Computing resource integration access management method and system, electronic device and medium

Abstract

The present disclosure provides a computing power resource integrated access management method and system, electronic equipment and storage medium to solve the problem that computing power resources cannot be connected. The method comprises: connecting different computing power resources through the center node and each level node of the MEC architecture to perform life cycle management; dividing data connection into data domain and management domain; the data domain is implemented based on MPP distributed data architecture; the management domain completes the storage and operation processing of computing power resources and the management of nodes through computing power resource management; MPP distributed data architecture is used to realize distributed storage, management and calling of data; an electronic map is used to form a computing power resource map according to the demand planning layer and computing power tile of computing power resources; an API interface provided by the map is used to provide a search service; and a path planning API and computing power map data provided by the map are used to generate optimized computing power routing. The present disclosure provides a feasible implementation path for connecting computing power resources.

Description

Technical Field

[0001] This disclosure relates to the field of communication technology, specifically to an integrated access management method for computing resources, an integrated access management system for computing resources, an electronic device, and a computer-readable storage medium. Background Technology

[0002] The core of an intelligent society is "data + computing power + algorithms," where computing power is the foundational platform for intelligence, composed of various computing resources. The core purpose of computing power networks is to provide users with computing power resource services, but its implementation differs from the traditional "cloud computing" approach. Instead, it integrates computing power resources into communication networks, providing computing power resource services that best meet user needs in a more holistic manner.

[0003] The core value of networks is to improve efficiency. The internet has improved the efficiency of human collaboration, and computing networks have emerged to improve the collaborative efficiency of edge computing (device, cloud, and cloud). The difference is that traditional networks directly serve humans, while computing networks directly serve intelligent machines and indirectly serve humans through these machines.

[0004] The computing resources in a computing power network are diverse, widely distributed, and vary significantly in performance. Reliable and secure access to these resources, and the seamless integration of these massive amounts of computing power to provide users with abundant computing resources, has become a crucial aspect of computing power networks. Currently, existing communication networks cannot directly integrate computing resources, and building an independent network for managing and integrating these resources involves significant investment, a long construction period, and high complexity, making it difficult to implement at present. Summary of the Invention

[0005] To at least address the technical problem that existing communication networks cannot directly achieve the interconnection of computing resources, this disclosure provides an integrated access management method for computing resources, an integrated access management system for computing resources, an electronic device, and a computer-readable storage medium. It can utilize the convergence of MEC (Multi-access Edge Computing) and 5G networks, and based on the service interfaces of 5G networks and MEC, provide a feasible implementation path for interconnecting computing resources, realizing the access and management of various computing resources throughout the network.

[0006] Firstly, this disclosure provides an integrated access management method for computing resources, the method comprising:

[0007] Access to different computing resources: Connect different computing resources through the central node and MEC nodes at all levels of the MEC architecture, and complete the lifecycle management of computing resources through MEC;

[0008] Computing resource data integration: The integration of computing resource data is divided into a data domain and a management domain; the data domain is integrated based on the MPP (Massively Parallel Processing) distributed data architecture, providing data storage methods and data organization forms for the integrated computing resource data of the entire network; in the management domain, the storage and operation of computing resources, as well as the management of computing data nodes, are completed through computing resource management.

[0009] Computing power data storage: Adopting an MPP distributed data architecture, while preserving the independence of MEC data nodes at each level, it realizes distributed storage, overall management and data retrieval of data;

[0010] Computing power map implementation: Utilizing existing electronic map technology, layers and computing power tiles are planned according to the demand for computing power resources to form a computing power resource map, providing services for computing power resource search;

[0011] Computing resource search: Utilizing the computing resource map and based on the search API (Application Programming Interface) provided by the map, it offers a variety of search services for computing resources;

[0012] Computing resource routing: Provides optimized network connection routes for computing resource users and selected computing resources, generating optimized computing resource routes based on the path planning API provided by the map and computing resource map data.

[0013] Furthermore,

[0014] The computing resources include the computing resources of the cloud center, the computing resources in the MEC, and the customer computing resources accessed through the MEC.

[0015] The process of connecting different computing resources through the central node and various levels of MEC nodes in the MEC architecture, and managing the lifecycle of computing resources through MEC, includes:

[0016] The computing resources of the cloud center are accessed through the unified cloud management platform of the central node, and an authentication system based on the MEC management system is adopted. The lifecycle management of computing resources is realized through cloud computing network management.

[0017] For the computing resources in the MEC and the customer computing resources accessed through the MEC, based on the four deployment modes of the MEC's ​​distribution network elements and the MEC resource platform, the lifecycle management of the computing resources is realized by the MEC's ​​API gateway after the computing power gateway converts the protocol.

[0018] Furthermore,

[0019] The data interconnection of data domains based on the MPP distributed data architecture includes:

[0020] In terms of data processing, the computing power data processing tasks are distributed in parallel to multiple related nodes through MPP. After the data processing is completed on each node, the results of each part are aggregated to obtain the final result.

[0021] In terms of data structure, the MPP node data of local computing resources in MEC is organized in the form of a data mart.

[0022] Furthermore,

[0023] The management domain includes the computing power resource library management node and the computing power data center node of the central node in the MEC. The computing power resource library management node includes data processing nodes and data storage nodes.

[0024] The management of computing resources in the management domain, including storage and operation processing of computing resources and management of computing data nodes, includes:

[0025] The management of computing power resource libraries in different levels of MEC uses data processing middleware to process the real-time data of computing power resources provided by the MEC computing power gateway, forming MEC computing power mart data, and updating it to the local data nodes of the distributed data architecture MPP.

[0026] The data processing node completes the metadata processing and connects to the query terminal in the MEC computing power gateway to perform computing power resource queries and other operations.

[0027] The data storage nodes complete the storage of computing power data and the execution of SQL (Structured Query Language), including: the data storage nodes store computing power data on the data nodes of the MPP distributed data architecture according to the MPP data writing strategy, and the MPP data storage nodes convert the data into the corresponding format for storage and generate relevant indexes;

[0028] The central node manages all computing power data nodes in the network and sends real-time changes in computing power data to relevant MEC data nodes via a message mechanism.

[0029] Furthermore,

[0030] The data processing nodes include read-write nodes, read-only nodes, and monitoring nodes;

[0031] The read-write nodes are used to complete various data processing tasks for reading and writing computing power data. After the write processing is completed, a message mechanism is used to notify other read-only nodes to synchronize.

[0032] The read-only nodes are used to perform operations such as reading computing power data and are dynamically scheduled by the supervisory nodes;

[0033] The monitoring node is used to monitor the operational status of local computing resource processing, and to add and unload computing resource processing instances based on the monitoring results, so as to ensure that computing data processing meets the indicators set by the system.

[0034] Furthermore, the SQL execution work includes:

[0035] When performing SQL calculations, the SQL statement is first converted into logical execution units according to its semantics, and then broken down into specific physical execution units according to the distribution of the data. Finally, it is executed on the corresponding MEC data storage node to realize the mapping from SQL to specific data storage MEC.

[0036] Furthermore, by utilizing existing electronic map technology to plan layers and computing power tiles according to the demand for computing power resources, a computing power resource map is formed to provide services for computing power resource search, including:

[0037] Based on the content of map digitization, the needs of geographic information processing and storage, and the different query requirements of computing power resources, a corresponding layer scheme for computing power maps is formulated.

[0038] Digitally encode computing resources;

[0039] Determine the number N of zoom levels provided by the computing power map, take the map image with the highest zoom level and the largest map scale as the bottom layer of the pyramid, define the operation rules of computing power resource data, form the 0th layer tile matrix, generate the next layer of computing power tiles according to the tile segmentation rules, and define the data operation rules of the 1st layer of computing power tiles, forming the 1st layer tile matrix; generate the tile matrices of each layer in sequence until the entire tile pyramid is formed.

[0040] The computing power tiles are stored in a distributed mode. The computing power tiles of each MEC access area are stored on the local nodes of the MPP data nodes. The central node stores the computing power map data and the computing power resource data corresponding to the tile data. According to the dynamic changes of computing power resource data, the relevant tile storage nodes are notified through a message mechanism to keep the computing power resource data of the entire system synchronized. The computing power tile data of each MEC is provided through the tile server in the MEC.

[0041] Furthermore, the layer scheme of the computing power map includes:

[0042] General polygon layer: includes the structure of the base map, obtained from the general map;

[0043] Computing resource location layer: generated from the location data of computing resources, used to represent the location of computing resources, and supports various location-based queries and operations;

[0044] MEC Access Boundary Layer: Includes the range and boundaries of computing resources at different MEC access points, supporting various MEC-based queries and operations.

