Desktop system virtualization implementation system and method based on container technology

Abstract

The invention discloses a desktop system virtualization implementation system based on a container technology. The desktop system virtualization implementation system comprises a system facility layer, a containerization layer, a virtualization layer and a monitoring alarm layer which are interacted in sequence, wherein the system facility layer is a Linux operating system and is used for providing underlying calculation and storage resources; the containerization layer comprises a Docker, Kubernetes and a Harbor, the Docker is used for realizing lightweight resource isolation of a desktop system, the Kubernetes are used for automatic operation and maintenance and resource allocation of containerization desktop instances, and the Harbor is used as a private mirror image warehouse for storing container mirror images; the virtualization layer comprises a containerized desktop instance and a noVNC, the containerized desktop instance is a desktop operating system instance constructed based on a Docker mirror image, and the noVNC is used for realizing remote access of a browser side to the containerized desktop instance through HTML5 and WebSocket technologies; the monitoring alarm layer comprises an Export, a Prometheus Server, a Grafana and an Alertmanager, the Export collects system resource monitoring data, the Prometheus Server pulls and stores the monitoring data in a timed manner, the Grafana visually displays the monitoring data, and the Alertmanager receives alarm information and informs operation and maintenance personnel of the alarm information.

Description

TECHNICAL FIELD

[0001] The application relates to a container technology-based desktop system virtualization implementation system and method, and belongs to the technical field of computer networks. BACKGROUND

[0002] In recent years, with the rapid development of cloud computing, desktop operating system virtualization is considered to be a good method capable of realizing centralized management of user desktops, and the management cost is relatively less than that of traditional methods. For example, VMware View of VMware, XenDesktop of Citrix, and Huawei Cloud Desktop of Huawei are considered to be mainstream cloud computing desktop software products.

[0003] Citrix, an American company, is a leading company of cloud computing and virtualization technology. The solution provided by the company is mainly XenDesktop. The solution can deliver a Windows desktop system as a kind of on-demand service to users, and can quickly and safely deliver a single application or an entire desktop to all users in an enterprise, greatly simplifying desktop management. However, the solution has a high failure rate in use and is difficult to maintain, and the technical architecture is difficult to adapt to the trend of future cloud computing technology development and market demand, and therefore has been faced with elimination. A product coeval with XenDesktop of Citrix is VMware View developed by VMware. View uses PCoIP protocol for remote connection, and although the network delay between end-to-end is effectively reduced, there are a large number of bandwidth occupation problems.

[0004] The cloud desktop product of a domestic main company, such as FusionSphere of Huawei, combines a virtualization cloud platform and cloud management. FusionSphere mainly includes two core components of Compute and Manager. The Compute is responsible for virtualization and centralized management of CPU, memory and other basic implementations. The Manager is responsible for the entire system, and mainly manages servers, switches and other hardware facilities. However, FusionSphere is a product dominated by Huawei, and there are compatibility problems for non-Huawei hardware.

[0005] Through the analysis of the research status at home and abroad above, it can be known that the current most desktop system virtualization solutions generally have problems of high failure rate, heavy maintenance burden, large bandwidth occupation and hardware compatibility. SUMMARY

[0006] The application solves the technical problems of the prior art, and provides a container technology-based desktop system virtualization implementation system and method.

[0007] The technical scheme for solving the above technical problems of the present application is: a desktop system virtualization implementation system based on container technology, comprising a system facility layer, a containerization layer, a virtualization layer and a monitoring and alarm layer that interact in sequence; wherein: The system facility layer is a Linux operating system, used to provide underlying computing and storage resources; The containerization layer comprises Docker, Kubernetes and Harbor, wherein: the Docker is used to implement lightweight resource isolation of the desktop system, the Kubernetes is used to implement automatic operation and maintenance and resource allocation of the containerized desktop instance, and the Harbor is used as a private image repository to store container images; The virtualization layer comprises a containerized desktop instance and a noVNC, wherein: the containerized desktop instance is a desktop operating system instance built based on a Docker image, and the noVNC is used to realize remote access of the containerized desktop instance by the browser end through HTML5 and WebSocket technologies; The monitoring and alarm layer comprises an Exporter, a Prometheus Server, a Grafana and an Alertmanager, wherein: the Exporter collects system resource monitoring data, the Prometheus Server periodically pulls and stores the monitoring data, the Grafana visualizes the monitoring data, and the Alertmanager receives alarm information and notifies the operation and maintenance personnel.

