Google Distributed Cloud 是我们推出的一款解决方案，可将 Google Cloud基础设施和服务扩展到您的数据中心。我们提供的 Google Distributed Cloud 有两种配置：连接配置和经过网闸隔离配置，这两种配置均可在 Google 提供的硬件上运行。我们还以纯软件产品的形式提供可在您自己的硬件上运行的 Google Distributed Cloud。Google Distributed Cloud 软件可安装在 VMware 或裸金属上。本指南适用于在 VMware vSphere 环境中的您自己的硬件上运行的 Google Distributed Cloud 软件。
Google Distributed Cloud 基于 Google Kubernetes Engine (GKE)，具有自己的 Kubernetes 软件包，可扩展 GKE 以在本地环境中使用。借助 Google Distributed Cloud，您可以在本地创建、管理和升级 GKE 集群，同时还能使用Google Cloud 功能，并使用 Google 的基础设施在集群上大规模部署和运行容器化应用。
将 VERSION 替换为 1.0 到 1.16 之间的次要版本，并将该网址复制到浏览器的地址栏中。
工作原理
Google Distributed Cloud 扩展了 GKE，让您可以在自己的场所内在 vSphere 环境中创建 GKE 集群，并在 Google Cloud 中将它们与常规 GKE 集群以及其他环境中的集群一起作为舰队的一部分进行管理。
由于 Google Distributed Cloud 软件在您的数据中心内运行，而不是在Google Cloud上运行，因此除了 GKE 软件本身之外，您还需要安装一些管理员和控制平面软件。系统会在安装和升级过程中下载在数据中心内运行的软件。
下图显示了已完成安装的简化结果。
具有一个用户集群的 Google Distributed Cloud 架构
关键组件
Google Distributed Cloud 的 VMware 纯软件安装包括以下组件：
用户集群是实现应用的工作负载运行的地方，例如 GKE on Google Cloud。运行工作负载的节点称为工作器节点。用户集群还具有一个或多个控制平面节点。在上图中，用户集群有一个控制平面节点。
管理员集群可管理一个或多个用户集群。在上图中，管理员集群有三个控制平面节点。
管理员工作站是一个单独的机器，其中包含集群创建者和开发者管理其安装所需的工具：
通过从管理员工作站运行 gkectl，您可以创建和更新集群并执行一些其他管理任务
通过从管理员工作站运行 kubectl，您可以与管理员集群和用户集群交互，包括部署和管理工作负载
Google Cloud 控制台为您的 Google Cloud 项目（包括 VMware 上的集群）提供网页界面。作为在管理员工作站上运行命令的替代方法，您可以从 Google Cloud 控制台执行部分管理任务，包括集群创建。
集群管理员和开发者使用 kubectl 和虚拟 IP 地址 (VIP) 访问管理员集群和用户集群中的控制平面。您可以在创建集群期间配置 VIP。在用户集群中调用工作负载的用户和开发者使用 Service 和 Ingress VIP 地址。安装中的每个节点也有自己的 IP 地址。您可以参阅规划您的 IP 地址，详细了解 Google Distributed Cloud 的 IP 规划。
连接到舰队
所有 Google Distributed Cloud 集群都是舰队（Kubernetes 集群的逻辑分组）的成员。借助舰队，您的组织可以加强对各个集群以及整个集群组的管理，您的团队也可以采用 Google 所用的类似最佳实践。您可以在 Google Cloud 控制台中同时查看和管理舰队集群，并使用启用舰队的 GKE Enterprise 功能来帮助您大规模管理、控制和操作工作负载。您可以在 GKE Enterprise 部署选项中查看本地环境的可用舰队功能的完整列表。
每个舰队集群与 Google Cloud 的连接都由 Connect Agent 管理，Connect Agent 会在 Google Distributed Cloud 安装过程中部署。您可以查看 Connect Agent 概览，详细了解此代理的工作原理。
[[["易于理解","easyToUnderstand","thumb-up"],["解决了我的问题","solvedMyProblem","thumb-up"],["其他","otherUp","thumb-up"]],[["很难理解","hardToUnderstand","thumb-down"],["信息或示例代码不正确","incorrectInformationOrSampleCode","thumb-down"],["没有我需要的信息/示例","missingTheInformationSamplesINeed","thumb-down"],["翻译问题","translationIssue","thumb-down"],["其他","otherDown","thumb-down"]],["最后更新时间 (UTC)：2025-07-31。"],[],[],null,["Google Distributed Cloud is our solution that extends Google Cloud's infrastructure and\nservices into your data center. We offer Google Distributed Cloud in both connected and\nair-gapped configurations that run on Google-provided hardware. We also offer\nGoogle Distributed Cloud as a software-only product that runs on your own hardware.\nGoogle Distributed Cloud software can be installed on either VMware or bare metal. This\nguide is for Google Distributed Cloud software that runs on your own hardware in a\nVMware vSphere environment.\n\nGoogle Distributed Cloud is based on\n[Google Kubernetes Engine (GKE)](/kubernetes-engine/docs/concepts/kubernetes-engine-overview),\nwith its own Kubernetes package that extends GKE for use in an\non-premises environment. With Google Distributed Cloud you can create, manage, and\nupgrade GKE clusters on your own premises while using\nGoogle Cloud features, and deploy and operate containerized applications\non your clusters at scale using Google's infrastructure.\n\nInstalling Google Distributed Cloud software entitles you to use\nGKE Enterprise: an enterprise tier for GKE with\npowerful features for governing, managing, and operating containerized workloads\nat scale. You can find out more about GKE Enterprise and the\nfeatures available on VMware in the\n[GKE Enterprise (Anthos) technical overview](/kubernetes-engine/enterprise/docs/concepts/overview)\n\nThis page provides an overview of how Google Distributed Cloud works on VMware, giving\nyou the background you need before going on to a minimal or production\ninstallation.