[0045] Computing power resource type graph: Includes different computing power types, supports various queries and operations based on computing power type, including general computing power type, special computing power type, scientific computing type, engineering computing type, and intelligent computing type;

[0046] Access line layer: includes different access bandwidth types and access bandwidth indicators, and supports various queries and calculations based on access line type;

[0047] Tariff layer: Includes different tariff types and tariff ranges, supporting various tariff-based queries and calculations;

[0048] Resource Source Layer: Includes different sources of computing power resources, supporting various queries and operations based on resource source;

[0049] Composite graphics: New layers created by combining the above layers, supporting queries and operations based on composite features.

[0050] Secondly, this disclosure provides an integrated access management system for computing resources, the system comprising:

[0051] The access module is designed to enable access to different computing resources: it connects different computing resources through the central node and various levels of MEC nodes in the MEC architecture, and manages the lifecycle of computing resources through MEC.

[0052] The integration module is designed to integrate computing resource data: it divides the integration of computing resource data into a data domain and a management domain; it implements data integration in the data domain based on the MPP distributed data architecture, and provides data storage methods and data organization forms for the integrated computing resource data of the entire network; in the management domain, it completes the storage and operation processing of computing resources, as well as the management of computing data nodes through computing resource management.

[0053] The storage module is designed to store computing power data: it adopts an MPP distributed data architecture, which, while preserving the independence of MEC data nodes at each level, enables distributed storage, overall management, and data retrieval.

[0054] The computing power map module is designed to realize the computing power map by using existing electronic map technology to plan layers and computing power tiles according to the demand for computing power resources, forming a computing power resource map and providing services for computing power resource search.

[0055] The search module is designed to enable the search of computing resources: by utilizing the computing power map and based on the search API interface provided by the map, it provides various search services for computing resources.

[0056] The routing module is configured to implement computing resource routing: providing optimized network connection routes for computing resource users and selected computing resources, and generating optimized computing resource routes based on the path planning API provided by the map and computing resource map data.

[0057] Thirdly, this disclosure provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and when the processor runs the computer program stored in the memory, the processor executes the integrated access management method for computing resources as described in any of the first aspects.

[0058] Fourthly, this disclosure provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the integrated access management method for computing resources as described in any of the first aspects above.

[0059] Beneficial effects:

[0060] The integrated access management method, integrated access management system, electronic equipment and storage media provided in this disclosure, based on the equipment and network resources of existing telecommunications networks, the data integration and search capabilities based on electronic maps, and the data display capabilities based on geographical locations, construct an integrated access management method for computing resources based on computing power maps. This method realizes the access and management of various computing power resources throughout the network, as well as data integration and search functions, and provides a feasible path for the integration of computing power resources. It has positive significance for promoting the development and popularization of computing power networks. Attached Figure Description

[0061] Figure 1 A flowchart illustrating an integrated access management method for computing resources provided in Embodiment 1 of this disclosure;

[0062] Figure 2 A schematic diagram of two types of nodes in a management domain for computing power data integration provided in this embodiment of the disclosure;

[0063] Figure 3 This is a schematic diagram of an integrated access management architecture for computing resources provided in Embodiment 2 of this disclosure;

[0064] Figure 4 This is a schematic diagram illustrating data interconnection across data domains based on an MPP distributed data architecture, as provided in Embodiment 2 of this disclosure.

[0065] Figure 5 This is an architecture diagram of an integrated access management system for computing resources provided in Embodiment 3 of this disclosure;

[0066] Figure 6 This is an architectural diagram of an electronic device provided in Embodiment 4 of this disclosure. Detailed Implementation

[0067] To enable those skilled in the art to better understand the technical solutions of this disclosure, the disclosure will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments and drawings described herein are merely for explaining the invention and are not intended to limit the invention.

[0068] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this disclosure are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence; furthermore, in the absence of conflict, the embodiments and features in the embodiments of this disclosure can be arbitrarily combined with each other.

[0069] The terminology used in the embodiments of this disclosure is for the purpose of describing particular embodiments only and is not intended to be limiting of this disclosure. The singular forms “a,” “the,” and “the” as used in the embodiments of this disclosure and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise.

[0070] In the following description, the use of suffixes such as “module,” “part,” or “unit” to denote elements is solely for the purpose of illustrative purposes and has no specific meaning in itself. Therefore, “module,” “part,” or “unit” may be used interchangeably.

[0071] Computing resources are diverse, including CPUs (Central Processing Units), GPUs (Graphics Processing Units), cloud technologies, mobile edge computing, and enterprise edge computing, as well as heterogeneous computing, serverless computing, in-memory computing, and offloading of computing power. Only by managing and scheduling these computing resources in an integrated manner—that is, by achieving resource integration—can the advantages of the diversity and distribution of computing resources be fully utilized, providing users with a wealth of computing resource services. Currently, achieving resource integration faces the following challenges:

[0072] 1) Computing resources are widely distributed, diverse in type, and have very different characteristics. Secure access is a prerequisite for successful connection.

[0073] 2) Existing communication networks cannot directly connect computing resources. It is necessary to establish an independent network for managing and connecting computing resources, which involves large investments and long construction periods.

[0074] 3) The computing resources are located in a wide area and are numerous. Searching for the required computing resources from massive amounts of information requires the construction of a highly complex system.

[0075] 4) Traditional data storage architectures face significant challenges in addressing the unified storage and data processing of widely distributed, massive computing resources.

[0076] 5G's MEC (Multi-Access Edge Computing) combines CT (Communication Technology) and IT (Information Technology), making it an excellent platform for deploying computing power applications.

[0077] This disclosure fully utilizes the existing MEC resources of 5G networks and the powerful data integration and search capabilities of electronic maps to design a method for integrating computing resources based on computing power maps.

[0078] The following detailed embodiments illustrate the technical solutions of this disclosure and how they solve the technical problems existing in the prior art. It is understood that in the embodiments of this application, the executing entity may perform some or all of the steps in the embodiments of this application. These steps or operations are merely examples, and the embodiments of this application may also perform other operations or variations thereof. Furthermore, the steps may be executed in different orders as presented in the embodiments of this application, and it is not necessary to execute all the operations in the embodiments of this application. Moreover, the following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0079] Figure 1 This is a flowchart illustrating an integrated access management method for computing resources provided in Embodiment 1 of this disclosure, as shown below. Figure 1 As shown, the method includes:

[0080] Step S101: Access to different computing resources: Access different computing resources through the central node and MEC nodes at all levels of the MEC architecture, and complete the lifecycle management of computing resources through MEC;

[0081] Step S102: Data Integration of Computing Resources: The integration of computing resource data is divided into a data domain and a management domain; the data domain is integrated based on the MPP distributed data architecture, providing data storage methods and data organization forms for the integrated computing resource data of the entire network; in the management domain, the storage and operation processing of computing resources, as well as the management of computing data nodes, are completed through computing resource management.

[0082] Step S103: Computing power data storage: Adopting an MPP distributed data architecture, while preserving the independence of MEC data nodes at each level, it realizes distributed storage, overall management, and data retrieval of data;

[0083] Step S104: Implementation of computing power map: Using existing electronic map technology, layers and computing power tiles are planned according to the demand for computing power resources to form a computing power resource map, providing services for computing power resource search;

[0084] Step S105: Computing resource search: Utilize the computing resource map and provide various search services for computing resources based on the search API interface provided by the map;

[0085] Step S106: Computing resource routing: Provide optimized network connection routes for computing resource users and selected computing resources, and generate optimized computing resource routes based on the path planning API provided by the map and computing map data.

[0086] The goal of computing resource integration is to achieve lifecycle management of all computing resources across the network, including registration, online status, authentication, publication, status tracking, offline status, and metering, providing users with the necessary computing resource search services. Utilizing existing network infrastructure and technical resources to achieve computing resource integration is highly feasible. Therefore, this embodiment achieves the goal of computing resource integration based on MEC and a computing map constructed from computing resource data and a distributed map.

[0087] Based on the current MEC edge cloud platform architecture, the MEC edge cloud is mainly divided into three layers: the global central node, regional central / provincial capital nodes, and local core / edge nodes. The MEC framework enables the integration of computing resources. After implementing a computing power map, the path planning API based on the computing power map, and the network connection layer data of the computing power map, determines the optimal route from the user access node to the selected computing power resource. Furthermore, it utilizes the real-time traffic API provided by the map to achieve dynamic route selection based on dynamic link data, and leverages the AI ​​(Artificial Intelligence) path planning API provided by the map to achieve further optimized route selection based on artificial intelligence and machine learning.

