A heterogeneous multi-cloud computing power unified management computing power service platform and method

Abstract

The application relates to a computing power service platform for unified management of heterogeneous multi-cloud computing power, wherein a multi-cloud access layer is directly designed in a government affair cloud and heterogeneous computing power resources, and serves as a bridge between the heterogeneous computing power resources and an intranet of the government affair cloud. The multi-cloud access layer abstracts computing power resources in different cloud environments into unified resource types and accesses the intranet of the government affair cloud. An adapter is written for each computing power service provider, conversion between unified resource interface requests and original APIs of specific computing power service providers is realized, and the compatibility and flexibility of the platform are enhanced. The application realizes integration of heterogeneous computing power across platforms and across operators. Users can conveniently use computing demands of various scales and types, and the application has high flexibility and scalability. The application uses a full homomorphic encryption algorithm in the multi-cloud platform, ensures that data analysis and calculation across cloud platforms are carried out in an encrypted state, and further improves the security of the data.

Description

Technical Field

[0001] This invention relates to the field of cloud computing technology, and in particular to a computing power service platform and method for unified management of heterogeneous multi-cloud computing power. Background Technology

[0002] With the rapid development of the computing power industry, computing power has become a key force driving the digital transformation of the economy and society. Faced with the ever-increasing computing power demands for large-scale AI model training, connected vehicles, and edge computing, how to efficiently integrate and utilize diverse computing resources has become an urgent problem to be solved. The diversity, heterogeneity, and dispersion of computing resources make unified management and scheduling of these resources particularly important. To meet this need, building a public computing power service platform to achieve the convergence and supply of computing resources across regions, architectures, and service providers is of great significance for improving the utilization rate of computing resources, reducing the barriers and costs for users to access computing services, promoting the implementation of intelligent computing applications, and empowering the digital transformation of traditional industries.

[0003] Current computing power scheduling platforms provide computing power services to users by integrating and scheduling their own heterogeneous computing power resources, including various computing power collaborative processing systems such as CPUs, GPUs, FPGAs, and ASICs. However, these platforms mainly focus on the scheduling and management of their own computing power resources, lacking the ability to integrate heterogeneous computing power across platforms and carriers. Furthermore, current platforms cannot uniformly manage all general computing cloud computing power resources, supercomputing, and intelligent computing power resources within a region at the government or regional level, resulting in fragmented computing power resources and difficulty in forming a synergy. Computing power demanders (such as universities, research institutes, and startups) find it difficult to find suitable computing power resources, while computing power providers lack an open platform to showcase and provide their computing power resources, leading to a poor match between supply and demand in the computing power market. Summary of the Invention

[0004] In view of this, it is necessary to provide a computing power service platform for unified management of heterogeneous multi-cloud computing power in order to solve the above-mentioned defects of existing technologies.

[0005] To address the aforementioned problems, in a first aspect, embodiments of the present invention provide a computing power service platform for unified management of heterogeneous multi-cloud computing power, comprising:

[0006] The multi-cloud access layer connects to heterogeneous computing resources on one end and the government cloud intranet on the other. It is used to abstract computing resources from different cloud environments into a unified resource type and connect them to the government cloud intranet. The heterogeneous computing resources include public cloud computing resources and private cloud computing resources.

[0007] The resource management layer is used to unify access, management, and scheduling of heterogeneous multi-cloud computing resources, and provides a visual monitoring and operation interface.

[0008] Preferably, the public cloud computing resources are connected to the government cloud intranet via the public network, and the private cloud computing resources are connected to the government cloud intranet via fiber optic or dedicated line networks.

[0009] Preferably, the multi-cloud access layer includes a resource abstraction module, which is used to abstract computing resources from different cloud environments into a unified resource type; it is also used to write an adapter for each computing service provider to convert the unified resource interface request into the native API call of the specific computing service provider.

