Added details on network communication. (#295)

oilbeater · web-flow · commit 6ce6d8dcd324 · 2025-07-08T10:40:53.000+08:00
Signed-off-by: Mengxin Liu &lt;liumengxinfly@gmail.com&gt;
diff --git a/docs/advance/offload-mellanox.en.md b/docs/advance/offload-mellanox.en.md
@@ -230,7 +230,7 @@ This plugin creates device plugin endpoints based on the configurations given in
   - `devices`: Target Devices' device Hex code as string
   - `drivers`: Target device driver names as string
 
-`selectors` also supports VF selection based on `pciAddresses`, `acpiIndexes` and other parameters. For more detailed configuration, please refer to [SR-IOV ConfigMap configuration](https://github.com/k8snetworkplumbingwg/sriov-network-device- plugin/tree/v3.6.2?tab=readme-ov-file#configurations)
+`selectors` also supports VF selection based on `pciAddresses`, `acpiIndexes` and other parameters. For more detailed configuration, please refer to [SR-IOV ConfigMap configuration](https://github.com/k8snetworkplumbingwg/sriov-network-device-plugin/tree/v3.6.2?tab=readme-ov-file#configurations)
 
 Please read the [SR-IOV device plugin](https://github.com/intel/sriov-network-device-plugin) to deploy:
 
diff --git a/docs/advance/windows.en.md b/docs/advance/windows.en.md
diff --git a/docs/advance/windows.md b/docs/advance/windows.md
diff --git a/docs/guide/subnet.en.md b/docs/guide/subnet.en.md
@@ -62,6 +62,8 @@ In the Kubernetes network specification, it is required that Nodes can communica
 To achieve this in Overlay network mode, Kube-OVN creates a `join` Subnet and creates a virtual NIC `ovn0`
 at each node that connect to the `join` subnet, through which the nodes and Pods can communicate with each other.
 
+All network communication between Pods and Nodes will go through the `ovn0` network interface. When a Node accesses a Pod, it enters the virtual network via the `ovn0` interface, and the virtual network then connects to the host network through the `ovn0` interface.
+
 The configuration of this Subnet is specified at installation time, you can refer to [Built-in Network Settings](setup-options.en.md#built-in-network-settings) for more details.
 To change the CIDR of the Join Subnet after installation please refer to [Change Join CIDR](../ops/change-join-subnet.en.md).
 
diff --git a/docs/guide/subnet.md b/docs/guide/subnet.md
@@ -57,6 +57,8 @@ spec:
 为了在 Overlay 网络模式下达到这个目的， Kube-OVN 创建了一个 `join` 子网，
 并在每个 Node 节点创建了一块虚拟网卡 ovn0 接入 `join` 子网，通过该网络完成节点和 Pod 之间的网络互通。
 
+所有 Pods 和 Nodes 之间的网络通信都会通过 ovn0 网卡进行，Node 访问 Pod 时通过 ovn0 网卡进入虚拟网络，虚拟网络通过 ovn0 网卡进入主机网络。
+
 该子网的配置为安装时指定，可以参考[内置网络设置](setup-options.md#_2)，如果要在安装后修改。
 join 子网的 CIDR 请参考[修改 Join 子网](../ops/change-join-subnet.md)
 
diff --git a/docs/index.en.md b/docs/index.en.md
@@ -25,7 +25,7 @@ It currently supports [Subnet Management](guide/subnet.en.md), [Static IP Alloca
 [Distributed/Centralized Gateways](guide/subnet.en.md#overlay-subnet-gateway-settings), [Underlay/Overlay Hybrid Networks](start/underlay.en.md),
 [VPC Multi-Tenant Networks](vpc/vpc.en.md), [Cross-Cluster Interconnect](advance/with-ovn-ic.en.md), [QoS Management](guide/qos.en.md),
 [Multi-NIC Management](advance/multi-nic.en.md), [ACL](guide/subnet.en.md#subnet-acl), [Traffic Mirroring](guide/mirror.en.md),
-ARM Support, [Windows Support](advance/windows.en.md), and many more.
+ARM Support, and many more.
 