\n\nSupported versions\n\nThis documentation covers all supported versions of Google Distributed Cloud.\nWhere relevant, we also retain limited information for older, unsupported\nversions in this documentation. Version-specific differences in requirements and\nbehavior are noted in the documentation. Similarly, when a new feature becomes\navailable, the supported version for the feature is documented.\n\nFor a list of the supported minor versions and available patches, see\n[Versioning](/kubernetes-engine/distributed-cloud/vmware/docs/version-history).\n\nStarting with the 1.29 release, we no longer create a directory for the previous\nminor release. Differences in behavior and are noted in the documentation.\nSimilarly, when a new feature becomes available, the supported version for\nthe feature is documented.\n\nYou can find the complete documentation for an earlier version by adding the\nminor version number in the following URL: \n\n https://cloud.google.com/anthos/clusters/docs/on-prem/\u003cvar translate=\"no\"\u003eVERSION\u003c/var\u003e/\n\nReplace \u003cvar translate=\"no\"\u003eVERSION\u003c/var\u003e with a minor version from 1.0 to 1.16 and\ncopy the URL to the address bar in your browser.\n\nHow it works\n\nGoogle Distributed Cloud extends\n[GKE](/kubernetes-engine/docs/concepts/kubernetes-engine-overview)\nto let you create GKE clusters in a\n[vSphere environment](https://docs.vmware.com/en/VMware-vSphere/index.html)\non your own premises, and manage them in Google Cloud along with regular\nGKE clusters and clusters in other environments as part of a\n[fleet](/anthos/fleet-management/docs).\n\nBecause the Google Distributed Cloud software runs in your data center rather than on\nGoogle Cloud, it requires you to install some admin and control plane software\nin addition to the GKE software itself. The software that runs in\nyour data center is downloaded as part of the installation and upgrade\nprocesses.\n\nThe following diagram shows the simplified result of a completed installation.\n[](/static/kubernetes-engine/distributed-cloud/vmware/docs/images/diagram-overview.svg) Google Distributed Cloud architecture with one user cluster\n\nKey components\n\nThe following components make up a software-only installation on VMware of\nGoogle Distributed Cloud:\n\n- A user cluster is where the workloads that implement your applications run,\n like in GKE on Google Cloud. The nodes that run your workloads\n are called *worker nodes*. A user cluster also has one or more control plane\n nodes. In the preceding diagram, the user cluster has one control plane node.\n\n- The admin cluster manages one or more user clusters. In the preceding diagram,\n the admin cluster has three control plane nodes.\n\n- The admin workstation is a separate machine that includes the tools that\n cluster creators and developers need to manage their installation:\n\n - Running `gkectl` from the admin workstation lets you create and update clusters and perform some other administrative tasks\n - Running `kubectl` from the admin workstation lets you interact with your admin and user clusters, including deploying and managing workloads\n- The Google Cloud console provides a web interface for your Google Cloud project,\n including your clusters on VMware. You can perform a subset of administrative\n tasks, including cluster creation, from the Google Cloud console as an\n alternative to running commands on the admin workstation.\n\n- Cluster admins and developers use `kubectl` and virtual IP addresses (VIPs)\n to access the control planes in the admin and user clusters. You configure\n VIPs during cluster creation. Users and developers calling workloads in your\n user clusters use Service and Ingress VIPs. Each node in the installation\n also has its own IP address. You can learn more about IP planning for\n Google Distributed Cloud in [Plan your IP addresses](/kubernetes-engine/distributed-cloud/vmware/docs/how-to/plan-ip-addresses).