[0010] Preferably, the resource management layer includes:

[0011] The resource registration module provides resource registration services, registering the connected computing resources into a unified resource pool.

[0012] The resource pool management module is used to integrate and manage computing resources in different cloud environments to establish a unified resource pool; it provides API interfaces for the system to query, allocate and operate resources; and it provides supercomputing resource management capabilities, including queue management, job management and file management.

[0013] The file storage management module provides file storage synchronization, file storage creation, file storage deletion, file storage search, and file storage retrieval functions;

[0014] The object storage management module provides functions for object storage synchronization, object storage creation, object storage deletion, object storage search, and object storage retrieval.

[0015] Preferably, the resource management layer further includes:

[0016] The health check module is used to periodically check the status of computing resources and update the resource status based on the check results.

[0017] Preferably, the resource management layer further includes:

[0018] The VPC management module provides VPC resource management functions for heterogeneous computing power platforms, including: VPC synchronization, VPC creation, VPC deletion, VPC search, and VPC retrieval; where VPC stands for Virtual Private Cloud.

[0019] The EIP management module provides subnet resource management functions for heterogeneous computing platforms, including subnet synchronization, subnet creation, subnet deletion, subnet search, and subnet retrieval; where EIP stands for Elastic Public IP.

[0020] Preferably, the resource management layer further includes:

[0021] The security group management module provides functions such as security group synchronization, security group creation, security group deletion, security group search, and security group retrieval. It is also used to implement fully homomorphic encryption algorithms on multi-cloud platforms, enabling cross-cloud platform data analysis and computation while data is encrypted.

[0022] Preferably, the use of a fully homomorphic encryption algorithm on a multi-cloud platform specifically includes:

[0023] When a customer obtains the cloud and supercomputing resources they need, the system automatically generates a public and private key pair.

[0024] When a customer stores raw data on the server, the server uses RSA homomorphic encryption, Paillier homomorphic encryption, or SEAL homomorphic encryption to encrypt the raw data using a public key and generate ciphertext.

[0025] When a client needs to view the results, they submit a decryption request to the server, which then uses its private key to decrypt the encrypted calculation results.

[0026] Secondly, embodiments of the present invention provide a heterogeneous multi-cloud computing power network management method for a computing power service platform, comprising:

[0027] The computing resources of different cloud environments are abstracted into a unified resource type and connected to the government cloud intranet; wherein, the heterogeneous computing resources include public cloud computing resources and private cloud computing resources.

[0028] The public cloud computing resources are connected to the government cloud intranet via the public network, while the private cloud computing resources are connected to the government cloud intranet via fiber optic or dedicated line networks.

[0029] This invention provides a unified computing power service platform for heterogeneous multi-cloud computing power management. It abstracts computing resources from different cloud environments into a unified resource type and connects them to the government cloud intranet. This enables cross-platform and cross-carrier heterogeneous computing power integration. It facilitates users' computing needs of different scales and types, offering high flexibility and scalability. Attached Figure Description

[0030] Figure 1 This is an architecture diagram of the computing power service platform for unified management of heterogeneous multi-cloud computing power provided by the present invention.

[0031] Figure 2 A structural block diagram of the computing power service platform for unified management of heterogeneous multi-cloud computing power provided by the present invention;

[0032] Figure 3 Flowchart of the heterogeneous multi-cloud computing power network management method for the computing power service platform provided by the present invention. Detailed Implementation

[0033] Preferred embodiments of the present invention will now be described in detail with reference to the accompanying drawings, which form part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not intended to limit the scope of the present invention.

[0034] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0035] Current computing power scheduling platforms provide computing power services to users by integrating and scheduling their own heterogeneous computing power resources, including various computing power collaborative processing systems such as CPUs, GPUs, FPGAs, and ASICs. However, these platforms mainly focus on the scheduling and management of their own computing power resources, lacking the ability to integrate heterogeneous computing power across platforms and carriers. Furthermore, current platforms cannot uniformly manage all general computing cloud computing power resources, supercomputing, and intelligent computing power resources within a region at the government or regional level, resulting in fragmented computing power resources and difficulty in forming a synergy. Computing power demanders (such as universities, research institutes, and startups) find it difficult to find suitable computing power resources, while computing power providers lack an open platform to showcase and provide their computing power resources, leading to a poor match between supply and demand in the computing power market.

