Lightweight multi-domain hyper-fusion platform construction method and system compatible with heterogeneous hardware resources

Abstract

The invention belongs to the technical field of cloud computing, and discloses a lightweight multi-domain hyper-fusion platform construction method compatible with heterogeneous hardware resources, which adopts a lightweight hyper-fusion platform architecture compatible with the heterogeneous hardware resources, converts physical resources into virtual resources on a virtual machine based on a KVM (Kernel-based Virtual Machine), shields the difference of underlying hardware, and improves the performance of the platform. According to the method, heterogeneous storage resources are integrated based on Ceph, and storage devices of different manufacturers are integrated into a unified storage pool through a distributed storage technology without being limited by types of bottom hardware devices. Through the SDN technology, the OpenStack network service can decouple a network control plane and hardware, and unified management of different network devices is realized. Through a virtualization technology and an automatic management tool, a simplified elastic expansion solution is provided, and changes of resource demands are easily coped with.

Description

Technical Field

[0001] The present invention belongs to the technical field of cloud computing, and in particular relates to a method and system for constructing a lightweight multi-domain hyper-convergence platform compatible with heterogeneous hardware resources. Background Art

[0002] As cloud computing is combined with technologies in other cutting-edge fields, the use of hyper-converged architecture to provide cloud computing services has gradually become the mainstream deployment form in the industry. It is a delivery model with advantages such as simplified management, elastic expansion, and flexibility and agility. Hyper-converged architecture is based on a standard universal hardware platform, and implements the convergence of computing, storage, and network through software definition, and realizes the technical architecture of a software-defined data center centered on virtualization. It can simplify the deployment and application of IT infrastructure, reduce the operating costs of enterprises, more efficiently utilize computing, storage, and network resources, improve overall resource utilization, and more effectively protect data security through centralized management and monitoring. However, the existing hyper-converged cloud technology still has the following problems.

[0003] (1) Current hyper-converged cloud computing platforms are usually built on standardized homogeneous servers and are difficult to be compatible with heterogeneous hardware resource scenarios. Standard servers use hardware standards that are widely recognized in the industry, allowing hyper-converged platforms to be deployed in a variety of hardware environments with high compatibility and flexibility. The hardware components of standard servers (such as CPU, memory, storage devices, etc.) are usually interchangeable and upgradeable, making maintenance and management simpler and more efficient. In addition, standard servers are usually cheaper due to large-scale production and market competition. This makes the construction cost of hyper-converged platforms lower and can reduce the total cost of ownership through modular expansion of standard servers. The modular design of standard servers allows enterprises to gradually expand computing, storage, and network resources according to business needs. This flexibility allows hyper-converged platforms to start small and expand to large-scale deployment as the business grows. Therefore, current hyper-converged architectures are mostly built on homogeneous standard servers.

[0004] However, as business requirements continue to increase, heterogeneous service clusters are introduced into cloud computing services because of their diverse hardware resources, which can provide the best computing power for different types of workloads. In homogeneous clusters, all nodes have the same hardware configuration, which may result in some tasks not being able to fully utilize the advantages of specific hardware. For example, I / O-intensive tasks may not achieve the best performance on compute-intensive nodes. Heterogeneous clusters can select the most suitable nodes based on the characteristics of the task, thereby improving overall efficiency. Different hardware platforms have different performance when processing specific types of tasks. Heterogeneous clusters can perform performance tuning for different tasks. For example, compute-intensive tasks can be scheduled to high-performance CPU or GPU nodes, while I / O-intensive tasks can be scheduled to nodes with high-throughput storage systems. Through this targeted optimization, heterogeneous clusters can maximize the potential of various types of hardware to improve performance. Therefore, building a hyper-converged cloud platform compatible with heterogeneous hardware resources can effectively improve the performance and quality of cloud services and improve resource utilization efficiency.

[0005] The traditional hyper-converged platform hardware deployment method is bound to the hardware, and the software deployment method has low resource utilization efficiency. There are two ways to deploy traditional hyper-converged platforms: one is the hyper-converged all-in-one delivery method, and the other is the hyper-converged software delivery method. The hyper-converged all-in-one delivery method delivers pre-integrated hardware devices, including computing, storage, and network devices, but is usually bound to the manufacturer, and cannot freely choose hardware suppliers and configurations according to needs, and cannot expand and upgrade hardware configurations according to business needs, relying on the manufacturer's upgrade path; the hyper-converged software delivery method is usually deployed on metal bare metal in the form of an operating system, or deployed in a virtual machine on an existing virtualization platform. However, both software deployment methods rely on a complete virtualization layer, and need to create a control virtual machine on each physical node to manage the computing and storage resources of the node. Since the virtual machine needs to run a complete operating system instance and the virtualization layer needs to simulate hardware resources, it leads to high performance overhead and low node resource utilization.

[0006] With the popularization of distributed storage technology and virtualization technology, more and more computing and data processing needs have emerged. The platform deployment and management under the traditional cloud computing model are cumbersome and complex, and cannot meet the requirements of full resource utilization and flexible service expansion. In order to solve these challenges, the hyper-converged architecture was introduced, which allows the server's computing, storage, and network resources to be integrated and uniformly managed in a software-defined manner. By expanding hyper-converged nodes to expand service resources, the deployment and management work is simplified, the utilization and flexibility of resources are improved, and the management and labor costs are reduced.

[0007] The basic architecture of a hyper-converged platform consists of three parts: computing virtualization, storage virtualization, and network virtualization [1]. Computing virtualization uses virtualization technology to abstract physical computing resources into virtual machines (VMs) or containers, allowing multiple workloads to run on the same hardware resources. Storage virtualization virtualizes local storage resources on nodes (such as SSDs, HDDs, and NVMe) into a shared storage pool through distributed storage technology, providing high-performance, scalable data services. Network virtualization virtualizes physical network resources through software-defined networking (SDN) technology, allowing multiple virtual networks to coexist on the same physical network.

