] S233. Determine other allocable computing power resource nodes based on the service type of the unloading task, and obtain the addresses of the allocable computing power resource nodes.

[0158] S234. Generate an uninstallation notification data packet and send the uninstallation notification data packet to each of the allocable computing power resource nodes according to the address, so that the allocable computing power resource nodes can obtain their assigned uninstallation tasks through the uninstallation notification data packet and return a response result.

[0159] Preferably, the unloading notification data consists of data fields corresponding to each of the allocatable computing power resource nodes, and each data field includes an encoding bit, a first interface bit, and a second interface bit.

[0160] The encoding bit consists of an encoding segmentation identifier and the encoding value of the unloading task. The first interface bit is used to provide a data interface for the allocable computing power resource node to obtain the unloading task assigned to it. The second interface bit is used to provide an interface for the allocable computing power resource node to return a response result after obtaining the unloading task.

[0161] By employing the technical means of this invention, if the computing resources already purchased by a user do not meet the actual computing power requirements during the user's use of computing resources, resources can be expanded according to the user's needs, and professional expansion strategies can be provided to ensure the normal operation of the user's business applications, without the need to repurchase other computing resources and migrate data and applications.

[0162] It should be noted that all process steps of the computing resource management method provided in the embodiments of the present invention are executed by a computing resource management system of the above embodiments. The working principles and beneficial effects of the two correspond one-to-one, so they will not be described again.

[0163] This invention also provides a computing resource management device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the computing resource management method as described in any of the above embodiments.

[0164] It should be noted that the computing resource management device provided in this embodiment of the invention is used to execute all the process steps of the computing resource management method in the above embodiment. The working principles and beneficial effects of the two are one-to-one, so they will not be described again.

[0165] This invention also provides a computer-readable storage medium, which includes a stored computer program, wherein the computer program, when running, controls the device where the computer-readable storage medium is located to execute the computing resource management method as described in any of the above embodiments.

[0166] This invention also provides a computer program product, which includes a computer program or computer instructions. When the computer program or computer instructions are executed by a processor, they implement the computing resource management method as described in any of the above embodiments.

[0167] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0168] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A computing resource management system, characterized in that, This includes a resource demand platform, proxy gateways, and computing resource pools; The resource demand platform is used to obtain the demand description text input by the user; and to parse the demand description text to obtain the business type and computing resource parameters of the user's demand. The proxy gateway is used to obtain the corresponding computing resource node from the computing resource pool it proxies, configure it, and generate the access address of the computing resource node according to the service type and the computing resource parameters. The resource demand platform includes a resource demand platform interface, a business type recommendation module, and a parameter parsing module; The resource requirement platform interface is used to obtain the requirement description text input by the user; wherein, the requirement description text is either a business type description text or a resource parameter description text; When the requirement description text is a business type description text, the business type recommendation module is used to parse the business type description text to obtain the business type; and, Search a preset business knowledge base to obtain the computing power resource parameters corresponding to the business type; wherein, the preset business knowledge base records the correspondence between business types and computing power resource parameters; When the demand description text is a resource parameter description text, the parameter parsing module is used to parse the resource parameter description text, obtain the computing power resource parameters, and send them to the business type recommendation module; The business type recommendation module is also used to search the business knowledge base to obtain the business type corresponding to the computing power resource parameters. The resource demand platform also includes an address encoding module, and the proxy gateway includes a computing resource configuration module. The address encoding module is used to determine the address of the corresponding proxy gateway according to the service type, so as to establish a communication link between the resource demand platform and the proxy gateway; The business type recommendation module is also used to send the computing power resource parameters to the computing power resource configuration module; The computing power resource configuration module is used to match the corresponding computing power resource node in the proxy computing power resource pool according to the computing power resource parameters, and generate the access address of the computing power resource node.

2. The computing resource management system as described in claim 1, characterized in that, The resource demand platform interface is also used to display the demand description text, the business type, the computing power resource parameters, and the access address of the computing power resource node.

3. The computing resource management system as described in any one of claims 1 to 2, characterized in that, The proxy gateway also includes a resource monitoring module and a computation offloading module; The resource monitoring module is used to monitor the operating status information of each computing power resource node; wherein, the operating status information is either a normal operating status or an abnormal operating status; The computing offloading module is used to acquire computing resource nodes that are not operating normally as nodes to be offloaded; and to allocate the configured services of the nodes to be offloaded to other available computing resource nodes.

4. The computing resource management system as described in claim 3, characterized in that, The computational unloading module is specifically used for: Identify computing resource nodes that are not operating normally and designate them as nodes to be unloaded; Analyze the configured services of the node to be uninstalled and denote them as uninstallation tasks; Based on the service type of the unloading task, other allocable computing power resource nodes are determined as allocable computing power resource nodes, and the addresses of the allocable computing power resource nodes are obtained. An uninstallation notification data packet is generated and sent to each of the allocable computing power resource nodes according to the address, so that the allocable computing power resource nodes can obtain their assigned uninstallation tasks through the uninstallation notification data packet and return a response result.

5. The computing resource management system as described in claim 4, characterized in that, The uninstallation notification data consists of data fields corresponding to each of the allocable computing power resource nodes, and each of the data fields includes an encoding bit, a first interface bit, and a second interface bit. The encoding bit consists of an encoding segmentation identifier and the encoding value of the unloading task. The first interface bit is used to provide a data interface for the allocable computing power resource node to obtain the unloading task assigned to it. The second interface bit is used to provide an interface for the allocable computing power resource node to return a response result after obtaining the unloading task.

6. A method for managing computing resources, characterized in that, include: Obtain the user's input description of their requirements; Based on the requirement description text, the user's required business type and computing resource parameters are parsed to obtain the data. Configure the corresponding computing resource nodes according to the business type and the computing resource parameters, and generate the access address of the computing resource nodes; The requirement description text is either a business type description text or a resource parameter description text; The step of parsing the user's service type and computing resource parameters based on the requirement description text includes: When the requirement description text is a business type description text, the business type description text is parsed to obtain the business type; Search a preset business knowledge base to obtain the computing power resource parameters corresponding to the business type; wherein, the preset business knowledge base records the correspondence between business types and computing power resource parameters; When the demand description text is a resource parameter description text, the resource parameter description text is parsed to obtain the computing power resource parameters; Search the business knowledge base to obtain the business type corresponding to the computing power resource parameters; The step of configuring corresponding computing resource nodes according to the service type and the computing resource parameters, and generating the access address of the computing resource nodes, includes: Determine the corresponding computing resource pool based on the business type; Match the corresponding computing resource nodes in the computing resource pool according to the computing resource parameters; Generate the access address of the computing resource node.

7. The computing resource management method as described in claim 6, characterized in that, The method further includes: Monitor the operational status information of each computing resource node; wherein the operational status information is either a normal operating status or an abnormal operating status; Identify computing resource nodes that are not operating normally and designate them as nodes to be unloaded; The configured services of the node to be unloaded are allocated to other available computing resource nodes.

8. A computing resource management device, characterized in that, It includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor, when executing the computer program, implements the computing resource management method as described in any one of claims 6 to 7.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to perform the computing resource management method as described in any one of claims 6 to 7.

10. A computer program product, characterized in that, The computer program product includes a computer program or computer instructions, which, when executed by a processor, implement the computing resource management method as described in any one of claims 6 to 7.

Images