[0008] Preferably, the containerized desktop instance is built through a Dockerfile, specifically based on an image of an Ubuntu-desktop system, and the image is pre-uploaded to the Harbor repository to realize cross-platform reuse.

[0009] Preferably, the implementation of the remote access comprises a service layer, an access layer and a user end, wherein: The service layer comprises a Kubernetes cluster, a containerized desktop instance, an authentication service and a VNC protocol, the authentication service verifies the user identity through an authentication token, and the VNC protocol transmits the desktop picture and input / output control; The access layer converts the traffic through a WebSocket proxy and forwards it to the noVNC; The user end realizes the access and control of the remote desktop through the browser based on HTML5 Canvas and WebSocket.

[0010] Preferably, in the monitoring alarm layer, the Exporter supports multiple types of monitoring objects, including container resources, server hardware and network status; Grafana generates dynamic monitoring dashboards by connecting the Prometheus data source and triggers alarm rules.

[0011] The application also provides a container technology-based desktop system virtualization implementation method, specifically comprising the following steps: (1) Building a container platform: install a Linux system in a server, deploy Docker, Kubernetes and Harbor after configuring the system environment, and form a containerization layer; (2) Packaging a container image: build a container image containing a desktop operating system through Dockerfile, upload it to the Harbor repository, and run the containerized desktop instance through Docker and Kubernetes; (3) Deploying a monitoring alarm system: build Docker images of Grafana, Alertmanager, Exporter and PrometheusServer, deploy them through docker-compose, and realize real-time monitoring and alarm of system resources; (4) Realizing remote access: after authentication, users access the containerized desktop instance through noVNC, WebSocket proxy and VNC protocol through a browser.

[0012] Preferably, in step (1), the Linux system uses CentOS-7-x86_64 version, and the system environment configuration includes closing the firewall, writing the domain name resolution and permanently disabling the SWAP swap partition.

[0013] Preferably, in step (2), the container image is an Ubuntu-desktop image, which is built through the docker build command and uploaded to the Harbor repository through the docker load command.

[0014] Preferably, in step (3), the deployment of the monitoring alarm system includes: building the images of each component through Dockerfile, writing the docker-compose file to define the network and dependency relationship between the components, and executing the docker-compose up -d command to complete the automated deployment.

[0015] The application has the following advantages: The application adopts container technology to replace traditional virtualization technology, realizes light weight resource isolation and automatic operation and maintenance, compared with traditional virtual machine, the containerization deployment is faster in starting speed, and by utilizing the function of Kubernetes in automatically managing containers, the failure rate can be effectively reduced, the burden of operation and maintenance personnel is reduced, by utilizing Docker and Kubernetes cluster to deploy containers, there is no hardware compatibility problem, the starting time is short, the response speed is fast, and the virtual desktop can be quickly deployed; (2) The application integrates Prometheus+Grafana to build full-link monitoring, and the system resources are monitored in real time, so that the system resource bottleneck and system program failure are prevented, and the failure is handled in time; (3) The application realizes quick browser remote desktop by noVNC protocol, and the design is light, and the bandwidth occupation is relatively low; (4) The remote desktop is convenient and fast to use on various devices in the network environment. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 It is a system architecture schematic diagram of the application; Figure 2 It is a remote login container desktop schematic diagram; Figure 3 It is a web login remote desktop principle diagram; Figure 4 It is a resource monitoring alarm layer architecture diagram; Figure 5 It is a Kubernetes and docker installation completion diagram; Figure 6 It is a diagram of writing Dockerfile and building; Figure 7 It is a diagram of writing and building and then uploading image warehouse; Figure 8 It is a diagram of building exporter image; Figure 9 It is a diagram of building alert image; Figure 10 It is a diagram of building grafana image; Figure 11 It is a diagram of building Prometheus image; Figure 12 It is a diagram of deployment monitoring; Figure 13 It is a diagram of visual configuration interface of monitoring alarm system; Figure 14 It is a diagram of web login remote desktop. DETAILED DESCRIPTION EMBODIMENT