[0088] This disclosure embodiment achieves the integration of computing resources based on a computing power map. It utilizes existing MEC resources in communication networks and distributed storage technology, and realizes the integration of computing resource data based on computing resource requirements. Furthermore, the computing resource data is used through the API interface provided by map technology to enable the search and routing of computing resources.

[0089] Furthermore,

[0090] The computing resources include the computing resources of the cloud center, the computing resources in the MEC, and the customer computing resources accessed through the MEC.

[0091] The process of connecting different computing resources through the central node and various levels of MEC nodes in the MEC architecture, and managing the lifecycle of computing resources through MEC, includes:

[0092] The computing resources of the cloud center are accessed through the unified cloud management platform of the central node, and an authentication system based on the MEC management system is adopted. The lifecycle management of computing resources is realized through cloud computing network management.

[0093] For the computing resources in the MEC and the customer computing resources accessed through the MEC, based on the four deployment modes of the MEC's ​​distribution network elements and the MEC resource platform, the lifecycle management of the computing resources is realized by the MEC's ​​API gateway after the computing power gateway converts the protocol.

[0094] The access to computing resources primarily occurs at two types of nodes. The first type is computing resources in large cloud centers, accessed through the unified cloud management platform of the central node. This access utilizes a certificate authentication system based on the MEC management framework. The lifecycle management of computing resources, including registration, deployment, authentication, publication, status, decommissioning, and metering, is implemented through the cloud computing power network management system. Computing data is sent to the computing power data center, processed, mapped onto the computing power map, and saved to the computing power map data center. The second type includes computing resources within the MEC and customer computing resources accessed through the MEC. Based on four deployment modes, different types of computing resource access services are provided to users.

[0095] In Mode 1, the MEC's ​​offloading network elements (UPF - User Plane Function) / GW-U - Gateway for User and the MEC resource platform are all deployed in aggregation or municipal core layer data centers. Therefore, different quality computing resources are configured through these offloading network elements. Higher-level MEC deployments offer higher performance and reliability of computing resources, but also higher latency and cost. Lower-level MEC deployments have computing resources closer to customers, resulting in lower latency and relatively lower cost.

[0096] In Mode 2, the MEC's ​​offloading network elements UPF / GW-U and platform are all deployed in the customer-side access room or on-site customer room, and the offloading network elements and platform are dedicated to a specific customer. The computing resources deployed in Mode 2 are mainly targeted at specific customers or specific scenarios.

[0097] In Mode 3, the MEC's ​​offloading network elements UPF / GW-U are deployed in the core / aggregation data centers, while the platform is deployed in the customer-side access / field-level edge data centers. Mode 3 supports various types of low-cost computing resources.

[0098] Mode 4) provides access to computing resources for areas with only 4G or 5G NSA (Non-Standalone) network configurations.

[0099] The lifecycle of computing resources in Modes 1-4, including registration, online status, authentication, publication, status tracking, offline status, and metering, is implemented through the MEC's ​​API Gateway after protocol conversion by the computing power gateway. For computing resources accessed via leased lines, authentication is performed using user access tokens. For access via terminals, security protection for computing resource access points is achieved through authentication, encryption, and verification by the core network.

[0100] Furthermore,

[0101] The data interconnection of data domains based on the MPP distributed data architecture includes:

[0102] In terms of data processing, the computing power data processing tasks are distributed in parallel to multiple related nodes through MPP. After the data processing is completed on each node, the results of each part are aggregated to obtain the final result.

[0103] In terms of data structure, the MPP node data of local computing resources in MEC is organized in the form of a data mart.

[0104] The data domain performs various operations on the integrated network-wide computing resource data. Data integration within the data domain is implemented based on the MPP distributed data architecture. In data processing, MPP distributes computing data processing tasks in parallel across multiple relevant nodes. After data processing is completed on each node, the results from each node are aggregated to obtain the final result.

[0105] Computing resources are distributed across large cloud centers, data centers (DCs) and virtual systems at various levels of MECs, and user access points and nodes via four MEC access modes. The first goal of computing resource integration is to unify the management and scheduling of different computing resources, providing users with optimized scheduling and allocation of computing resources across all access nodes and regions, fully utilizing the network's computing resources to provide better service. The second goal is to implement hierarchical distributed management of the vast number of computing resources, dynamically aggregating lifecycle management data such as registration, online status, authentication, publication, status, offline status, and metering of computing resources at access nodes, forming unified data for the entire network's computing resources, available to all computing resource management nodes. To achieve these goals, computing resource integration is divided into two domains: a data domain and a management domain.

[0106] The data domain performs various operations on the integrated network-wide computing resource data, such as searching, matching, and analyzing data operations required for computing resource matching and optimization. Dynamic data on the registration, online status, authentication, publication, status, offline status, and metering lifecycle management of computing resources at each level of the MEC are sent by the MEC computing gateway to the distributed database for data domain integration and aggregation.

[0107] In this embodiment, data integration across the data domain is implemented using an MPP (Massively Parallel Processing) distributed data architecture.

[0108] MPP consists of a large number of SMP (Symmetric Multi-Processing) service nodes working together through a node interconnection network. Each node only accesses its own local resources. As a database with a completely share-nothing structure, it has very good scalability and is suitable for storing and processing the computing power of dynamically changing massive MEC nodes.

[0109] In data processing (such as data querying and analysis required for computing resource optimization), MPP distributes computing data processing tasks in parallel across multiple related nodes. After data processing is completed on each node, the results of each part are aggregated to obtain the final result (for example, the Greenplum distributed database uses an MPP architecture, where the host, operating system, memory, and storage are self-controlled and not shared. Each node is a separate database). The Hadoop architecture has similar functionality, but Hadoop has advantages in processing unstructured and semi-structured data. Since computing resource data is structured data, Hadoop's performance is relatively slow and cannot meet the requirements of real-time processing of computing resources. Therefore, using MPP to process computing resource data with relational data structures is more efficient.

[0110] In terms of data structure, since each MEC primarily uses local MEC computing power data, neighboring MEC computing power data, and cloud center computing power data, the geographical location and other characteristics of computing power resources that each MEC focuses on vary significantly. Furthermore, the computing power data needs to be further mapped to different layers of the computing power data map to facilitate queries from different dimensions (e.g., computing power, cost, latency) to meet the diverse needs of computing power resource users. Therefore, a data mart structure is adopted to organize the MPP node data of relevant local computing power resources within the MEC.

[0111] Data marts (DMs) can meet specific needs by storing data in a multi-dimensional way, including defining dimensions, metrics to be calculated, and the hierarchy of dimensions, generating data cubes that meet the needs of different layers of a computing power data map.

[0112] The computing power data is organized in a star schema within the data mart, containing fact tables and various supporting dimension tables. The fact tables describe the most relevant computing power data within the local MEC of the data mart, such as computing power resource data within the local MEC, third-party computing power resource data accessed through the local MEC, and computing power resource data from surrounding MECs. The fact tables are highly indexed (e.g., each column is indexed), making data retrieval and reading from the fact tables extremely fast.

[0113] Dimension tables are built on top of fact tables and contain non-intensive computing power data. They are linked to the fact table via foreign keys. Users store additional information about the data in different columns of the fact table, such as the position of nodes with similar computing power resource node characteristics in the dimension table base.

[0114] The detailed data of each data node is contained in the data mart star schema, which serves as a description of the data resources and the source of raw data for subsequent operations.

[0115] Furthermore,

[0116] The management domain includes the computing power resource library management node and the computing power data center node of the central node in the MEC. The computing power resource library management node includes data processing nodes and data storage nodes.

[0117] The management domain manages the storage and operation of computing resources, as well as the management of computing data nodes, through computing resource management.

[0118] The management of computing power resource libraries in different levels of MEC uses data processing middleware to process the real-time data of computing power resources provided by the MEC computing power gateway, forming MEC computing power mart data, and updating it to the local data nodes of the distributed data architecture MPP.

[0119] The data processing node completes the metadata processing and connects to the query terminal in the MEC computing power gateway to perform computing power resource queries and other operations.

[0120] The data storage nodes complete the storage of computing power data and SQL execution, including: the data storage nodes store computing power data on the data nodes of the MPP distributed data architecture according to the MPP data writing strategy, and the MPP data storage nodes convert the data into the corresponding format for storage and generate relevant indexes;

[0121] The central node manages all computing power data nodes in the network and sends real-time changes in computing power data to relevant MEC data nodes via a message mechanism.

[0122] The management domain for integrated computing power data is divided into two types of nodes: the computing power resource repository management in MEC and the computing power data center of the central node, such as... Figure 2 As shown.