[0036] In view of this, this invention proposes a computing power service platform and method for unified management of heterogeneous multi-cloud computing power. It abstracts computing resources from different cloud environments into a unified resource type and connects them to the government cloud intranet. This achieves cross-platform and cross-carrier heterogeneous computing power integration. It facilitates users' computing needs of different scales and types, and has high flexibility and scalability. The following will elaborate and describe this through several embodiments.

[0037] Figure 1 This is an architecture diagram of the computing power service platform for unified management of heterogeneous multi-cloud computing power provided by the present invention, with reference to... Figure 1 The computing power service platform for unified management of heterogeneous multi-cloud computing power provided by this invention includes at least:

[0038] The multi-cloud access layer connects to heterogeneous computing resources on one end and the government cloud intranet on the other. It is used to abstract computing resources from different cloud environments into a unified resource type and connect them to the government cloud intranet. The heterogeneous computing resources include public cloud computing resources and private cloud computing resources.

[0039] The resource management layer is used to unify access, management, and scheduling of heterogeneous multi-cloud computing resources, and provides a visual monitoring and operation interface.

[0040] Reference Figure 1 This invention employs Domain-Driven Design (DDD) principles, abstracting the concept of a multi-cloud access layer. One end of the multi-cloud access layer interfaces with heterogeneous computing resources, enabling communication and interaction with various cloud environments (such as public and private clouds) supporting different technology stacks and protocols. It requires designing access schemes tailored to the specific requirements of each cloud environment. The other end of the multi-cloud access layer provides a unified abstract interface for accessing the government cloud intranet. For the upstream government cloud environment, the multi-cloud access layer provides a unified interface, allowing applications and services within the government cloud to access and use heterogeneous computing resources in a standardized manner. This abstraction layer hides the complexity of the underlying cloud environment, freeing upper-layer applications from concern themselves with the specific cloud implementation details.

[0041] The multi-cloud access layer includes a resource abstraction module, which abstracts computing resources from different cloud environments into a unified resource type. It also develops adapters for each computing service provider, converting unified resource interface requests into native API calls specific to that provider, and converting the response results back to a unified format. The adapters implement necessary logic to handle functional differences and naming inconsistencies between cloud service providers, ensuring that upper-layer applications can perform cross-cloud resource operations through unified interface calls. The module receives resource requests from the application layer, parses them, and maps them to appropriate resource abstraction types and configurations.

[0042] Understandably, different cloud service providers may offer different types of resources, such as virtual machines, containers, and bare metal servers. To manage and schedule these resources uniformly, the system abstracts them, defining a unified set of resource types, such as computing resources, storage resources, and network resources. These abstracted resource types are called resource abstract types. When the application layer (user interface) sends a resource request to the system, the system first receives the request. Then, the system parses the request to understand what resources the application layer needs, including the type, quantity, and specifications of the resources. Next, the system maps the request to the previously defined resource abstract types and configurations based on this information. In this way, regardless of the specific resource requested by the application layer, the system can find the matching computing resources through resource abstraction and configuration, and perform unified scheduling and management.

[0043] In this embodiment of the invention, for private cloud resources, a dedicated network line is designed to connect the government cloud intranet server and the private cloud provider's cloud host server. This dedicated line connection ensures data security and transmission efficiency. For supercomputing resources (usually bare metal resources), they are directly connected in a bare-metal configuration, with public IP ports exposed and root privileges provided for higher-level management and configuration. For public cloud resources, the intelligent computing resource management interface provided by the public cloud provider is used to access the multi-cloud access layer. These interfaces typically include APIs for account management and resource operations, enabling the multi-cloud access layer to manage and schedule public cloud resources in a standardized manner.