[0008] (2) Existing technical solutions

[0009] 1) Hyper-convergence platform construction solution

[0010] Major hyper-convergence vendors use different technologies to build hyper-convergence platforms. These solutions have their own characteristics in architecture, functions and deployment modes. The following is a detailed introduction to the hyper-convergence platform construction solutions of several well-known hyper-convergence vendors:

[0011] Nutanix[2] is one of the pioneers of hyper-converged architecture. Its hyper-converged platform is based on its core Acropolis Operating System (AOS), which integrates computing, storage, and network resources in a software-defined manner to provide a highly integrated solution. Nutanix's hyper-converged operating system AOS adopts a distributed storage architecture, integrates local storage resources in the cluster into a global storage pool, and manages storage and data services through the Nutanix ControllerVM (CVM). Nutanix's virtualization management program Acropolis Hypervisor (AHV) is developed based on KVM and is deeply integrated in AOS, providing a lightweight virtualization solution. Nutanix's centralized management platform Prism management platform provides cluster management, performance monitoring, automated operation and maintenance, backup and recovery, and other functions.

[0012] VMware's[3] hyper-convergence solution is based on its core technology vSAN (vSphere Storage Area Network), which integrates computing, storage, and network resources in the vSphere virtualization platform in a software-defined manner. VMware vSphere is the core of its hyper-convergence architecture and is responsible for computing virtualization management. vSphere runs virtual machines through the ESXi Hypervisor and provides powerful virtualization capabilities. vSAN integrates local storage in the cluster into a shared storage pool, supporting functions such as data distribution, data protection, and disaster recovery. vSAN is deeply integrated in the vSphere environment to provide a consistent storage management experience. NSX is VMware's software-defined networking (SDN) solution that supports network virtualization, security policy management, and network traffic optimization. It is an important component of VMware's hyper-convergence platform.

[0013] HPE SimpliVity is a hyper-converged solution from Hewlett Packard Enterprise (HPE) that integrates computing, storage and network resources. The core platform of SimpliVity, SimpliVity DataVirtualization Platform, provides data virtualization, optimization and protection functions. DVP achieves data compression and data deduplication through dedicated hardware acceleration cards, which improves storage efficiency and reduces storage requirements. SimpliVity is deployed on HPE's ProLiant servers, which provide powerful computing performance and scalability, and support virtualization and data storage integration.

[0014] The current hyper-converged cloud computing platform has poor compatibility with heterogeneous hardware resources and difficulty in elastic expansion. The existing hyper-converged cloud platform is designed to run in a homogeneous environment, assuming that all nodes have similar hardware and software configurations. This design simplifies resource management and scheduling, but also brings difficulties in compatibility with heterogeneous resources. Many hyper-converged solutions are highly dependent on specific hardware configurations during deployment and operation, making it difficult to support hardware from different manufacturers or with different configurations. Secondly, hyper-converged platforms are usually designed with fixed architectures and expansion units, which limits the flexibility of expansion. Hyper-converged platforms usually use nodes as expansion units, and each node contains computing, storage, and network resources. This fixed unit expansion method is difficult to cope with the expansion of different resource requirements. In addition, the expansion of many hyper-converged platforms is linear, requiring the hardware configuration of new nodes to be consistent with existing nodes, which also limits the flexible expansion of hyper-converged platforms customized according to business characteristics.

[0015] The traditional hyper-converged platform hardware deployment method is tied to the vendor, hardware selection is limited, and the software deployment method has low resource utilization efficiency. The traditional hyper-converged platform deployment methods can generally be divided into two types: one is the hyper-converged all-in-one delivery method, and the other is the hyper-converged software delivery method. There are significant differences between these two methods in terms of architecture, flexibility, resource utilization, etc. First, the hyper-converged all-in-one delivery method is achieved by delivering pre-integrated hardware devices, which usually contain computing, storage and network resources to form a highly integrated overall solution. The advantage of this method is that it simplifies the deployment process and hardware configuration work because the hardware and software have been integrated and optimized by the vendor. However, the limitation of this method is that it is usually closely tied to the hardware and software ecosystem of a specific vendor, and users have low flexibility in hardware selection and configuration. Specifically, users cannot freely choose hardware vendors or customize hardware configurations according to their own business needs. In addition, scalability is also limited. It is difficult for enterprises to expand or upgrade hardware on their own according to actual business development needs. Instead, they need to rely on the upgrade path and support plan predefined by the vendor, which to some extent limits the scalability and autonomy of the system. Secondly, the hyper-convergence software delivery method is to deploy the software components of the hyper-convergence platform directly on the bare metal server in the form of an operating system, or deploy them in a virtual machine through a virtualization platform. Compared with the all-in-one delivery method, the software delivery method provides greater flexibility, allowing users to deploy hyper-convergence software on the basis of existing hardware and supporting the integration of different hardware platforms. However, this deployment method relies on a complete virtualization layer, that is, one or more control virtual machines (Controller VM) need to be created on each physical node to manage the computing and storage resources of the node. This means that the virtualization layer not only introduces additional complexity, but also brings certain performance overhead. Specifically, the virtual machine needs to run a complete operating system instance, involving the consumption of resources such as memory and CPU. In addition, the virtualization layer needs to simulate the underlying hardware resources during the resource management process, which will cause certain performance loss. The overhead of the virtualization layer not only reduces the utilization of node resources, but also may affect the response speed and throughput of the overall system, especially under resource-intensive workloads, where performance bottlenecks are more obvious. Therefore, although the software delivery method provides greater flexibility, challenges in resource utilization and performance still exist, especially in scenarios that require high-performance and high-density computing.