 **Extreme Performance:**
 
diff --git a/docs/index.md b/docs/index.md
@@ -22,8 +22,7 @@ OVN 和 OVS 有着悠久的历史，早在 Kubernetes 诞生前就已经出现
 借助 OVS/OVN 在 SDN 领域成熟的能力，Kube-OVN 将网络虚拟化的丰富功能带入云原生领域。目前已支持[子网管理](guide/subnet.md)，
 [静态 IP 分配](guide/static-ip-mac.md)，[分布式/集中式网关](guide/subnet.md#overlay)，[Underlay/Overlay 混合网络](start/underlay.md)，
 [VPC 多租户网络](vpc/vpc.md)，[跨集群互联网络](advance/with-ovn-ic.md)，[QoS 管理](guide/qos.md)，
-[多网卡管理](advance/multi-nic.md)，[ACL 网络控制](guide/subnet.md#acl)，[流量镜像](guide/mirror.md)，ARM 支持，
-[Windows 支持](advance/windows.md)等诸多功能。
+[多网卡管理](advance/multi-nic.md)，[ACL 网络控制](guide/subnet.md#acl)，[流量镜像](guide/mirror.md)，ARM 支持。
 
 **极致的性能：**
 
diff --git a/docs/reference/feature-stage.md b/docs/reference/feature-stage.md
@@ -67,7 +67,6 @@
 | [自定义 VPC 互联](../vpc/vpc-peering.md) | false | Alpha | 1.10      |           |
 | [Mellanox Offload](../advance/offload-mellanox.md) | false | Alpha | 1.8       |           |
 | [芯启源 Offload](../advance/offload-corigine.md) | false | Alpha | 1.10      |           |
-| [Windows 支持](../advance/windows.md) | false | Alpha | 1.10      |           |
 | [DPDK 支持](../advance/dpdk.md) | false | Alpha | 1.10      |           |
 | [OpenStack 集成](../advance/dpdk.md) | false | Alpha | 1.9       |           |
 | [单个 Pod 固定 IP/Mac](../guide/static-ip-mac.md) | true  | GA    | 1.8       |           |
diff --git a/docs/vpc/vpc-internal-lb.en.md b/docs/vpc/vpc-internal-lb.en.md
@@ -79,7 +79,7 @@ If both `selector` and `endpoints` are configured, the `selector` configuration
 `OVN` supports health checks for load balancer endpoints, for IPv4 load balancers only.
 When health checks are enabled, the load balancer uses only healthy endpoints.
 
-[[Health Checks](https://www.ovn.org/support/dist-docs/ovn-nb.5.html)](https://www.ovn.org/support/dist-docs/ovn-nb.5.html)
+[[Health Checks](https://www.ovn.org/support/dist-docs/ovn-nb.5.html)](<https://www.ovn.org/support/dist-docs/ovn-nb.5.html>)
 
 Add a health check to `SwitchLBRule` based on the health check of the `ovn` load balancer.While creating the `SwitchLBRule`, obtain a reusable `vip` from the corresponding `VPC` and `subnet` as the detection endpoint and associate the corresponding `IP_Port_Mappings` and `Load_Balancer_Health_Check` to the corresponding load balancer.
 
diff --git a/docs/vpc/vpc-internal-lb.md b/docs/vpc/vpc-internal-lb.md
@@ -84,7 +84,7 @@ spec:
 `OVN` 支持 `IPv4` 的负载平衡器服务终端的运行状况检查。
 启用运行状况检查后，负载平衡器会对服务终端的状态进行检测维护，并仅使用运行状况良好的服务终端。
 
-[[Health Checks](https://www.ovn.org/support/dist-docs/ovn-nb.5.html)](https://www.ovn.org/support/dist-docs/ovn-nb.5.html)
+[[Health Checks](https://www.ovn.org/support/dist-docs/ovn-nb.5.html)](<https://www.ovn.org/support/dist-docs/ovn-nb.5.html>)
 