[0044] To ensure the uniqueness and identifiability of resources across systems and cloud environments, a unified resource identification system was invented and designed. Each resource identifier is unique and clearly reflects the essential attributes of the resource, such as type, status, location, and specifications. This identification system is compatible with existing standards (such as OpenStack and VMware), which is beneficial for system integration and interoperability.

[0045] This invention integrates computing resources from different computing service providers within a government cloud environment. Upstream, it connects to the enterprise account system within the government cloud, and downstream, it connects to computing resources, facilitating enterprises to directly order and purchase computing resources through the government platform. By connecting public and private clouds to the government cloud intranet via a large computing network, the computing service platform can achieve the following functions:

[0046] Supports access to the computing resources of public cloud computing service providers through the public network environment;

[0047] It supports accessing the computing resources of private cloud computing service providers through intranet network environments such as fiber optic cables and dedicated lines;

[0048] It supports the ability to visualize the overall network overview, including CPU usage, number of accelerator cards, memory capacity, and storage capacity;

[0049] Supports the number of cloud computing service providers, the number of resource pools, the cumulative number of scheduling requests, and the cumulative number of deployed tasks;

[0050] Supports the cyclical usage trends of computing resources such as CPU, accelerator cards, and memory;

[0051] Supports the distribution of computing power across general computing, intelligent computing, and supercomputing scenarios;

[0052] Supports the presentation of computing power access-related data, and supports custom queries of information such as the cumulative number of customers, the cumulative number of computing network tasks, the cumulative number of connected computing power service providers, and the cumulative number of connected resource pools;

[0053] Supports daily customer count visualization and periodic display of network task quantity;

[0054] It supports customer industry analysis by attribute and displays the time-series progress.

[0055] This invention provides a unified computing power service platform for heterogeneous multi-cloud computing power management. It abstracts computing resources from different cloud environments into a unified resource type and connects them to the government cloud intranet. This enables cross-platform and cross-carrier heterogeneous computing power integration. It facilitates users' computing needs of different scales and types, offering high flexibility and scalability.

[0056] The resource management layer can be deployed on a government cloud server, using the same intranet environment and adhering to the same network security policies as the government cloud. The resource management layer connects to downstream computing resources such as public and private clouds through a multi-cloud access layer. In a preferred embodiment of the invention, the resource management layer further includes the following modules:

[0057] The resource registration module provides resource registration services, registering connected computing resources into a unified resource pool, including resource type, status, and performance metrics. Basic verification is performed during the registration process to ensure the submitted information is complete and legal, preventing invalid or malicious resource registrations. A resource information database is established to store detailed information and real-time status of all registered resources, ensuring high availability and data consistency. Resource information is effectively indexed, supporting fast querying and filtering, such as searching by resource type, status, and tags.

[0058] The resource pool management module integrates and manages computing resources across different cloud environments, establishing a unified resource pool. It provides API interfaces for the system to query, allocate, and operate resources. Resource discovery and management plugins are developed for different computing platforms (such as public and private clouds) to capture and synchronize resource information. New resource creation and changes are detected using APIs or event notification services provided by the underlying platform, and these are registered in the resource pool. The system can query, allocate, and operate resources via APIs, maintaining operational consistency. Supercomputing resource management capabilities are provided, including queue management, job management, and file management. Queue management includes queue synchronization, queue creation, queue deletion, queue search, and queue retrieval. Job management allows for job submission via command lines, job scripts, and scheduler scripts, supporting instance selection, resource requirements setting, and working directory configuration. File management enables storage in a file system format, supporting viewing file lists, creating folders, uploading files, downloading files, deleting files / folders, previewing files, and editing files.