\n\nConnecting to the fleet\n\nAll Google Distributed Cloud clusters are members of a [fleet](/anthos/fleet-management/docs):\na logical grouping of Kubernetes clusters. Fleets let your organization uplevel management from individual clusters to entire groups of clusters, and can help your teams adopt similar best practices to those used at Google. You can view and manage fleet clusters together in the Google Cloud console, and use fleet-enabled GKE Enterprise features to help you manage, govern, and operate your workloads at scale. You can see a complete list of available fleet features for on-premises environments in [GKE Enterprise deployment options](/anthos/deployment-options#features_available_on_anthos_clusters_outside_of_google_cloud).\n\nEach fleet cluster's connection to Google Cloud is managed by a Connect Agent, which is deployed as part of the Google Distributed Cloud installation process. You can learn more about how this agent works in the [Connect Agent overview](/anthos/fleet-management/docs/connect-agent).\n\nFleet membership is also used to manage Google Distributed Cloud pricing, as described in the next section.\n\nFor a deeper discussion of GKE Enterprise features and how they\nwork together, see the\n[GKE Enterprise technical overview](/anthos/docs/concepts/overview).\n\nPricing\n\nGKE clusters on-premises created as part of Google Distributed Cloud are\nbilled per vCPU as part of GKE Enterprise. You enable the\nEnterprise tier by enabling the\n[Anthos API](https://console.cloud.google.com/apis/library/anthos.googleapis.com)\nin your Google Cloud project..\n\nFor full pricing information, including how to contact sales, see\n[GKE pricing](/kubernetes-engine/pricing).\n\nInstalling Google Distributed Cloud on VMware\n\nBecause the Google Distributed Cloud software runs in your own infrastructure, it is\nhighly configurable to meet your particular organizational and use case needs:\nyou can choose from a range of supported load balancing modes, vSphere\nconfigurations, IP addressing options, security features, connectivity options,\nand more. This means that setting up Google Distributed Cloud involves making decisions\nbefore and during installation in consultation with your networking, vSphere,\nand application teams to ensure that your installation meets your needs. This\ndocumentation set includes guides to help your team make these decisions.\n\nHowever, if you just need to see Google Distributed Cloud in action, we also provide a\nbasic installation path for a small test installation where we've made a lot of\nthese choices for you, letting you quickly get a workload up and running.\n\nIn each case, the installation process is as follows:\n\n1. **Plan your installation**. Minimally this includes ensuring you can meet the resource and vSphere requirements for Google Distributed Cloud, as well as planning your IP addresses.\n2. **Set up your on-premises environment** to support Google Distributed Cloud, including setting up vSphere inventory objects and your connection to Google.\n3. **Set up Google Cloud resources**, including the Google Cloud project you will use when setting up and managing Google Distributed Cloud.\n4. **Create an admin workstation** with the resources and tools you need to create clusters.\n5. **Create an admin cluster** to create, manage, and update user clusters.\n6. **Create user clusters** to run your actual workloads.\n\nWhat's next?\n\n- To start a minimal proof-of-concept installation, see [Set up minimal infrastructure](/kubernetes-engine/distributed-cloud/vmware/docs/how-to/minimal-infrastructure).\n- To review some of the considerations needed to plan a Google Distributed Cloud installation, start a production installation, or both, see the [installation overview](/kubernetes-engine/distributed-cloud/vmware/docs/how-to/install-overview)."]]