 根据 `ovn` 负载均衡器的运行状况检查，对 `SwitchLBRule` 添加健康检查。在创建 `SwitchLBRule` 的同时，从对应的 `VPC` 和 `subnet` 中获取一个可复用的 `vip` 作为检测端点，并添加对应的 `ip_port_mappings` 和 `load_balancer_health_check` 到对应的负载均衡器上。
 
diff --git a/docs/vpc/vpc.en.md b/docs/vpc/vpc.en.md
@@ -16,6 +16,10 @@ and can be configured separately with Subnet CIDRs, routing policies, security p
 
 ![](../static/network-topology.png)
 
+## Implementation Principle
+
+In Kube-OVN, each VPC is mapped to a [Logical Router](https://man7.org/linux/man-pages/man5/ovn-nb.5.html#Logical_Router_TABLE){: target="_blank" } in OVN. Multiple logical routers are independent network units, and each logical router can be associated with its own [Logical Switches](https://man7.org/linux/man-pages/man5/ovn-nb.5.html#Logical_Switch_TABLE){: target="_blank" }, thereby having independent network ports and IP address spaces. Traffic from different VPCs is distinguished and isolated by different Datapath IDs when traversing tunnels, and forwarding is performed based on Datapath IDs. This ensures IP address space isolation, allowing the same IP addresses to be used in different VPCs without conflicts. For tunnel encapsulation formats, refer to [OVN Architecture Design Decisions](https://man7.org/linux/man-pages/man7/ovn-architecture.7.html#DESIGN_DECISIONS){: target="_blank" }.  
+
 ## Creating Custom VPCs
 
 Create two VPCs:
diff --git a/docs/vpc/vpc.md b/docs/vpc/vpc.md
@@ -13,6 +13,10 @@ Kube-OVN 支持多租户隔离级别的 VPC 网络。不同 VPC 网络相互独
 
 ![](../static/network-topology.png)
 
+## 实现原理
+
+在 Kube-OVN 中，每个 VPC 会映射到 OVN 中的一个[逻辑路由器](https://man7.org/linux/man-pages/man5/ovn-nb.5.html#Logical_Router_TABLE){: target="_blank" }。多个逻辑路由器之间是相互独立的网络单元，每个逻辑路由器下可以关联各自的[逻辑交换机](https://man7.org/linux/man-pages/man5/ovn-nb.5.html#Logical_Switch_TABLE){: target="_blank" }，进而有独立的网络端口和 IP 地址空间。不同 VPC 下的流量在经过隧道时 OVN 会通过不同的 Datapath ID 进行区分和隔离，并通过 Datapath ID 进行转发，从而可以做到 IP 地址空间相互隔离，不同 VPC 下可以使用相同的 IP 地址而不会冲突。隧道封装格式请参考[OVN Architecture Design Decisions](https://man7.org/linux/man-pages/man7/ovn-architecture.7.html#DESIGN_DECISIONS){: target="_blank" }。
+
 ## 创建自定义 VPC
 
 创建两个 VPC：
diff --git a/mkdocs.yml b/mkdocs.yml
@@ -72,7 +72,6 @@ nav:
       - 芯启源网卡 Offload 支持: "advance/offload-corigine.md"
       - 云脉芯联网卡 Offload 支持: "advance/offload-yunsilicon.md"
       - 中科驭数网卡 Offload 支持: "advance/offload-yusur.md"
-      - Windows 支持: "advance/windows.md"
       - DPDK 支持: "advance/dpdk.md"
       - OpenStack 集成: "advance/with-openstack.md"
       - OVN IPsec 支持: "advance/ovn-ipsec.md"
@@ -222,7 +221,6 @@ plugins:
             芯启源网卡 Offload 支持: Offload with Corigine
             云脉芯联网卡 Offload 支持: Offload with Yunsilicon
             中科驭数网卡 Offload 支持: Offload with YUSUR
-            Windows 支持: Windows Support
             DPDK 支持: DPDK Support
             OpenStack 集成: Integration with OpenStack
             总体架构: Architecture