[0016] Through the above analysis, the problems and defects of the prior art are as follows:

[0017] (1) The current hyper-converged cloud computing platform has poor compatibility with heterogeneous hardware resources and is difficult to expand elastically.

[0018] (2) The hardware deployment method of traditional hyper-converged platforms is bound to the manufacturer, which limits the hardware selection, and the software deployment method has low resource utilization efficiency. Summary of the invention

[0019] In view of the problems existing in the prior art, the present invention provides a method for constructing a lightweight multi-domain hyper-convergence platform that is compatible with heterogeneous hardware resources.

[0020] The present invention is implemented as follows: a method for constructing a lightweight multi-domain hyper-convergence platform compatible with heterogeneous hardware resources includes:

[0021] Step 1: The lightweight multi-domain hyper-converged platform architecture compatible with heterogeneous hardware resources consists of four modules: computing virtualization, storage virtualization, network virtualization and unified management platform;

[0022] Step 2: Computing virtualization virtualizes the computing resources of physical servers into virtual machines based on KVM technology, supporting multiple workloads in heterogeneous hardware environments;

[0023] Step 3: Storage virtualization integrates the storage resources of different nodes into a global distributed storage pool based on Ceph, supports the integration of heterogeneous storage devices, and provides block storage, object storage, and file storage services;

[0024] Step 4: Network virtualization uses SDN technology to achieve dynamic configuration of virtual networks, network isolation, and network traffic optimization across data centers;

[0025] Step 5: A unified management platform is created based on OpenStack to implement cross-domain resource scheduling, management, and monitoring, and support resource integration of multiple data centers and hybrid cloud architectures;

[0026] Step 6: The container-based lightweight hyper-converged platform deployment method uses containerization technology to abstract the virtualization layer and management services of the multi-domain hyper-converged cloud platform into lightweight containers.

[0027] Furthermore, the computing virtualization virtualizes heterogeneous computing resources into virtual machine services based on KVM technology and a custom XML virtual machine definition file; converts user virtual machine instructions into hardware-level instruction execution by calling the underlying hardware virtualization extension of the processor, and provides a virtual CPU for each virtual machine, and realizes concurrent operation of multiple virtual machines through the KVM kernel's context switching between the physical CPU and the virtual CPU.

[0028] Furthermore, the storage virtualization integrates the storage devices on multiple service clusters into a global storage pool based on Ceph, supports centralized management and allocation of storage resources, and pools the heterogeneous storage resources of multi-domain service clusters into unified SSD and HDD storage resources by building different storage pools and storage layers.

[0029] Furthermore, the network virtualization abstracts physical network resources into virtual networks through software definition, so that multiple virtual machines can share the same underlying network infrastructure. Network virtualization separates virtual networks from physical networks through technologies such as virtual switches, virtual routers, and tunnel protocols, thereby achieving flexible network configuration and management.

[0030] Furthermore, the unified management platform divides the entire cloud platform into multiple domains based on OpenStack, and all domains are managed through a unified control plane; the scheduler is responsible for selecting suitable computing nodes in different domains to start virtual machines; when a user requests to create a virtual machine, the global controller will pass the request to the scheduler, and the scheduler will select a suitable domain to deploy the virtual machine based on factors such as resource status, geographical location, network latency, etc. in different regions.

[0031] Furthermore, the container-based lightweight hyper-converged platform deployment method abstracts the virtualization layer and management services of a lightweight multi-domain hyper-converged cloud platform compatible with heterogeneous hardware resources into lightweight containers through containerization technology, thereby achieving decoupling of the platform from the hardware, reducing the performance overhead of the virtualization layer, and improving the utilization efficiency of computing and storage resources.

[0032] Another object of the present invention is to provide a lightweight multi-domain hyper-convergence platform construction system compatible with heterogeneous hardware resources, including:

[0033] Computing virtualization module, which is used to virtualize the computing resources of physical servers into virtual machines based on KVM technology, and supports various workloads in heterogeneous hardware environments;

[0034] Storage virtualization module, which is used for storage virtualization. It integrates the storage resources of different nodes into a global distributed storage pool based on Ceph, supports the integration of heterogeneous storage devices, and provides block storage, object storage, and file storage services.

[0035] Network virtualization module, which is used for network virtualization and uses SDN technology to achieve dynamic configuration of virtual networks, network isolation, and network traffic optimization across data centers;

[0036] The unified management platform module is used to create a unified management platform based on OpenStack to implement cross-domain resource scheduling, management and monitoring, and support resource integration of multiple data centers and hybrid cloud architectures; the container-based lightweight hyper-converged platform deployment method uses containerization technology to abstract the virtualization layer and management services of the multi-domain hyper-converged cloud platform into lightweight containers.

[0037] Another object of the present invention is to provide a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the method for constructing a lightweight multi-domain hyper-converged platform compatible with heterogeneous hardware resources.

[0038] Another object of the present invention is to provide a computer-readable storage medium storing a computer program, which, when executed by a processor, enables the processor to execute the steps of the method for constructing a lightweight multi-domain hyper-converged platform compatible with heterogeneous hardware resources.

[0039] Another object of the present invention is to provide an information data processing terminal, which is used to implement the lightweight multi-domain hyper-converged platform construction system that is compatible with heterogeneous hardware resources.