[0123] The management of computing power resource libraries in different levels of MEC uses data processing middleware to process the real-time data of computing power resources provided by the MEC computing power gateway, forming MEC computing power mart data, and updating it to the local data nodes of the distributed data architecture MPP.

[0124] The computing resource repository management consists of two types of nodes: data processing and data storage. The data processing nodes process metadata and connect to the query terminal in the MEC computing power gateway to perform computing resource queries and other operations.

[0125] The data storage nodes complete the storage of computing power data and SQL execution. In terms of data storage, the data storage nodes store data on the data nodes of the MPP distributed data architecture according to the MPP data writing strategy. The MPP data storage nodes then convert the data into the corresponding format for storage and generate relevant indexes.

[0126] The central node's computing power data center manages all computing power data nodes in the network, including registration, startup, maintenance, and deregistration, as well as the aggregation of computing power data and the sending of real-time changes in computing power data to relevant MEC data nodes that have subscribed to the message via a messaging mechanism, thus achieving data consistency across the entire network.

[0127] Furthermore,

[0128] The data processing nodes include read-write nodes, read-only nodes, and monitoring nodes;

[0129] The read-write nodes are used to complete various data processing tasks for reading and writing computing power data. After the write processing is completed, a message mechanism is used to notify other read-only nodes to synchronize.

[0130] The read-only nodes are used to perform operations such as reading computing power data and are dynamically scheduled by the supervisory nodes;

[0131] The monitoring node is used to monitor the operational status of local computing resource processing, and to add and unload computing resource processing instances based on the monitoring results, so as to ensure that computing data processing meets the indicators set by the system.

[0132] Since the vast majority of computing power data involves read operations, data processing nodes are divided into read-write nodes, read-only nodes, and supervisory nodes. Only one read-write node is deployed per node, capable of handling both reading and writing of computing power data. After write processing is complete, a message mechanism is used to notify other read-only nodes for synchronization. Read-only nodes can only perform read operations on computing power data and do not require a synchronization mechanism. Multiple read-only nodes can be deployed, dynamically scheduled by supervisory nodes. Supervisory nodes monitor the operational status of local computing power resource processing and add or remove computing power resource processing instances based on the monitoring results to ensure that computing power data processing meets the system's set metrics.

[0133] Furthermore, the SQL execution work includes:

[0134] When performing SQL calculations, the SQL statement is first converted into logical execution units according to its semantics, and then broken down into specific physical execution units according to the distribution of the data. Finally, it is executed on the corresponding MEC data storage node to realize the mapping from SQL to specific data storage MEC.

[0135] Data storage nodes handle both storing computing power data and executing SQL queries. For data storage, data storage nodes store data on MPP distributed data architecture nodes according to the MPP data writing strategy. The MPP data storage nodes then convert the data into the corresponding format and generate relevant indexes. During SQL computation, the SQL statement is first semantically converted into logical execution units, and then broken down into specific physical execution units based on the data distribution. These units are then executed on the corresponding MEC data storage nodes, achieving a mapping from SQL to specific MEC data storage, avoiding data transmission and copying, and thus achieving good query performance.

[0136] Furthermore, by utilizing existing electronic map technology to plan layers and computing power tiles according to the demand for computing power resources, a computing power resource map is formed to provide services for computing power resource search, including:

[0137] Based on the content of map digitization, the needs of geographic information processing and storage, and the different query requirements of computing power resources, a corresponding layer scheme for computing power maps is formulated.

[0138] Digitally encode computing resources;

[0139] Determine the number N of zoom levels provided by the computing power map, take the map image with the highest zoom level and the largest map scale as the bottom layer of the pyramid, define the operation rules of computing power resource data, form the 0th layer tile matrix, generate the next layer of computing power tiles according to the tile segmentation rules, and define the data operation rules of the 1st layer of computing power tiles, forming the 1st layer tile matrix; generate the tile matrices of each layer in sequence until the entire tile pyramid is formed.

[0140] The computing power tiles are stored in a distributed mode. The computing power tiles of each MEC access area are stored on the local nodes of the MPP data nodes. The central node stores the computing power map data and the computing power resource data corresponding to the tile data. According to the dynamic changes of computing power resource data, the relevant tile storage nodes are notified through a message mechanism to keep the computing power resource data of the entire system synchronized. The computing power tile data of each MEC is provided through the tile server in the MEC.

[0141] To meet the needs of displaying, querying, and locating computing resources during use, the computing resources are digitally encoded, consisting of the following parts:

[0142] {MEC code} + {MEC internal identifier}

[0143] The tile map pyramid model is a multi-resolution hierarchical model. From the bottom to the top of the tile pyramid, the resolution decreases, but the geographical area represented remains the same. Users can query computing resources at different zoom levels within the same area. The implementation of computing power tiles first determines the number N of zoom levels provided by the computing power map. Then, the map image with the highest zoom level and largest map scale is used as the bottom layer of the pyramid, i.e., layer 0. The computational rules for computing power resource data are defined, forming the layer 0 tile matrix. Based on the layer 0 map image, layer 1 computing power tiles are generated according to the tile segmentation rules, and the data computation rules for layer 1 computing power tiles are defined, forming the layer 1 tile matrix. The same method is used to generate the layer 2 tile matrix, and so on, until the (N-1)th layer, constituting the entire tile pyramid.

[0144] Furthermore, the layer scheme of the computing power map includes:

[0145] General polygon layer: includes the structure of the base map, obtained from the general map;

[0146] Computing resource location layer: generated from the location data of computing resources, used to represent the location of computing resources, and supports various location-based queries and operations;

[0147] MEC Access Boundary Layer: Includes the range and boundaries of computing resources at different MEC access points, supporting various MEC-based queries and operations.

[0148] Computing power resource type graph: Includes different computing power types, supports various queries and operations based on computing power type, including general computing power type, special computing power type, scientific computing type, engineering computing type, and intelligent computing type;

[0149] Access line layer: includes different access bandwidth types and access bandwidth indicators, and supports various queries and calculations based on access line type;

[0150] Tariff layer: Includes different tariff types and tariff ranges, supporting various tariff-based queries and calculations;

[0151] Resource Source Layer: Includes different sources of computing power resources, supporting various queries and operations based on resource source;

[0152] Composite graphics: New layers created by combining the above layers, supporting queries and operations based on composite features.

[0153] In the aforementioned layer scheme, the structure of the base map, including elements representing regional geographical features such as major city streets and highways, buildings, reservoirs and lakes, oceanic isobaths and rivers, parks and green spaces, is obtained from a general map. Different types of computing power include, for example, general-purpose computing power (e.g., CPU), dedicated computing power (e.g., GPU, general AI, image processing), scientific computing, engineering computing, and intelligent computing. Different access bandwidth types and metrics include, for example, SDN (Software Defined Network) based DCN (Data Communication Network) and DCI (Data Center Interconnection) networks, SD-WAN Overlay networks for user-cloud connections, Network for AI (Network for AI) networks, IB, RDMA, and larger-scale non-blocking DCNs for data center internal networks, CFN (Compute First Network) for user data-to-computing power connections, and Metro Fabric and 5G URLLC networks supporting lower latency and higher bandwidth for user data-to-computing power connections. The sources of computing power resources include, for example, large-scale cloud computing, telecom cloud computing, enterprise cloud computing, local cloud computing, third-party computing power resources, personal computing power resources, and terminal computing power resources. The aforementioned layering scheme can satisfy queries and calculations based on various features. Computing power searches are implemented through the built-in API interface of the computing power map, such as queries based on computing power resource codes and computing power map queries based on computing power resource features (city areas), allowing queries for information on various computing power resources, such as nearby personally shared computing power resources.

[0154] This disclosure fully utilizes the existing MEC resources of 5G networks and the powerful data integration and search capabilities of electronic maps to design an integrated access management method for computing resources, which has the following advantages:

[0155] 1) By leveraging the integration of MEC and 5G networks, and based on the service interfaces of 5G networks and MEC, access and management of computing resources can be achieved;

[0156] 2) Building a computing power resource integration platform by utilizing existing MEC hardware resources and open network resources has the advantages of short construction cycle and low investment;

[0157] 3) By leveraging the information integration and search capabilities of distributed maps, the connection and search of computing resource information were realized;

[0158] 4) By utilizing a distributed data storage architecture, while ensuring the independence of MEC data nodes at all levels, the integration and unified management of computing resources are realized.

[0159] To more clearly describe the technical solution of this disclosure, Embodiment 2 of this disclosure also provides an integrated access management method for computing resources; comprising three parts: 1) overall architecture; 2) method for integrating computing resources in MEC; 3) method for unified management of computing resources based on computing power map.

