#Introduction

#Deployment Model

Due to considerations of service performance, stability, security isolation, latency, location, and cluster scalability, enterprises typically deploy services across multiple clusters in different data centers (across regions and clouds).

The platform's multi-cluster service mesh is implemented based on Istio's multiple control planes deployment model. In this model, the service mesh can manage multiple primary clusters, each with its own Istio control plane. Services deployed in various clusters under the mesh can discover each other and share the load.

#Features

#Integrated with Istio's Native Capabilities

• Failover: If a service in one cluster fails and triggers a circuit breaker, part of the traffic intended for that service is redirected to healthy pods in clusters in the disaster recovery region to ensure overall service load performance stability.
• Weighted Distribution: Traffic is routed to clusters in multiple regions based on configured weights, achieving multi-region load balancing.

#Unified UI for Easy Configuration and Management

• You can create a service mesh or manage clusters by configuring simple parameters via the UI.
• The platform provides a unified operation and configuration management interface for the multi-cluster service mesh. You can view and manage the configurations and monitoring data of different clusters within the mesh on a single page.
• The platform automatically deploys Istio east-west traffic gateways for all clusters in a multi-network mode service mesh. You only need to add the cluster to the mesh.

#Cross-Cluster Visualization

• Compare traffic monitoring data of services in different clusters within the mesh.
• View cross-cluster service topology.
• View cross-cluster call chains.

#Application Scenarios

#Cross-Cluster Failover

Deploy services in clusters located in different regions within the mesh. If services in a cluster in one region fail to handle traffic, the platform can smoothly shift the traffic to clusters in disaster recovery regions according to the pre-configured failover strategy, ensuring normal business operations.

#Cross-Cluster Weighted Distribution

• Scenario 1: Load traffic proportionally to multiple clusters

  • When services have specific regional requirements and can only meet availability, latency, and other needs by being deployed in specific regions, you can adjust the cluster's traffic distribution weight to route a designated proportion of traffic to the clusters in those regions.
  • When service operation demands high performance from the hosting cluster, you can adjust the cluster's traffic distribution weight to route more traffic to clusters with better performance.

• Scenario 2: Cross-cluster service migration When enterprises need to migrate services across clusters (across clouds or regions), they can add the target cluster to the mesh and deploy the services in that cluster. Then, by configuring traffic distribution weights for the source and target clusters through the platform, services can be smoothly migrated to the target cluster without interrupting business operations. During this process, you can monitor the service performance in the target cluster and adjust weights in real-time.