The backend endpoints of Kubernetes-native Services are randomly distributed across nodes. Consequently, when Service requests are distributed to nodes across node groups, these requests may fail to reach the nodes or may not be answered promptly. You can configure a Service topology to expose an application on an edge node only to the current node or nodes in the same edge node pool. This topic describes how a Service topology works and how to configure a Service topology.
In edge computing, edge nodes are classified into groups by zone, region, or other logical attribute, such as CPU architecture, Internet service provider (ISP), or cloud service provider. Nodes in different groups are isolated from each other in one way or another. For example, these nodes may be disconnected, may not share the same resources, may have heterogeneous resources, or may run applications that are independently deployed.
How a Service topology works
To solve the preceding issues, ACK@Edge provides a feature to manage the topology of endpoints of Kubernetes-native Services. You can configure a Service topology to specify how endpoints of a Service are accessed. For example, you can configure a Service topology to expose an application on an edge node only to the current node or nodes in the same edge node pool. The following figure shows how a Service topology works.
- Service 1 is associated with Pod 2 and Pod 3. The
annotation:"openyurt.io/topologyKeys: kubernetes.io/zone"annotation specifies the node pool that is allowed to access Service 1.
- Pod 2 is deployed on Node 2 and Pod 3 is deployed on Node 4. Node 2 belongs to Node Pool A and Node 3 belongs to Node Pool B.
- Pod 3 and Pod 1 do not belong to the same node pool. As a result, when Pod 1 accesses Service 1, the traffic is forwarded only to Pod 2. The traffic is not forwarded to Pod 3.
You must add the Service topology annotation to a Service when you create the Service. Otherwise, the Service topology does not take effect. If you add the annotation after the Service is created, the Service topology does not take effect. You must delete and recreate the Service.
Method 1: Configure a Service topology in the console
To create a Service that can be accessed only by nodes in the node pool where the
Service is deployed, you need only to add an annotation to the Service. For example,
you can set Name to
openyurt.io/topologyKeys and Value to
kubernetes.io/zone. For more information about how to create a Service, see Manage Services.
Method 2: Configure a Service topology by using a CLI
Create a Service that uses the topological domain of a specific node pool. The following code block is an example of the YAML template:
apiVersion: v1 kind: Service metadata: annotations: openyurt.io/topologyKeys: kubernetes.io/zone name: my-service-nodepool namespace: default spec: ports: - port: 80 protocol: TCP targetPort: 8080 selector: app: nginx sessionAffinity: None type: ClusterIP
|Annotation key||Annotation value||Description|
|openyurt.io/topologyKeys||kubernetes.io/hostname||Specifies that the Service can be accessed only by the node where the Service is deployed.|
|openyurt.io/topologyKeys||kubernetes.io/zone or openyurt.io/nodepool||Specifies that the Service can be accessed only by the nodes in the node pool where the Service is deployed. If the version of the edge Kubernetes cluster is 1.18 or later, we recommend that you use openyurt.io/nodepool.|
|-||-||Specifies that access to the Service is unlimited.|