[0160] Part 1: Overall Architecture

[0161] The goal of computing resource integration is to achieve lifecycle management of all computing resources across the network, including registration, online status, authentication, publication, status tracking, offline status, and metering, providing users with the necessary computing resource search services. Utilizing existing network infrastructure and technical resources to achieve computing resource integration is highly feasible. Therefore, this embodiment achieves the goal of computing resource integration based on MEC (Multi-access Edge Computing) and a computing resource map constructed from computing resource data and a distributed map.

[0162] MEC edge cloud is mainly divided into three levels: network-wide central nodes, regional central / provincial capital nodes, and local core / edge nodes. Based on the MEC framework, the overall architecture of integrated access management of computing resources in this embodiment is as follows: Figure 3 As shown: including:

[0163] S1: Access to different computing resources

[0164] Different computing resources are accessed through the central node of the MEC architecture and MECs at all levels. The MEC is used to complete the lifecycle management of computing resources, including registration, online, authentication, release, status, offline and metering.

[0165] S2: Data Integration of Computing Resources

[0166] The integrated computing resources are divided into two domains: the data domain and the management domain. The data domain provides data storage methods and data organization formats for the integrated network's computing resources. The management domain manages the computing resources, handling storage, operations, and the management of computing data nodes.

[0167] S3: Computing power data storage

[0168] The computing power data storage adopts an MPP distributed data architecture, which preserves the independence of each level of MEC data nodes while realizing distributed data storage and overall management / data retrieval.

[0169] S4: Implementation of Computing Power Map

[0170] The computing power map utilizes existing electronic map technology to plan layers and computing power tiles according to the demand for computing power resources, forming a computing power resource map and providing services for computing power resource search.

[0171] S5: Computing Resource Search

[0172] The computing power resource search utilizes the computing power map and provides various search services for computing power resources based on the search API interface provided by the map (the search type varies depending on the map search API).

[0173] S6: Computing Resource Routing

[0174] The computing resource access routing provides optimized network connection routes for computing resource users and selected computing resources, generating optimized computing resource routes based on the path planning API provided by the map and computing resource map data.

[0175] Part Two: Methods for Implementing Integrated Computing Resources in MEC

[0176] 1. Access and registration of computing resources

[0177] According to the 5G MEC edge cloud platform architecture, large cloud centers, including public clouds, private clouds and telecom clouds, are generally managed and accessed through a unified cloud management platform at the group-level central node.

[0178] The regional center deploys systems or network elements such as MEPM (Mobile Edge Platform Manager) and MEP (Mobile Edge Platform). MEPM is responsible for the management of ME_ICT-IaaS virtualization resources, MEP access collaboration platform, and node service management for all edge nodes within the region / province of this node.

[0179] Local core nodes / edge nodes correspond to the core / aggregation / field access data center nodes in various cities where MEC is deployed. ME_ICT-IaaS, MEP, and ME-VAS are deployed within these nodes to support specific customer business applications. Core / aggregation nodes typically support distributed / shared services, while field access nodes typically support dedicated services for a specific customer.

[0180] Therefore, computing resources are mainly accessed through two types of nodes, such as Figure 2 As shown. The first category is computing power resources in large cloud centers, which are accessed through a unified cloud management platform at the central node. These computing power resources are typically connected via dedicated lines, therefore an authentication system based on a certificate management framework (MEC) is used for access. The lifecycle management of computing power resources, including registration, deployment, authentication, publication, status, decommissioning, and metering, is implemented through the cloud computing power network management system. Computing power data is sent to the computing power data center, processed there, mapped onto a computing power map, and then saved to the computing power map data center.

[0181] The second category comprises computing resources within the MEC and customer computing resources accessed through the MEC. MEC node deployment adopts a converged ICT deployment model of "network + platform," with the network side providing traffic offloading and configuration capabilities, and the platform using the self-developed CUC-MEC platform. Customers' own applications can also be deployed based on the CUC-MEC platform. MEC service deployment primarily employs four modes: Mode 1) Offloading and sharing + platform sharing deployment mode; Mode 2) Offloading and dedicated + platform dedicated deployment mode; Mode 3) Offloading and sharing + platform dedicated deployment mode; and Mode 4) Platform deployment (DP) mode. Therefore, this second category of computing resources provides users with different types of computing resource access services based on these four deployment modes.

[0182] In Mode 1, the offloading network elements UPF / GW-U and the MEC resource platform are all deployed in aggregation or municipal core layer data centers. Therefore, different quality computing resources are configured through these offloading network elements. Higher-level MECs offer higher performance and reliability due to better overall data center conditions, abundant hardware resources and access bandwidth, and high reliability, but also higher latency and cost. Lower-level MECs, being closer to customers, offer lower latency and relatively lower cost. Mode 1 provides a wide variety of computing resource quality and pricing options, good reliability, and a broad service range, making it suitable for operators providing various public computing resource services.

[0183] In Mode 2, the offloading network elements UPF / GW-U and platform of MEC are all deployed in the customer-side access room or on-site customer room, and the offloading network elements and platform are dedicated to a specific customer. The computing resources deployed in Mode 2 are mainly for specific customers or specific scenarios, such as roadside units (RSUs) used for vehicle short-range communication, or computing resources provided by enterprise clouds for customers of large enterprises or industrial parks.

[0184] In Mode 3, the offloading network elements UPF / GW-U of the MEC are deployed in the core / aggregation data center, while the platform is deployed in the customer-side access / on-site edge data center. Mode 3 allows for the sharing of offloading network elements, thus supporting various types of low-cost computing resources, such as enterprise cloud computing resources, third-party computing resources, and customer-shared computing resources.

[0185] Mode 4) targets MEC services in 4G / NSA areas. Therefore, it provides access to computing resources in areas with only 4G or 5G NSA network deployments.

[0186] The lifecycle of computing resources in Modes 1-4, including registration, online status, authentication, publication, status tracking, offline status, and metering, is implemented through the MEC's ​​API Gateway after protocol conversion by the computing power gateway. For computing resources accessed via leased lines, authentication is performed using user access tokens. For those accessed via terminal access points, security protection for the computing resource access points is achieved through authentication, encryption, and verification by the core network.

[0187] 2. Methods for integrating computing resources

[0188] Computing resources are distributed across large cloud centers, data centers (DCs) at various levels of MECs, virtual systems, and user access points and nodes via four MEC access modes. The first goal of computing resource integration is to unify the management and scheduling of different computing resources, providing users with optimized scheduling and allocation of computing resources across all access nodes and regions, fully utilizing the network's computing resources to provide better service. The second goal is to implement hierarchical distributed management of the vast number of computing resources, dynamically aggregating lifecycle management data such as registration, online status, authentication, publication, status, offline status, and metering of computing resources at access nodes, forming unified data for the entire network's computing resources, available to all computing resource management nodes. To achieve these goals, computing resource integration is divided into two domains: a data domain and a management domain.

[0189] 1) Data Domain

[0190] The data domain performs various operations on the integrated network-wide computing resource data, such as searching, matching, and analyzing data operations required for computing resource matching and optimization. Dynamic data on the registration, online status, authentication, publication, status, offline status, and metering lifecycle management of computing resources at each level of the MEC are sent by the MEC computing gateway to the distributed database for data domain integration and aggregation.

[0191] In this embodiment, data connectivity across data domains is implemented using an MPP distributed data architecture, such as... Figure 4 As shown.

[0192] MPP (Massively Parallel Processing) consists of a large number of SMP (Symmetric Multi-Processing) service nodes working together through a node interconnection network. Each node only accesses its own local resources. As a database with a completely share-nothing structure, it has very good scalability and is suitable for storing and processing the computing power of dynamically changing massive MEC nodes.

[0193] In data processing (such as data querying and analysis required for computing resource optimization), MPP distributes computing data processing tasks in parallel across multiple related nodes. After data processing is completed on each node, the results of each part are aggregated to obtain the final result (for example, the Greenplum distributed database uses an MPP architecture, where the host, operating system, memory, and storage are self-controlled and not shared. Each node is a separate database). The Hadoop architecture has similar functionality, but Hadoop has advantages in processing unstructured and semi-structured data. Since computing resource data is structured data, Hadoop's performance is relatively slow and cannot meet the requirements of real-time processing of computing resources. Therefore, using MPP to process computing resource data with relational data structures is more efficient.

[0194] In terms of data structure, since each MEC primarily uses local MEC computing power data, neighboring MEC computing power data, and cloud center computing power data, the geographical location and other characteristics of computing power resources that each MEC focuses on vary significantly. Furthermore, the computing power data needs to be further mapped to different layers of the computing power data map to facilitate queries from different dimensions (e.g., computing power, cost, latency) to meet the diverse needs of computing power resource users. Therefore, a data mart structure is adopted to organize the MPP node data of relevant local computing power resources within the MEC.