[0017] This embodiment provides a desktop system virtualization implementation system based on container technology, such as... Figure 1 As shown, it includes the system infrastructure layer, containerization layer, virtualization layer, and monitoring and alarm layer that interact sequentially; wherein: The system infrastructure layer is a Linux operating system, which provides underlying computing and storage resources; The containerization layer includes Docker, Kubernetes, and Harbor, where: Docker is used to implement lightweight resource isolation for desktop systems, Kubernetes is used for automated operation and resource allocation of containerized desktop instances, and Harbor serves as a private image repository to store container images; The virtualization layer includes containerized desktop instances and noVNC, where: containerized desktop instances are desktop operating system instances built based on Docker images, and noVNC is used to enable remote access to containerized desktop instances from the browser through HTML5 and WebSocket technologies; The monitoring and alerting layer includes Exporter, Prometheus Server, Grafana, and Alertmanager. Specifically, Exporter collects system resource monitoring data, Prometheus Server periodically pulls and stores the monitoring data, Grafana visualizes the monitoring data, and Alertmanager receives alert information and notifies operations and maintenance personnel.

[0018] The containerized desktop instance described above is built using a Dockerfile, specifically an image based on the Ubuntu-desktop system, and the image is pre-uploaded to the Harbor repository for cross-platform reuse.

[0019] The implementation of the aforementioned remote access includes a service layer, an access layer, and a user client, wherein: The service layer includes a Kubernetes cluster, containerized desktop instances, authentication services, and the VNC protocol. The authentication service verifies the user's identity through authentication tokens, and the VNC protocol transmits the desktop screen and input / output controls. The access layer transforms traffic through a WebSocket proxy and forwards it to noVNC; Users can access and control remote desktops through a browser using HTML5 Canvas and WebSocket.

[0020] In the above monitoring alarm layer, the exporter supports multiple types of monitoring objects, including container resources, server hardware and network status; Grafana generates dynamic monitoring dashboards by connecting the Prometheus data source and triggers alarm rules.

[0021] The application also provides a container technology-based desktop system virtualization implementation method, specifically comprising the following steps: (1) Build a container platform: install a Linux system in the server, deploy Docker, Kubernetes and Harbor after configuring the system environment, and form a containerization layer; (2) Package container images: build container images containing desktop operating systems through Dockerfile, upload them to the Harbor repository, and run containerized desktop instances through Docker and Kubernetes; (3) Deploy a monitoring and alarm system: build Docker images of Grafana, Alertmanager, Exporter and PrometheusServer, deploy them through docker-compose, and realize real-time monitoring and alarm of system resources; (4) Implement remote access: after authentication, users access containerized desktop instances through noVNC, WebSocket proxy and VNC protocol through a browser.

[0022] In step (1) above, the Linux system uses CentOS-7-x86_64 version, and the system environment configuration includes closing the firewall, writing domain name resolution and permanently disabling the SWAP swap partition.

[0023] In step (2) above, the container image is an Ubuntu-desktop image, which is built by the docker build command and uploaded to the Harbor repository by the docker load command.

[0024] In step (3) above, the deployment of the monitoring and alarm system includes: building each component image through Dockerfile, writing docker-compose file to define the network and dependency relationship between components, and executing docker-compose up-d command to complete automatic deployment.