[0040] In combination with the above technical solutions and the technical problems solved, the advantages and positive effects of the technical solutions to be protected by the present invention are as follows:

[0041] First, it adopts a lightweight hyper-converged platform architecture that is compatible with heterogeneous hardware resources. Based on KVM, physical resources are converted into virtual resources on virtual machines, shielding the differences in the underlying hardware. Based on Ceph, heterogeneous storage resources are integrated. Through distributed storage technology, storage devices from different manufacturers are integrated into a unified storage pool, which is not restricted by the type of underlying hardware devices. Through SDN technology, OpenStack's network service can decouple the network control plane from the hardware and achieve unified management of different network devices. Through virtualization technology and automated management tools, it provides a simplified elastic expansion solution to easily cope with changes in resource requirements.

[0042] Through containerization technology, the virtualization layer and management services of the multi-domain hyper-converged cloud platform in (1) are abstracted into lightweight containers, realizing the decoupling of the platform from the hardware. Administrators can flexibly select hardware vendors and configurations based on business needs, and are no longer restricted to the hardware integration solutions of specific manufacturers. In addition, container technology is more lightweight than traditional virtual machines, and there is no need to run a complete operating system instance for each physical node. Through the containerized KVM virtualization layer, the performance overhead of the virtualization layer is reduced, and the utilization efficiency of computing and storage resources is improved.

[0043] The existing hyper-converged cloud platform is designed to run in a homogeneous environment, with poor compatibility of heterogeneous hardware resources and difficulty in elastic expansion. In addition, the existing hyper-converged platform deployment methods have defects. In terms of hardware deployment methods, they are bound to manufacturers and hardware selection is limited. In terms of software deployment methods, resource utilization efficiency is low.

[0044] The present invention adopts a lightweight hyper-converged platform architecture that is compatible with heterogeneous hardware resources. Based on virtualization technology that is compatible with heterogeneous resources, heterogeneous computing, storage, and network resources are managed in a unified manner, the differences in underlying hardware are shielded, and expansion is easy, solving the problems of poor compatibility of heterogeneous resources and difficulty in elastic expansion in existing solutions. In addition, through containerization technology, the virtualization layer and management services of the lightweight multi-domain hyper-converged cloud platform that is compatible with heterogeneous hardware resources are abstracted into lightweight containers, realizing the decoupling of the platform and hardware, reducing the performance overhead of the virtualization layer, and solving the problems of the existing hyper-converged platform hardware deployment method being bound to manufacturers, limited hardware selection, and low resource utilization efficiency in the software deployment method.

[0045] Second, the expected benefits and commercial value of the technical solution of the present invention after transformation are: through a lightweight hyper-converged platform architecture compatible with heterogeneous hardware resources, the underlying hardware differences are shielded, so that the platform can run seamlessly in different hardware environments. This will greatly reduce the dependence of enterprises on a single hardware manufacturer, provide a wider range of hardware selection space, and reduce hardware costs and procurement risks. Through the introduction of virtualization technology and containerized management, the platform has a strong elastic expansion capability. The system can dynamically expand computing, storage and network resources according to business needs to ensure that while resource utilization is improved, the business can quickly respond to expansion needs. Elastic expansion reduces resource redundancy in traditional systems and reduces unnecessary hardware investment. The use of containerization technology to abstract the virtualization layer and management services into lightweight containers significantly reduces the performance overhead of the virtualization layer, thereby improving the resource utilization efficiency of the platform. Compared with traditional hyper-converged platforms, containerized management services are more lightweight and more efficient to deploy and manage. More efficient resource utilization can improve user experience and enhance customer satisfaction.

[0046] Third, the technical solution of the present invention proposes a method for constructing a lightweight multi-domain hyper-converged platform that is compatible with heterogeneous hardware resources, which successfully solves the problems of poor compatibility, low resource integration and management efficiency, and poor scalability of multi-domain hyper-converged platforms in heterogeneous hardware environments in the prior art, and has achieved significant technical progress in performance and applicability. When facing different hardware environments, traditional hyper-converged architectures often need to rely on a unified hardware platform or expensive dedicated hardware to achieve resource virtualization and centralized management, which leads to limitations in hardware selection and increased costs. The lightweight hyper-converged solution of the present invention provides the ability to efficiently integrate computing, storage, and network resources in heterogeneous hardware environments by being compatible with multiple technologies such as KVM, Ceph, and SDN, significantly improving the compatibility of the platform.

[0047] In terms of computing virtualization, the present invention virtualizes the computing resources of different physical servers into virtual machines based on KVM virtualization technology, which is not restricted by specific hardware, thereby realizing computing resource sharing and centralized management in heterogeneous hardware environments. This design greatly improves the utilization rate of hardware resources, reduces the dependence on hardware uniformity, solves the shortcomings of traditional platforms in terms of compatibility, and saves hardware investment costs for enterprises. At the same time, KVM technology has the advantage of low overhead, ensures the performance of virtual machines and the stability of the system, and is suitable for diversified business needs.

[0048] In terms of storage integration, the present invention adopts the Ceph distributed storage system, and realizes the integration and unified management of different storage device types by virtualizing the storage resources of different nodes. The high scalability and flexibility of Ceph enable the platform to support multiple service types such as block storage, object storage and file storage, meeting the diversified storage needs of enterprise-level applications. Through the distributed characteristics and high availability mechanism of Ceph, this solution effectively improves the fault tolerance and reliability of the storage system, and solves the technical problem that traditional storage integration solutions are difficult to achieve elastic expansion and reliability assurance.

[0049] The present invention adopts SDN technology in network virtualization to realize dynamic configuration, isolation and optimization of network resources, which is particularly suitable for multi-domain architecture and cross-data center network traffic control requirements. Through SDN, the platform can intelligently schedule and isolate network traffic between different domains, meeting the security and flexibility requirements of different business scenarios. Traditional network virtualization technology is often difficult to cope with complex requirements across data centers, and the SDN solution of the present invention significantly enhances the applicability and performance of network virtualization and improves the overall utilization efficiency of network resources.