[0195] Data marts (DMs) can meet specific needs by storing data in a multi-dimensional way, including defining dimensions, metrics to be calculated, and the hierarchy of dimensions, generating data cubes that meet the needs of different layers of a computing power data map.

[0196] The computing power data is organized in a star schema within the data mart, containing fact tables and various supporting dimension tables. The fact tables describe the most relevant computing power data within the local MEC of the data mart, such as computing power resource data within the local MEC, third-party computing power resource data accessed through the local MEC, and computing power resource data from surrounding MECs. The fact tables are highly indexed (e.g., each column is indexed), making data retrieval and reading from the fact tables extremely fast.

[0197] Dimension tables are built on top of fact tables and contain non-intensive computing power data. They are linked to the fact table via foreign keys. Users store additional information about the data in different columns of the fact table, such as the position of nodes with similar computing power resource node characteristics in the dimension table base.

[0198] The detailed data of each data node is contained in the data mart star schema, which serves as a description of the data resources and the source of raw data for subsequent operations.

[0199] 2) Management Domain

[0200] The management domain for integrated computing power data is divided into two types of nodes: the computing power resource repository management in MEC and the computing power data center of the central node, such as... Figure 2 As shown.

[0201] The management of computing power resource libraries in different levels of MEC uses data processing middleware to process the real-time data of computing power resources provided by the MEC computing power gateway, forming MEC computing power mart data, and updating it to the local data nodes of the distributed data architecture MPP.

[0202] The computing resource repository management consists of two types of nodes: data processing and data storage. Data processing nodes process metadata and connect to the query endpoint in the MEC computing power gateway to perform queries and other calculations on computing resources. Because the vast majority of computing data is read-based, data processing nodes are divided into read-write nodes, read-only nodes, and monitoring nodes. Only one read-write node is deployed per node, capable of performing various data processing operations, including reading and writing. After writing, it notifies other read-only nodes of synchronization via a message mechanism. Read-only nodes can only perform read operations on computing data, such as queries and statistics, and do not require a synchronization mechanism. Multiple read-only nodes can be deployed and dynamically scheduled by monitoring nodes. Monitoring nodes monitor the operational status of local computing resource processing and add or remove computing resource processing instances based on the monitoring results to ensure that computing data processing meets the system's set metrics.

[0203] Data storage nodes handle both storing computing power data and executing SQL queries. For data storage, data storage nodes store data on MPP distributed data architecture nodes according to the MPP data writing strategy. The MPP data storage nodes then convert the data into the corresponding format and generate relevant indexes. During SQL computation, the SQL statement is first semantically converted into logical execution units, and then broken down into specific physical execution units based on the data distribution. These units are then executed on the corresponding MEC data storage nodes, achieving a mapping from SQL to specific MEC data storage, avoiding data transmission and copying, and thus achieving good query performance.

[0204] The central node's computing power data center manages all computing power data nodes in the network, including registration, startup, maintenance, and deregistration, as well as the aggregation of computing power data and the sending of real-time changes in computing power data to relevant MEC data nodes that have subscribed to the message via a messaging mechanism, thus achieving data consistency across the entire network.

[0205] Part Three: Implementation Method of Unified Management of Computing Resources Based on Computing Power Map

[0206] 1. Data layer processing method for distributed computing power resource maps

[0207] Based on the content of map digitization, the needs of geographic information processing and storage, and the different query requirements of computing resources, corresponding computing power map layer (Coverage) schemes are formulated. The following is one computing power resource layer scheme:

[0208] 1) General area layer: This includes the structure of the base map, such as major city streets and highways, buildings, reservoirs and lakes, ocean areas and rivers, parks and green spaces, and other elements that represent the geographical features of the region, which are obtained from the general map.

[0209] 2) Computing resource location layer: generated from the location data of computing resources, used to represent the location of computing resources, and supports various location-based queries and operations.

[0210] 3) MEC Access Boundary Layer: The range and boundary of computing resources at different MEC access points, supporting various MEC-based queries and operations.

[0211] 4) Computing Power Resource Type Graphics: Different computing power types, such as general-purpose computing power (e.g., CPU), specialized computing power (e.g., GPU, general AI, image processing), scientific computing, engineering computing, and intelligent computing. Supports various queries and operations based on computing power type.

[0212] 5) Access Line Layer: Different access bandwidth types and metrics, such as SDN-based DCN and DCI networks, SD-WAN Overlay networks for user-cloud connectivity, Network for AI (Network for AI) networks, Intra-Data Center (IB) networks, RDMA networks, and larger-scale non-blocking DCNs, Compute First Network (CFN) for user data-to-computing power connectivity, and MetroFabric and 5G URLLC networks supporting lower latency and higher bandwidth for user data-to-computing power connections. Supports various queries and operations based on access line type.

[0213] 6) Tariff Layer: Different tariff types and tariff ranges support various tariff-based queries and calculations.

[0214] 7) Resource Source Layer: Different sources of computing power resources, such as large cloud, telecom cloud, enterprise cloud, local cloud, third-party computing power resources, personal computing power resources, and terminal computing power resources, supporting various queries and operations based on resource source.

[0215] 8) Combined graphics: New layers can be created by combining the above layers, supporting queries and calculations based on combined features.

[0216] 2. Digital Coding Scheme for Computing Power Resource Map

[0217] To meet the needs of displaying, querying, and locating computing resources during use, the computing resources are digitally encoded, consisting of the following parts:

[0218] {MEC code} + {MEC internal identifier}

[0219] 3. Implementation method of computing power tiles

[0220] The tile map pyramid model is a multi-resolution hierarchical model. From the bottom to the top of the tile pyramid, the resolution decreases, but the geographical range it represents remains unchanged. Users can query computing resources at different zoom levels within the same area.

[0221] First, determine the number N of zoom levels provided by the computing power map. Then, use the map image with the highest zoom level and largest map scale as the bottom layer of the pyramid, i.e., layer 0. Define the calculation rules for computing power resource data, for example, layer 0 displays the total number of computing power resources within a tile, forming the layer 0 tile matrix. Based on the layer 0 map image, generate layer 1 computing power tiles according to the tile segmentation rules, and then define the data calculation rules for layer 1 computing power tiles, for example, displaying the data based on the first query condition of the computing power resource user (e.g., the largest data resource service range), forming the layer 1 tile matrix. Use the same method to generate the layer 2 tile matrix; ...; and so on, until the (N-1)th layer, forming the entire tile pyramid.

[0222] 4. Storage Implementation Method of Computational Tile

[0223] The storage of computing power tiles adopts a distributed model. The computing power tiles of each MEC access area are stored on the local node of the MPP data node. The central node stores the computing power map data and the computing power resource data corresponding to the tile data. Based on the dynamic changes of computing power resource data, it notifies the relevant tile storage nodes through a message mechanism to maintain the synchronization of computing power resource data throughout the system. The computing power tile data of each MEC is provided as computing power tile services through the tile server in the MEC.

[0224] 5. Computing resource search methods for computing power maps

[0225] The computing power map's computing power resource search method is implemented through the map's built-in API interface. For example, it can perform queries based on computing power resource codes and computing power map queries based on computing power resource characteristics (urban areas). It can query information on various computing power resources, such as nearby personal shared computing power resources.

[0226] 6. Routing search method for computing power maps

[0227] The API for path planning based on the computing power map uses network connection layer data from the computing power map to determine the optimal route from the user access node to the selected computing power resource.

[0228] Routing planning implements routing for various network links based on network connectivity layer data. For example, applications with high reliability requirements can choose Metro Fabric and 5G URLLC network routing, while cost-sensitive users can choose routes composed of low-cost links.

[0229] By leveraging the real-time traffic API provided by the map, dynamic route selection can be achieved based on dynamic link data of the route, and the AI ​​route planning API provided by the map can achieve more optimized route selection based on artificial intelligence and machine learning.

[0230] This disclosure utilizes existing MEC resources in communication networks and distributed storage technology to achieve the integration of computing resource data based on computing power resource requirements. The computing resource data is then used through API interfaces provided by map technology to enable the search and routing of computing resources. Because it is implemented based on existing network resources and electronic map technologies, it has good feasibility and is a feasible method for achieving computing resource integration, which is of positive significance for promoting the development and popularization of computing networks.