[0025] In addition to the above embodiments, the application can also have other implementation manners. Any technical solution formed by equivalent replacement or equivalent transformation falls within the protection scope of the application.

Claims

1. A desktop system virtualization implementation system based on container technology, characterized in that: This includes the sequentially interacting system infrastructure layer, containerization layer, virtualization layer, and monitoring and alarm layer; among which: The system infrastructure layer is a Linux operating system, which provides underlying computing and storage resources; The containerization layer includes Docker, Kubernetes, and Harbor, where: Docker is used to implement lightweight resource isolation for desktop systems, Kubernetes is used for automated operation and maintenance and resource allocation of containerized desktop instances, and Harbor serves as a private image repository to store container images; The virtualization layer includes containerized desktop instances and noVNC, wherein: the containerized desktop instance is a desktop operating system instance built based on a Docker image, and noVNC is used to enable remote access to the containerized desktop instance by the browser through HTML5 and WebSocket technologies; The monitoring and alarm layer includes Exporter, Prometheus Server, Grafana, and Alertmanager. Specifically, Exporter collects system resource monitoring data, Prometheus Server periodically pulls and stores the monitoring data, Grafana visualizes the monitoring data, and Alertmanager receives alarm information and notifies operations and maintenance personnel.

2. The desktop system virtualization implementation system based on container technology according to claim 1, characterized in that: The containerized desktop instance is built using a Dockerfile, specifically an image based on the Ubuntu-desktop system, and the image is pre-uploaded to the Harbor repository for cross-platform reuse.

3. The desktop system virtualization implementation system based on container technology according to claim 1, characterized in that: The implementation of remote access includes a service layer, an access layer, and a user terminal, wherein: The service layer includes a Kubernetes cluster, containerized desktop instances, authentication services, and the VNC protocol. The authentication service verifies the user's identity through authentication tokens, and the VNC protocol transmits the desktop screen and input / output controls. The access layer transforms traffic through a WebSocket proxy and forwards it to noVNC; Users can access and control remote desktops through a browser using HTML5 Canvas and WebSocket.

4. The desktop system virtualization implementation system based on container technology according to claim 1, characterized in that: In the monitoring and alerting layer, Exporter supports various types of monitoring objects, including container resources, server hardware, and network status; Grafana generates dynamic monitoring dashboards by connecting to Prometheus data sources and triggers alert rules.

5. A method for implementing desktop system virtualization based on container technology according to any one of claims 1-4, characterized in that: Specifically, the following steps are included: (1) Build a container platform: Install a Linux system on the server, configure the system environment, and then deploy Docker, Kubernetes and Harbor to form a containerization layer; (2) Package the container image: Build a container image containing a desktop operating system using Dockerfile, upload it to the Harbor repository, and run the containerized desktop instance using Docker and Kubernetes; (3) Deploy a monitoring and alarm system: Build Docker images of Grafana, Alertmanager, Exporter and PrometheusServer, and deploy them through docker-compose to achieve real-time monitoring and alarm of system resources; (4) Enable remote access: After being authenticated through a browser, users can access the containerized desktop instance through noVNC, WebSocket proxy and VNC protocol.

6. The method for implementing desktop system virtualization based on container technology according to claim 5, characterized in that, In step (1), the Linux system selected is CentOS-7-x86_64, and the system environment configuration includes disabling the firewall, writing domain name resolution, and permanently disabling the SWAP swap partition.

7. The method for implementing desktop system virtualization based on container technology according to claim 5, characterized in that, In step (2), the container image is an Ubuntu-desktop image, which is built using the docker build command and uploaded to the Harbor repository using the docker load command.

8. The method for implementing desktop system virtualization based on container technology according to claim 5, characterized in that, In step (3), the deployment of the monitoring and alarm system includes: building images of each component through Dockerfile, writing docker-compose files to define the network and dependency relationships between components, and executing the docker-compose up -d command to complete the automated deployment.

Images