[0059] The file storage management module provides file storage synchronization, file storage creation, file storage deletion, file storage search, and file storage retrieval functions. Through API interfaces, it enables unified access and management of heterogeneous computing resources from various computing power providers, obtains resource information, and can initiate operations such as resource creation, modification, and deletion.

[0060] The object storage management module provides functions for object storage synchronization, creation, deletion, lookup, and retrieval. It supports using storage as a file system, allowing operations such as viewing file lists, creating folders, uploading files, downloading files, and deleting files / folders. For large files, it supports uploading in chunks; it supports retrieving dataset lists and model lists; and it supports creating storage within a resource pool (cluster) and allocating a specified amount of space. It also supports viewing storage information.

[0061] The health check module is used to periodically check the status of computing resources and update the resource status based on the check results. For example, if a computing resource is automatically marked as faulty or under maintenance if it experiences N consecutive lost heartbeats or abnormal performance indicators.

[0062] The VPC management module provides VPC resource management functions for heterogeneous computing power platforms, including: VPC synchronization, VPC creation, VPC deletion, VPC search, and VPC retrieval; where VPC stands for Virtual Private Cloud.

[0063] The EIP management module provides subnet resource management functions for heterogeneous computing platforms, including subnet synchronization, subnet creation, subnet deletion, subnet search, and subnet retrieval; where EIP stands for Elastic Public IP.

[0064] In a preferred embodiment of the present invention, the resource management layer further includes:

[0065] The security group management module provides functions such as security group synchronization, security group creation, security group deletion, security group search, and security group retrieval. It is also used to implement fully homomorphic encryption algorithms on multi-cloud platforms. For situations where data may be stored in different cloud service providers, fully homomorphic encryption allows users to perform data calculations and analyses without exposing the data content. At the same time, cross-cloud platform data analysis can be performed directly in encrypted form, avoiding the security risks associated with data decryption.

[0066] In this embodiment, a fully homomorphic encryption algorithm is used on the multi-cloud platform. Specifically, when a customer obtains the required cloud resources and supercomputing resources, the system automatically generates a pair of public and private keys. When the customer stores the original data on the server, the server uses RSA homomorphic encryption, Paillier homomorphic encryption, or SEAL homomorphic encryption to encrypt the original data using the public key and generate ciphertext. When the customer needs to view the results, they submit a decryption request to the server, and the server uses the private key to decrypt the encrypted calculation results.

[0067] Figure 2 This is a structural block diagram of the computing power service platform for unified management of heterogeneous multi-cloud computing power provided by the present invention. (Refer to...) Figure 2 The platform provides operational services, core services, and basic services. By utilizing functions within the core and basic services, such as computing power coding and account systems, it achieves effective integration and utilization of resources. Figure 2 As shown, the computing power service platform ( Figure 2 The medium-power computing scheduling service platform also provides the following functions

[0068] Resource Classification: When defining resource identifiers, in addition to distinguishing between broad categories such as hosts, storage, and networks, it is also necessary to refine them to specific subcategories, such as CPU model, GPU model, memory, hard drive type (SSD, HDD), network bandwidth, etc.

[0069] Tagging System: Introducing a tagging system allows users to customize tags (e.g., Environment: Production, Testing, Development; Project: XX System; Region: Asia), enhancing the searchability and management flexibility of resources.

[0070] Version management: The resource identifier contains version information, which makes it easy to track changes in resources over time, manage upgrades, and maintain historical records.

[0071] Interface Definition: Define a standard API interface, based on protocols such as RESTful, providing functions such as resource registration, querying, status updating, and task scheduling. Use standard protocols such as OAuth 2.0 and JWT for user authentication and access control to ensure that each API call is initiated by a legitimate user. Ensure that PUT and PATCH requests are idempotent, meaning that multiple executions have the same effect, avoiding inconsistent resource states due to duplicate requests.

[0072] Version control: Supports API version management mechanisms, such as / v1, / v2, and supports backward compatibility, facilitating iteration without interrupting old services.