[0231] Embodiment 3 of this disclosure also provides an integrated access management system for computing resources, such as Figure 5 As shown, the system includes:

[0232] Access module 11 is configured to enable access to different computing resources: different computing resources are accessed through the central node and MEC nodes at all levels of the MEC architecture, and the lifecycle management of computing resources is completed through MEC;

[0233] The integration module 12 is configured to realize the integration of computing resource data: the integration of computing resource data is divided into data domain and management domain; the data domain is integrated based on the MPP distributed data architecture, and the storage method and data organization form of the integrated computing resource data of the entire network are provided; in the management domain, the storage and operation processing of computing resources, as well as the management of computing data nodes are completed through computing resource management.

[0234] Storage module 13 is configured to realize computing power data storage: it adopts an MPP distributed data architecture, and realizes distributed storage, overall management and data retrieval of data while retaining the independence of MEC data nodes at each level;

[0235] The computing power map module 14 is configured to implement the computing power map: using existing electronic map technology, planning layers and computing power tiles according to the demand for computing power resources, forming a computing power resource map, and providing services for computing power resource search;

[0236] Search module 15 is configured to perform computing power resource search: using a computing power map and based on the search API interface provided by the map, it provides a variety of search services for computing power resources;

[0237] The routing module 16 is configured to implement computing resource routing: provide optimized network connection routes for computing resource users and selected computing resources, and generate optimized computing resource routes based on the path planning API provided by the map and computing map data.

[0238] Furthermore,

[0239] The computing resources include the computing resources of the cloud center, the computing resources in the MEC, and the customer computing resources accessed through the MEC.

[0240] The access module 11 is specifically configured as follows:

[0241] The computing resources of the cloud center are accessed through the unified cloud management platform of the central node, and an authentication system based on the MEC management system is adopted. The lifecycle management of computing resources is realized through cloud computing network management.

[0242] For the computing resources in the MEC and the customer computing resources accessed through the MEC, based on the four deployment modes of the MEC's ​​distribution network elements and the MEC resource platform, the lifecycle management of the computing resources is realized by the MEC's ​​API gateway after the computing power gateway converts the protocol.

[0243] Furthermore, the connection module 12 includes:

[0244] The data domain integration unit is configured as follows:

[0245] In terms of data processing, the computing power data processing tasks are distributed in parallel to multiple related nodes through MPP. After the data processing is completed on each node, the results of each part are aggregated to obtain the final result.

[0246] In terms of data structure, the MPP node data of local computing resources in MEC is organized in the form of a data mart.

[0247] Furthermore, the connection module 12 also includes a management domain connection unit;

[0248] The management domain includes the computing power resource library management node and the computing power data center node of the central node in the MEC. The computing power resource library management node includes data processing nodes and data storage nodes.

[0249] The management domain connection unit is configured as follows:

[0250] The management of computing power resource libraries in different levels of MEC uses data processing middleware to process the real-time data of computing power resources provided by the MEC computing power gateway, forming MEC computing power mart data, and updating it to the local data nodes of the distributed data architecture MPP.

[0251] The data processing node completes the metadata processing and connects to the query terminal in the MEC computing power gateway to perform computing power resource queries and other operations.

[0252] The data storage nodes complete the storage of computing power data and SQL execution, including: the data storage nodes store computing power data on the data nodes of the MPP distributed data architecture according to the MPP data writing strategy, and the MPP data storage nodes convert the data into the corresponding format for storage and generate relevant indexes;

[0253] The central node manages all computing power data nodes in the network and sends real-time changes in computing power data to relevant MEC data nodes via a message mechanism.

[0254] Furthermore,

[0255] The data processing nodes include read-write nodes, read-only nodes, and monitoring nodes;

[0256] The read-write nodes are used to complete various data processing tasks for reading and writing computing power data. After the write processing is completed, a message mechanism is used to notify other read-only nodes to synchronize.

[0257] The read-only nodes are used to perform operations such as reading computing power data and are dynamically scheduled by the supervisory nodes;

[0258] The monitoring node is used to monitor the operational status of local computing resource processing, and to add and unload computing resource processing instances based on the monitoring results, so as to ensure that computing data processing meets the indicators set by the system.

[0259] Furthermore, the management domain connection unit is also configured as follows:

[0260] When performing SQL calculations, the SQL statement is first converted into logical execution units according to its semantics, and then broken down into specific physical execution units according to the distribution of the data. Finally, it is executed on the corresponding MEC data storage node to realize the mapping from SQL to specific data storage MEC.

[0261] Furthermore, the computing power map module 14 is specifically configured as follows:

[0262] Based on the content of map digitization, the needs of geographic information processing and storage, and the different query requirements of computing power resources, a corresponding layer scheme for computing power maps is formulated.

[0263] Digitally encode computing resources;

[0264] Determine the number N of zoom levels provided by the computing power map, take the map image with the highest zoom level and the largest map scale as the bottom layer of the pyramid, define the operation rules of computing power resource data, form the 0th layer tile matrix, generate the next layer of computing power tiles according to the tile segmentation rules, and define the data operation rules of the 1st layer of computing power tiles, forming the 1st layer tile matrix; generate the tile matrices of each layer in sequence until the entire tile pyramid is formed.

[0265] The computing power tiles are stored in a distributed mode. The computing power tiles of each MEC access area are stored on the local nodes of the MPP data nodes. The central node stores the computing power map data and the computing power resource data corresponding to the tile data. According to the dynamic changes of computing power resource data, the relevant tile storage nodes are notified through a message mechanism to keep the computing power resource data of the entire system synchronized. The computing power tile data of each MEC is provided through the tile server in the MEC.

[0266] Furthermore, the layer scheme of the computing power map includes:

[0267] General polygon layer: includes the structure of the base map, obtained from the general map;

[0268] Computing resource location layer: generated from the location data of computing resources, used to represent the location of computing resources, and supports various location-based queries and operations;

[0269] MEC Access Boundary Layer: Includes the range and boundaries of computing resources at different MEC access points, supporting various MEC-based queries and operations.

[0270] Computing power resource type graph: Includes different computing power types, supports various queries and operations based on computing power type, including general computing power type, special computing power type, scientific computing type, engineering computing type, and intelligent computing type;

[0271] Access line layer: includes different access bandwidth types and access bandwidth indicators, and supports various queries and calculations based on access line type;

[0272] Tariff layer: Includes different tariff types and tariff ranges, supporting various tariff-based queries and calculations;

[0273] Resource Source Layer: Includes different sources of computing power resources, supporting various queries and operations based on resource source;

[0274] Composite graphics: New layers created by combining the above layers, supporting queries and operations based on composite features.

[0275] Based on the equipment and network resources of existing telecommunications networks, the data integration and search capabilities of electronic maps, and the data display capabilities based on geographical locations, this disclosure constructs a method for integrating computing resources based on computing power maps. This method enables the access and management of various computing resources across the entire network, as well as data integration and search functions. It provides a feasible path for integrating computing resources and has positive significance for promoting the development and popularization of computing power networks.

[0276] The integrated computing resource access management system of this disclosure is used to implement the integrated computing resource access management method in Method Embodiment 1 and Method Embodiment 2, so the description is relatively simple. For details, please refer to the relevant descriptions in Method Embodiment 1 and Method Embodiment 2 above, which will not be repeated here.

[0277] In addition, such as Figure 6 As shown, Embodiment 4 of this disclosure also provides an electronic device, including a memory 100 and a processor 200. The memory 100 stores a computer program. When the processor 200 runs the computer program stored in the memory 100, the processor 200 executes the various possible methods described above.

[0278] The memory 100 is connected to the processor 200. The memory 100 can be a flash memory, a read-only memory, or another type of memory. The processor 200 can be a central processing unit or a microcontroller.

[0279] Furthermore, embodiments of this disclosure also provide a computer-readable storage medium storing a computer program, which is executed by a processor using the various possible methods described above.

[0280] The computer-readable storage medium includes volatile or non-volatile, removable or non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, computer program modules or other data). Computer-readable storage media includes, but is not limited to, RAM (Random Access Memory), ROM (Read-Only Memory), EEPROM (Electrically Erasable Programmable Read-Only Memory), flash memory or other memory technologies, CD-ROM (Compact Disc Read-Only Memory), Digital Video Disc (DVD) or other optical disc storage, cartridges, magnetic tapes, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and is accessible to a computer.

[0281] It is understood that the above embodiments are merely exemplary embodiments used to illustrate the principles of this disclosure, and this disclosure is not limited thereto. For those skilled in the art, various modifications and improvements can be made without departing from the spirit and substance of this disclosure, and these modifications and improvements are also considered to be within the scope of protection of this disclosure.