[0050] In terms of system management, the present invention realizes the centralized management and scheduling of cross-domain resources based on OpenStack, and has the functions of monitoring, scheduling, and expansion, so that the platform can still provide efficient management capabilities in multiple data centers and hybrid cloud environments. In addition, the platform deployment adopts containerization technology, abstracting virtualization and management services into lightweight container modules, which are lightweight and highly scalable, and support flexible resource configuration and rapid deployment. Compared with the traditional architecture, the present invention significantly reduces the complexity of the system, simplifies the deployment process, and improves the scalability and maintenance convenience of the system.

[0051] In summary, the present invention not only achieves technical breakthroughs in heterogeneous hardware compatibility, resource integration, and management efficiency, but also improves the scalability and deployment convenience of the platform through lightweight and containerized design, providing an efficient and stable solution for the industrial application of multi-domain hyper-converged platforms. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1It is a flow chart of a method for constructing a lightweight multi-domain hyper-converged platform compatible with heterogeneous hardware resources provided by an embodiment of the present invention.

[0053] Figure 2 It is a structural block diagram of a lightweight multi-domain hyper-convergence platform construction system compatible with heterogeneous hardware resources provided by an embodiment of the present invention.

[0054] Figure 3 This is a diagram of the architecture of a multi-domain hyper-converged platform for heterogeneous hardware resources provided by an embodiment of the present invention.

[0055] Figure 4 It is a container architecture diagram of the lightweight deployment method provided by an embodiment of the present invention.

[0056] Figure 5 It is a diagram showing the application function of a virtual cloud host in a prototype system provided by an embodiment of the present invention.

[0057] Figure 6 This is a diagram showing the multi-domain heterogeneous cloud collaboration function of the prototype system provided by an embodiment of the present invention.

[0058] Figure 7 This is a diagram showing unified management of resources of a prototype system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0059] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0060] like Figure 1 As shown, a method for constructing a lightweight multi-domain hyper-converged platform compatible with heterogeneous hardware resources provided by an embodiment of the present invention includes the following steps:

[0061] S101, a lightweight multi-domain hyper-converged platform architecture compatible with heterogeneous hardware resources, consists of four modules: computing virtualization, storage virtualization, network virtualization and unified management platform;

[0062] S102, computing virtualization virtualizes the computing resources of the physical server into virtual machines based on KVM technology, supporting various workloads in heterogeneous hardware environments;

[0063] S103, storage virtualization integrates storage resources of different nodes into a global distributed storage pool based on Ceph, supports the integration of heterogeneous storage devices, and provides block storage, object storage, and file storage services;

[0064] S104, network virtualization uses SDN technology to achieve dynamic configuration of virtual networks, network isolation, and cross-data center network traffic optimization;

[0065] S105, a unified management platform based on OpenStack, implements cross-domain resource scheduling, management and monitoring, and supports resource integration of multiple data centers and hybrid cloud architectures;

[0066] S106, the container-based lightweight hyper-converged platform deployment method abstracts the virtualization layer and management services of the multi-domain hyper-converged cloud platform into lightweight containers through containerization technology.

[0067] The multi-domain hyper-convergence platform construction method of the present invention, centered on the goal of being compatible with heterogeneous hardware resources, adopts four modules of computing virtualization, storage virtualization, network virtualization and unified management platform to achieve resource integration and efficient management. First, in the computing virtualization module, the computing resources of the physical server are virtualized into virtual machines based on KVM technology. KVM virtualization supports a variety of workloads under different hardware architectures, enabling the platform to run stably in a heterogeneous hardware environment, while effectively managing and allocating computing resources, improving hardware compatibility and resource utilization.

[0068] Next, in the storage virtualization module, Ceph is used to integrate the storage resources of each node into a global distributed storage pool. The Ceph storage system can connect to different types of storage devices, including SSDs, HDDs, etc., to achieve unified management of heterogeneous storage devices. In this way, the platform provides block storage, object storage, and file storage services, allowing data to be efficiently switched and shared between different storage types, meeting the data storage requirements of different application scenarios, and achieving high availability and fault tolerance.

[0069] In the network virtualization module, the present invention realizes the dynamic configuration, isolation and traffic optimization of virtual networks through SDN technology. SDN technology can support network configuration across data centers, so that resource intercommunication, traffic regulation and isolation can be achieved between different data centers. In addition, the SDN network can dynamically adjust routing and bandwidth, support load balancing and optimize network traffic, improve the efficiency of network resource utilization and ensure data security isolation, and meet the diverse needs in a multi-domain hyper-converged environment.

[0070] Finally, the unified management platform implements cross-domain resource scheduling, management, and monitoring based on OpenStack. Through containerization technology, the virtualization layer and management services are abstracted into lightweight container deployment. This lightweight design enables the platform to support resource integration of multiple data centers and hybrid cloud architectures while reducing resource overhead. Containerized deployment makes the system easy to expand and migrate, realizes flexible cross-domain resource management and monitoring, and is suitable for complex application environments with heterogeneous hardware, improving the overall performance and applicability of the platform.

[0071] The computing virtualization provided in the embodiment of the present invention virtualizes heterogeneous computing resources into virtual machine services based on KVM technology and a custom XML virtual machine definition file; by calling the underlying hardware virtualization extension of the processor, the user virtual machine instructions are converted into hardware-level instruction execution, and a virtual CPU is provided for each virtual machine. The KVM kernel switches the context of the physical CPU and the virtual CPU to achieve concurrent operation of multiple virtual machines.