[0073] Resource discovery: Implement a computing resource discovery mechanism to discover host, GPU, storage, and network resources on the cloud platform through APIs and services.

[0074] The computing power service platform for unified management of heterogeneous multi-cloud computing power provided by this invention can achieve the following beneficial effects:

[0075] 1. Unify the specifications of computing power products from multiple computing power operators within the region, unify the form of computing power products, and unify the delivery method of computing power, so as to facilitate users to meet computing needs of different scales and types, and have high flexibility and scalability.

[0076] 2. The platform has optimized the process for users to purchase computing power, providing a complete service from selection, ordering, configuration to usage. Users can complete the purchase and configuration of computing resources with simple operations on the platform, eliminating the need for cumbersome communication with multiple computing power providers. The platform automatically allocates and configures computing resources based on the user's selections and configuration requirements, significantly improving purchasing efficiency.

[0077] 3. The billing system has been optimized. The platform will record users' usage and billing records in detail, including usage time, amount used, and cost. All this information will be presented to users in the form of a bill, allowing them to view their usage at any time and ensuring transparent and unhidden costs.

[0078] 4. The platform connects the account systems of enterprises and various computing power providers within the region through a unified account system. This allows the government to easily manage enterprise accounts, such as monitoring usage and calculating costs. It also provides a robust access control mechanism, ensuring that different users can only access resources and information within their authorized scope, thus guaranteeing data security and privacy.

[0079] Figure 3 The flowchart of the heterogeneous multi-cloud computing power network management method of the computing power service platform provided by the present invention is shown below. Figure 3 This invention provides a method for managing heterogeneous multi-cloud computing power networks in a computing power service platform, comprising:

[0080] The computing resources of different cloud environments are abstracted into a unified resource type and connected to the government cloud intranet; wherein, the heterogeneous computing resources include public cloud computing resources and private cloud computing resources.

[0081] The public cloud computing resources are connected to the government cloud intranet via the public network, while the private cloud computing resources are connected to the government cloud intranet via fiber optic or dedicated line networks.

[0082] Specifically, public cloud computing resources are typically provided publicly via the internet, featuring elasticity, scalability, and pay-as-you-go pricing. Public cloud computing service providers offer a series of API interfaces, such as account management, resource operations, and container operations. This invention adapts each public cloud API and employs domain-driven design principles to unify these different APIs into a single resource scheduling system. In this way, regardless of which public cloud service is used, management and scheduling can be performed through a unified interface.

[0083] Private cloud computing resources are typically deployed within enterprises or organizations, offering enhanced security and controllability. The government cloud establishes dedicated network lines with private cloud vendors within the region, using multimode fiber optic cables to connect the government cloud hosts to the private cloud vendors' data centers. In this way, the private cloud is extended into the government cloud's internal network environment, ensuring both the availability of computing resources and the security of the overall network environment.

[0084] The computing power service platform of this invention is deployed on a government cloud server, adopting the same intranet environment and network security policies as the government cloud. It communicates with the external public network through designated IPs and ports, ensuring both the closed nature of the government cloud intranet environment and interconnectivity with external computing resources.

[0085] In multi-cloud platforms, data security and privacy are paramount. After a customer acquires cloud and supercomputing resources, the system automatically generates a public and private key pair. When the customer stores raw data on the server, the data is automatically fully homomorphically encrypted using the public key. Fully homomorphic encryption allows computation and analysis of the data in an encrypted state without decryption. When the customer needs to view the results, they use the private key to decrypt the encrypted computation. Fully homomorphic encryption technology ensures the security and privacy of data across multiple cloud platforms. Even if data is distributed across multiple cloud platforms, only the data owner can decrypt and view the final results, effectively avoiding the risks of data leakage and misuse.

[0086] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the appended claims.