Claims

1. A method for integrated access management of computing resources, characterized in that, The method includes: Access to different computing resources: Connect different computing resources through the central node and MEC nodes at all levels of the multi-access edge computing MEC architecture, and complete the lifecycle management of computing resources through MEC; Computing resource data integration: The integration of computing resource data is divided into a data domain and a management domain; the data domain is integrated based on a massively parallel processing (MPP) distributed data architecture, providing data storage methods and data organization forms for the integrated computing resource data of the entire network; in the management domain, the storage and operation of computing resources, as well as the management of computing data nodes, are completed through computing resource management. Computing power data storage: Adopting an MPP distributed data architecture, while preserving the independence of MEC data nodes at each level, it realizes distributed storage, overall management and data retrieval of data; Computing power map implementation: Utilizing existing electronic map technology, layers and computing power tiles are planned according to the demand for computing power resources to form a computing power resource map, providing services for computing power resource search; Computing resource search: Utilizing a computing resource map and based on the search application programming interface (API) provided by the map, it offers various search services for computing resources; Computing resource routing: Provides optimized network connection routes for computing resource users and selected computing resources, generating optimized computing resource routes based on the path planning API provided by the map and computing resource map data; The method utilizes existing electronic map technology to plan layers and computing power tiles according to the demand for computing power resources, forming a computing power resource map to provide services for computing power resource search, including: Based on the content of map digitization, the needs of geographic information processing and storage, and the different query requirements of computing power resources, a corresponding layer scheme for computing power maps is formulated. Digitally encode computing resources; Determine the number N of zoom levels provided by the computing power map, take the map image with the highest zoom level and the largest map scale as the bottom layer of the pyramid, define the operation rules of computing power resource data, form the 0th layer tile matrix, generate the next layer of computing power tiles according to the tile segmentation rules, and define the data operation rules of the 1st layer of computing power tiles, forming the 1st layer tile matrix; generate the tile matrices of each layer in sequence until the entire tile pyramid is formed. The computing power tiles are stored in a distributed mode. The computing power tiles of each MEC access area are stored on the local nodes of the MPP data nodes. The central node stores the computing power map data and the computing power resource data corresponding to the tile data. According to the dynamic changes of computing power resource data, the relevant tile storage nodes are notified through a message mechanism to keep the computing power resource data of the entire system synchronized. The computing power tile data of each MEC is provided through the tile server in the MEC.

2. The method according to claim 1, characterized in that, The computing resources include the computing resources of the cloud center, the computing resources in the MEC, and the customer computing resources accessed through the MEC. The process of connecting different computing resources through the central node and various levels of MEC nodes in the MEC architecture, and managing the lifecycle of computing resources through MEC, includes: The computing resources of the cloud center are accessed through the unified cloud management platform of the central node, and an authentication system based on the MEC management system is adopted. The lifecycle management of computing resources is realized through cloud computing network management. For the computing resources in the MEC and the customer computing resources accessed through the MEC, based on the four deployment modes of the MEC's ​​distribution network elements and the MEC resource platform, the lifecycle management of the computing resources is realized by the MEC's ​​API gateway after the computing power gateway converts the protocol.

3. The method according to claim 1, characterized in that, The data interconnection of data domains based on the massively parallel processing (MPP) distributed data architecture includes: In terms of data processing, the computing power data processing tasks are distributed in parallel to multiple related nodes through MPP. After the data processing is completed on each node, the results of each part are aggregated to obtain the final result. In terms of data structure, the MPP node data of local computing resources in MEC is organized in the form of a data mart.

4. The method according to claim 1, characterized in that, The management domain includes the computing power resource library management node and the computing power data center node of the central node in the MEC. The computing power resource library management node includes data processing nodes and data storage nodes. The management of computing resources in the management domain, including storage and operation processing of computing resources and management of computing data nodes, includes: The management of computing power resource libraries in different levels of MEC uses data processing middleware to process the real-time data of computing power resources provided by the MEC computing power gateway, forming MEC computing power mart data, and updating it to the local data nodes of the distributed data architecture MPP. The data processing node completes the metadata processing and connects to the query terminal in the MEC computing power gateway to perform computing power resource queries and other operations. The data storage nodes complete the storage of computing power data and the execution of structured query language (SQL) tasks, including: the data storage nodes store computing power data on the data nodes of the MPP distributed data architecture according to the MPP data writing strategy, and the MPP data storage nodes convert the data into the corresponding format for storage and generate relevant indexes; The central node manages all computing power data nodes in the network and sends real-time changes in computing power data to relevant MEC data nodes via a message mechanism.

5. The method according to claim 4, characterized in that, The data processing nodes include read-write nodes, read-only nodes, and monitoring nodes; The read-write nodes are used to complete various data processing tasks for reading and writing computing power data. After the write processing is completed, a message mechanism is used to notify other read-only nodes to synchronize. The read-only nodes are used to perform operations such as reading computing power data and are dynamically scheduled by the supervisory nodes; The monitoring node is used to monitor the operational status of local computing resource processing, and to add and unload computing resource processing instances based on the monitoring results, so as to ensure that computing data processing meets the indicators set by the system.

6. The method according to claim 4, characterized in that, The SQL execution work includes: When performing SQL calculations, the SQL statement is first converted into logical execution units according to its semantics, and then broken down into specific physical execution units according to the distribution of the data. Finally, it is executed on the corresponding MEC data storage node, thus realizing the mapping from SQL to specific data storage MEC.

7. The method according to claim 1, characterized in that, The layer scheme of the computing power map includes: General polygon layer: includes the structure of the base map, obtained from the general map; Computing resource location layer: generated from the location data of computing resources, used to represent the location of computing resources, and supports various location-based queries and operations; MEC Access Boundary Layer: Includes the range and boundaries of computing resources at different MEC access points, supporting various MEC-based queries and operations; Computing power resource type graph: Includes different computing power types, supports various queries and operations based on computing power type, including general computing power type, special computing power type, scientific computing type, engineering computing type, and intelligent computing type; Access line layer: includes different access bandwidth types and access bandwidth indicators, and supports various queries and calculations based on access line type; Tariff layer: Includes different tariff types and tariff ranges, supporting various tariff-based queries and calculations; Resource Source Layer: Includes different sources of computing power resources, supporting various queries and operations based on resource source; Composite graphics: New layers created by combining the above layers, supporting queries and operations based on composite features.

8. An integrated access management system for computing resources, characterized in that, The system includes: The access module is designed to enable access to different computing resources: it connects different computing resources through the central node and various levels of MEC nodes in the MEC architecture, and manages the lifecycle of computing resources through MEC. The integration module is designed to integrate computing resource data: it divides the integration of computing resource data into a data domain and a management domain; it implements data integration in the data domain based on the MPP distributed data architecture, and provides data storage methods and data organization forms for the integrated computing resource data of the entire network; in the management domain, it completes the storage and operation processing of computing resources, as well as the management of computing data nodes through computing resource management. The storage module is designed to store computing power data: it adopts an MPP distributed data architecture, which, while preserving the independence of MEC data nodes at each level, enables distributed storage, overall management, and data retrieval. The computing power map module is designed to implement a computing power map: using existing electronic map technology, it plans layers and computing power tiles according to the demand for computing power resources to form a computing power resource map and provide services for computing power resource search. The search module is designed to enable the search of computing resources: by utilizing the computing power map and based on the search API interface provided by the map, it provides various search services for computing resources. The routing module is configured to implement computing resource routing: provide optimized network connection routes for computing resource users and selected computing resources, and generate optimized computing resource routes based on the path planning API provided by the map and computing resource map data; The computing power map module is specifically configured as follows: Based on the content of map digitization, the needs of geographic information processing and storage, and the different query requirements of computing power resources, a corresponding layer scheme for computing power maps is formulated. Digitally encode computing resources; Determine the number N of zoom levels provided by the computing power map, take the map image with the highest zoom level and the largest map scale as the bottom layer of the pyramid, define the operation rules of computing power resource data, form the 0th layer tile matrix, generate the next layer of computing power tiles according to the tile segmentation rules, and define the data operation rules of the 1st layer of computing power tiles, forming the 1st layer tile matrix; generate the tile matrices of each layer in sequence until the entire tile pyramid is formed. The computing power tiles are stored in a distributed mode. The computing power tiles of each MEC access area are stored on the local nodes of the MPP data nodes. The central node stores the computing power map data and the computing power resource data corresponding to the tile data. According to the dynamic changes of computing power resource data, the relevant tile storage nodes are notified through a message mechanism to keep the computing power resource data of the entire system synchronized. The computing power tile data of each MEC is provided through the tile server in the MEC.

9. An electronic device, characterized in that, It includes a memory and a processor, wherein the memory stores a computer program, and when the processor runs the computer program stored in the memory, the processor executes the integrated access management method for computing resources as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the integrated access management method for computing resources as described in any one of claims 1-7.

Images