[0072] The storage virtualization provided by the embodiment of the present invention integrates the storage devices on multiple service clusters into a global storage pool based on Ceph, supports centralized management and allocation of storage resources, and pools the heterogeneous storage resources of multi-domain service clusters into unified SSD and HDD storage resources by constructing different storage pools and storage layers.

[0073] The network virtualization provided in the embodiment of the present invention abstracts physical network resources into virtual networks in a software-defined manner, so that multiple virtual machines can share the same underlying network infrastructure. Network virtualization separates the virtual network from the physical network through technologies such as virtual switches, virtual routers, and tunnel protocols, thereby achieving flexible network configuration and management.

[0074] The unified management platform provided by the embodiment of the present invention divides the entire cloud platform into multiple domains based on OpenStack, and all domains are managed through a unified control plane; the scheduler is responsible for selecting suitable computing nodes in different domains to start virtual machines; when a user requests to create a virtual machine, the global controller will pass the request to the scheduler, and the scheduler will select a suitable domain to deploy the virtual machine based on factors such as resource status, geographical location, network delay, etc. in different regions.

[0075] The container-based lightweight hyper-converged platform deployment method provided in the embodiment of the present invention uses containerization technology to abstract the virtualization layer and management services of a lightweight multi-domain hyper-converged cloud platform compatible with heterogeneous hardware resources into lightweight containers, thereby achieving decoupling of the platform from the hardware, reducing the performance overhead of the virtualization layer, and improving the utilization efficiency of computing and storage resources.

[0076] like Figure 2 As shown, a lightweight multi-domain hyper-converged platform construction system compatible with heterogeneous hardware resources provided by an embodiment of the present invention includes:

[0077] Computing virtualization module, which is used to virtualize the computing resources of physical servers into virtual machines based on KVM technology, and supports various workloads in heterogeneous hardware environments;

[0078] Storage virtualization module, which is used for storage virtualization. It integrates the storage resources of different nodes into a global distributed storage pool based on Ceph, supports the integration of heterogeneous storage devices, and provides block storage, object storage, and file storage services.

[0079] Network virtualization module, which is used for network virtualization and uses SDN technology to achieve dynamic configuration of virtual networks, network isolation, and network traffic optimization across data centers;

[0080] The unified management platform module is used to create a unified management platform based on OpenStack to implement cross-domain resource scheduling, management and monitoring, and support resource integration of multiple data centers and hybrid cloud architectures; the container-based lightweight hyper-converged platform deployment method uses containerization technology to abstract the virtualization layer and management services of the multi-domain hyper-converged cloud platform into lightweight containers.

[0081] Another object of the present invention is to provide a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the method for constructing a lightweight multi-domain hyper-converged platform compatible with heterogeneous hardware resources.

[0082] Another object of the present invention is to provide a computer-readable storage medium storing a computer program, which, when executed by a processor, enables the processor to execute the steps of the method for constructing a lightweight multi-domain hyper-converged platform compatible with heterogeneous hardware resources.

[0083] Another object of the present invention is to provide an information data processing terminal, which is used to implement the lightweight multi-domain hyper-converged platform construction system that is compatible with heterogeneous hardware resources.

[0084] The present invention is specifically implemented:

[0085] The method for constructing a lightweight multi-domain hyper-converged platform for heterogeneous hardware resources provided by an embodiment of the present invention comprises the following steps:

[0086] (1) Lightweight multi-domain hyper-converged platform architecture compatible with heterogeneous hardware resources

[0087] To solve the problem that the existing hyper-convergence platform cannot effectively be compatible with heterogeneous hardware resources and has difficulty in elastic expansion, we build Figure 2The system architecture of the lightweight multi-domain hyper-converged platform compatible with heterogeneous hardware resources shown in the figure includes computing virtualization, storage virtualization, and network virtualization. Computing virtualization virtualizes the computing resources of physical servers into virtual machines based on KVM (Kernel-based Virtual Machine) technology, supporting a variety of workloads in heterogeneous hardware environments. Storage virtualization integrates the storage resources of different nodes into a global distributed storage pool based on Ceph, supports the integration of heterogeneous storage devices (such as HDD, SSD, NVMe), and provides block storage, object storage, and file storage services. Network virtualization uses SDN technology to achieve dynamic configuration of virtual networks, network isolation, and cross-data center network traffic optimization. Finally, a unified management platform is created based on OpenStack to achieve cross-domain resource scheduling, management, and monitoring, and support resource integration of multiple data centers and hybrid cloud architectures.

[0088] Computing resource virtualization is based on KVM to achieve compatibility with heterogeneous computing resources. KVM is a virtual machine monitor built on hardware-assisted virtualization technology. KVM's lowest hardware dependence is the CPU's hardware virtualization support, such as Intel's VT technology and AMD's AMD-V technology, so it can support almost all operating systems that can run on the corresponding hardware, such as Linux, Windows, FreeBSD, MacOS, etc. When a virtual machine creation request is issued, the hardware resource configuration details of the client virtual machine are defined based on the XML configuration file, and the QEMU user-mode process is called to simulate the virtual machine's hardware devices. The QEMU process passes the virtual machine's CPU instructions to KVM, and the KVM kernel module directly executes these instructions.

[0089] Storage virtualization is implemented based on the Ceph distributed storage system, which integrates storage devices on multiple service clusters into a global storage pool, supports centralized management and allocation of storage resources, and pools heterogeneous storage resources of multi-domain service clusters into unified SSD and HDD storage resources by building different storage pools and storage layers, so that user data can be allocated to storage devices with different performance and capacity according to read and write requirements, such as storing data with high access frequency in SSD or NVMe devices, while HDD stores infrequently accessed data. By integrating storage devices, creating storage pools, and enabling the RBD function, you can define the virtual disk of the virtual machine, and specify the created disk image as the disk of the virtual machine in the virtual machine xml definition file, so that the customer virtual machine can ignore hardware differences and use standardized storage resources.