[0087] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A computing power service platform for unified management of heterogeneous multi-cloud computing power, characterized in that, include: The multi-cloud access layer connects to heterogeneous computing resources on one end and the government cloud intranet on the other. It is used to abstract computing resources from different cloud environments into a unified resource type and connect them to the government cloud intranet. The heterogeneous computing resources include public cloud computing resources and private cloud computing resources. The resource management layer is used to unify access, management, and scheduling of heterogeneous multi-cloud computing resources, and provides a visual monitoring and operation interface.

2. The computing power service platform for unified management of heterogeneous multi-cloud computing power according to claim 1, characterized in that, The public cloud computing resources are connected to the government cloud intranet via the public network, while the private cloud computing resources are connected to the government cloud intranet via fiber optic or dedicated line networks.

3. The computing power service platform for unified management of heterogeneous multi-cloud computing power according to claim 1, characterized in that, The multi-cloud access layer includes a resource abstraction module, which is used to abstract computing resources from different cloud environments into a unified resource type; it is also used to write an adapter for each computing service provider to convert the unified resource interface request into the native API call of the specific computing service provider.

4. The computing power service platform for unified management of heterogeneous multi-cloud computing power according to claim 1, characterized in that, The resource management layer includes: The resource registration module provides resource registration services, registering the connected computing resources into a unified resource pool. The resource pool management module is used to integrate and manage computing resources in different cloud environments to establish a unified resource pool; it provides API interfaces for the system to query, allocate and operate resources; and it provides supercomputing resource management capabilities, including queue management, job management and file management. The file storage management module provides file storage synchronization, file storage creation, file storage deletion, file storage search, and file storage retrieval functions; The object storage management module provides functions for object storage synchronization, object storage creation, object storage deletion, object storage search, and object storage retrieval.

5. The computing power service platform for unified management of heterogeneous multi-cloud computing power according to claim 1, characterized in that, The resource management layer also includes: The health check module is used to periodically check the status of computing resources and update the resource status based on the check results.

6. The computing power service platform for unified management of heterogeneous multi-cloud computing power according to claim 1, characterized in that, The resource management layer also includes: The VPC management module provides VPC resource management functions for heterogeneous computing power platforms, including: VPC synchronization, VPC creation, VPC deletion, VPC search, and VPC retrieval; where VPC stands for Virtual Private Cloud. The EIP management module provides subnet resource management functions for heterogeneous computing platforms, including subnet synchronization, subnet creation, subnet deletion, subnet search, and subnet retrieval; where EIP stands for Elastic Public IP.

7. The computing power service platform for unified management of heterogeneous multi-cloud computing power according to claim 1, characterized in that, The resource management layer also includes: The security group management module provides functions such as security group synchronization, security group creation, security group deletion, security group search, and security group retrieval. It is also used to implement fully homomorphic encryption algorithms on multi-cloud platforms, enabling cross-cloud platform data analysis and computation while data is encrypted.

8. The computing power service platform for unified management of heterogeneous multi-cloud computing power according to claim 7, characterized in that, The use of fully homomorphic encryption algorithms on multi-cloud platforms specifically includes: When a customer obtains the cloud and supercomputing resources they need, the system automatically generates a public and private key pair. When a customer stores raw data on the server, the server uses RSA homomorphic encryption, Paillier homomorphic encryption, or SEAL homomorphic encryption to encrypt the raw data using a public key and generate ciphertext. When a client needs to view the results, they submit a decryption request to the server, which then uses its private key to decrypt the encrypted calculation results.

9. A method for managing heterogeneous multi-cloud computing power networks of a computing power service platform as described in any one of claims 1-8, characterized in that, include: The computing resources of different cloud environments are abstracted into a unified resource type and connected to the government cloud intranet; wherein, the heterogeneous computing resources include public cloud computing resources and private cloud computing resources. The public cloud computing resources are connected to the government cloud intranet via the public network, while the private cloud computing resources are connected to the government cloud intranet via fiber optic or dedicated line networks.

Images