[0090] Network virtualization abstracts physical network resources into virtual networks through software definition, allowing multiple virtual machines to share the same underlying network infrastructure. Network virtualization separates virtual networks from physical networks through technologies such as virtual switches, virtual routers, and tunnel protocols, thereby achieving flexible network configuration and management. The steps to build multiple shared virtual networks are as follows:

[0091] First, deploy a virtual switch on each physical node. The virtual switch connects the virtual network card of the virtual machine to the network and provides virtual network functions. Then configure the virtual network in the virtual switch so that virtual machines or containers in the same domain can communicate through the virtual network. Secondly, configure the tunnel protocol between multiple domains to establish cross-domain connections between virtual switches. The tunnel protocol encapsulates the network traffic of the virtual machine into tunnel traffic and transmits it from one domain to another to achieve cross-domain communication of the virtual network. Then deploy the SDN controller to manage all virtual networks in the multi-domain cluster. The SDN controller communicates with the virtual switch in each domain through the southbound interface and dynamically configures the network topology and traffic policy. The SDN controller can perform dynamic network scheduling based on the global network status to ensure the optimal path selection for cross-domain traffic. Finally, enable distributed routing in a multi-domain environment so that virtual machines or containers can communicate directly across domains. Through the distributed router, the network traffic of the virtual machine can be directly transmitted to the target domain through the tunnel, reducing the number of network hops and latency.

[0092] The unified management platform divides the entire cloud platform into multiple domains based on OpenStack. Each domain is an independent computing and storage resource area. All domains are managed through a unified control plane. The scheduler is responsible for selecting suitable computing nodes in different domains to start virtual machines. When a user requests to create a virtual machine, the global controller will pass the request to the scheduler, and the scheduler will select the appropriate domain to deploy the virtual machine based on factors such as resource status, geographical location, and network latency in different regions.

[0093] (2) Container-based lightweight hyper-converged platform deployment method

[0094] Through containerization technology, the virtualization layer and management services of the multi-domain hyper-converged cloud platform in (1) are abstracted into lightweight containers, which reduces resource overhead, improves resource utilization, and enhances the flexibility, scalability, and operation and maintenance efficiency of the system. The specific container structure is as follows: Figure 4As shown in the figure. First, containers are used to run KVM services, making virtual machine management more lightweight and flexible. Containerized KVM instances reduce the resource usage of the traditional virtualization layer because there is no need to create an independent control virtual machine (Control VM) for each node. Secondly, containerized Ceph is used to implement distributed storage services, supporting elastic expansion and automated management. The various components of Ceph are containerized and deployed and managed through container orchestration tools. One or more Ceph containers can run on each physical node to manage local storage devices. Then the SDN controller and virtual switch are deployed in containers to ensure that the network virtualization function is independent of the hardware and easy to expand. Containerization tools are used again to deploy and manage OpenStack components. Each OpenStack service runs in an independent container, providing management interfaces for computing, storage, and network resources. Finally, Kubernetes nodes are deployed on each physical node to manage containerized virtualization, storage, network, and management services. Kubernetes is responsible for scheduling these containers to ensure load balancing and high availability.

[0095] Through containerization technology and Kubernetes' scheduling capabilities, the platform and hardware are decoupled. Administrators can flexibly choose hardware vendors and configurations based on business needs, and are no longer restricted by hardware integration solutions from specific vendors. Kubernetes supports dynamic expansion and load balancing, and can automatically expand nodes and container instances based on resource requirements, solving the problem of limited hardware expansion on traditional hyper-converged platforms. In addition, container technology is more lightweight than traditional virtual machines, and there is no need to run a complete operating system instance for each physical node. Through the containerized KVM virtualization layer, the performance overhead of the virtualization layer is reduced, and the utilization efficiency of computing and storage resources is improved.

[0096] Relevant evidence of the technical effects achieved by the embodiments of the present invention.

[0097] Figure 5 The virtual cloud host application module provided by the prototype system of the present invention sends the user request to the virtualization manager of the present invention, and the virtualization manager creates virtual machine resources that meet the requirements according to the user request.

[0098] Figure 6 The multi-domain heterogeneous cloud collaboration module provided by the prototype system of the present invention proposes and implements a multi-cloud collaborative cloud computing resource management system architecture, realizes cross-domain collaborative management of computing and storage resources, data sharing, supports cross-domain collaborative computing, meets the basic needs of distributed application systems, and improves real-time and scalability.

[0099] Figure 7The unified resource management module provided by the prototype system of the present invention is a unified management platform based on dividing the entire cloud platform into multiple domains. Each domain is an independent computing and storage resource area, and all domains are managed through a unified control plane.

[0100] It should be noted that the embodiments of the present invention can be implemented by hardware, software, or a combination of software and hardware. The hardware part can be implemented using dedicated logic; the software part can be stored in a memory and executed by an appropriate instruction execution system, such as a microprocessor or dedicated design hardware. It can be understood by a person of ordinary skill in the art that the above-mentioned devices and methods can be implemented using computer executable instructions and / or contained in a processor control code, such as a carrier medium such as a disk, CD or DVD-ROM, a programmable memory such as a read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. Such code is provided on the carrier medium. The device and its modules of the present invention can be implemented by hardware circuits such as very large-scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, etc., or programmable hardware devices such as field programmable gate arrays, programmable logic devices, etc., can also be implemented by software executed by various types of processors, and can also be implemented by a combination of the above-mentioned hardware circuits and software, such as firmware.

[0101] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any modifications, equivalent substitutions and improvements made by any technician familiar with the technical field within the technical scope disclosed by the present invention and within the spirit and principle of the present invention should be covered by the protection scope of the present invention.

Claims

1. A method for constructing a lightweight multi-domain hyper-converged platform compatible with heterogeneous hardware resources, characterized in that: The following steps are involved: Step 1: The lightweight multi-domain hyper-converged platform architecture compatible with heterogeneous hardware resources consists of four modules: computing virtualization, storage virtualization, network virtualization and unified management platform; Step 2: Computing virtualization virtualizes the computing resources of physical servers into virtual machines based on KVM technology, supporting multiple workloads in heterogeneous hardware environments; Step 3: Storage virtualization integrates the storage resources of different nodes into a global distributed storage pool based on Ceph, supports the integration of heterogeneous storage devices, and provides block storage, object storage, and file storage services; Step 4: Network virtualization uses SDN technology to achieve dynamic configuration of virtual networks, network isolation, and cross-data center network traffic optimization; Step 5: A unified management platform is created based on OpenStack to implement cross-domain resource scheduling, management, and monitoring, and support resource integration of multiple data centers and hybrid cloud architectures; Step 6: The container-based lightweight hyper-converged platform deployment method uses containerization technology to abstract the virtualization layer and management services of the multi-domain hyper-converged cloud platform into lightweight containers.

2. The method for constructing a lightweight multi-domain hyper-converged platform compatible with heterogeneous hardware resources as claimed in claim 1, characterized in that: The computing virtualization virtualizes heterogeneous computing resources into virtual machine services based on KVM technology and a custom XML virtual machine definition file; converts user virtual machine instructions into hardware-level instruction execution by calling the underlying hardware virtualization extension of the processor, and provides a virtual CPU for each virtual machine, and realizes the concurrent operation of multiple virtual machines through the KVM kernel's context switching between the physical CPU and the virtual CPU.

3. The method for constructing a lightweight multi-domain hyper-converged platform compatible with heterogeneous hardware resources as claimed in claim 1, characterized in that: The storage virtualization integrates the storage devices on multiple service clusters into a global storage pool based on Ceph, supports centralized management and allocation of storage resources, and pools the heterogeneous storage resources of multi-domain service clusters into unified SSD and HDD storage resources by building different storage pools and storage layers.

4. The method for constructing a lightweight multi-domain hyper-converged platform compatible with heterogeneous hardware resources as claimed in claim 1, characterized in that: The network virtualization abstracts physical network resources into virtual networks in a software-defined manner, allowing multiple virtual machines to share the same underlying network infrastructure. Network virtualization separates virtual networks from physical networks through technologies such as virtual switches, virtual routers, and tunnel protocols, thereby achieving flexible network configuration and management.

5. The method for constructing a lightweight multi-domain hyper-converged platform compatible with heterogeneous hardware resources as claimed in claim 1, characterized in that: The unified management platform divides the entire cloud platform into multiple domains based on OpenStack, and all domains are managed through a unified control plane; the scheduler is responsible for selecting suitable computing nodes in different domains to start virtual machines; when a user requests to create a virtual machine, the global controller will pass the request to the scheduler, and the scheduler will select a suitable domain to deploy the virtual machine based on factors such as resource status, geographical location, network latency, etc. in different regions.

6. The method for constructing a lightweight multi-domain hyper-converged platform compatible with heterogeneous hardware resources as claimed in claim 1, characterized in that: The container-based lightweight hyper-converged platform deployment method uses containerization technology to abstract the virtualization layer and management services of a lightweight multi-domain hyper-converged cloud platform that is compatible with heterogeneous hardware resources into lightweight containers, thereby achieving decoupling of the platform from the hardware, reducing the performance overhead of the virtualization layer, and improving the utilization efficiency of computing and storage resources.

7. A lightweight multi-domain hyper-converged platform construction system compatible with heterogeneous hardware resources that implements the lightweight multi-domain hyper-converged platform construction method compatible with heterogeneous hardware resources as described in any one of claims 1 to 6, characterized in that: The lightweight multi-domain hyper-converged platform construction system compatible with heterogeneous hardware resources includes: Computing virtualization module, which is used to virtualize the computing resources of physical servers into virtual machines based on KVM technology, and supports various workloads in heterogeneous hardware environments; Storage virtualization module, which is used for storage virtualization. It integrates the storage resources of different nodes into a global distributed storage pool based on Ceph, supports the integration of heterogeneous storage devices, and provides block storage, object storage, and file storage services. Network virtualization module, which is used for network virtualization and uses SDN technology to achieve dynamic configuration of virtual networks, network isolation, and network traffic optimization across data centers; The unified management platform module is used to create a unified management platform based on OpenStack to implement cross-domain resource scheduling, management and monitoring, and support resource integration of multiple data centers and hybrid cloud architectures; the container-based lightweight hyper-converged platform deployment method uses containerization technology to abstract the virtualization layer and management services of the multi-domain hyper-converged cloud platform into lightweight containers.

8. A computer device, characterized in that: The computer device includes a memory and a processor, the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the method for constructing a lightweight multi-domain hyper-converged platform compatible with heterogeneous hardware resources as described in any one of claims 1-6.

9. A computer-readable storage medium storing a computer program, wherein when the computer program is executed by a processor, the processor executes the steps of the method for constructing a lightweight multi-domain hyper-converged platform compatible with heterogeneous hardware resources as described in any one of claims 1 to 6.

10. An information data processing terminal, characterized in that: The information data processing terminal is used to implement the lightweight multi-domain hyper-converged platform construction system compatible with heterogeneous hardware resources as described in claim 7.

Images