How ACK One Multi‑Cluster Gateway Enables Seamless Cross‑AZ and Multi‑Region Disaster Recovery

Overview

The enterprise business system is typically divided into three layers: access, application, and data. Each layer requires specific disaster‑recovery measures to achieve end‑to‑end business continuity.

Business Layer Architecture

Access layer: traffic entry point that forwards requests to the backend application layer based on routing rules.

Application layer: services that process requests and return results to upstream callers.

Data layer: storage services that provide and persist data for the application layer.

ACK One Multi‑Cluster Gateway Workflow

Create a fleet ("舰队").

Register clusters (ACK clusters or registered clusters) to the fleet for unified management.

Create a multi‑cluster gateway in the fleet using AlbConfig or MseIngressConfig.

Create an Ingress in the fleet, bind services from member clusters, and define routing rules.

Access services via the gateway’s domain name or IP, which forwards traffic to the appropriate cluster service.

Same‑City Cross‑AZ Active‑Active DR

ACK One supports both ALB and MSE multi‑cluster gateways to build a same‑city cross‑AZ active‑active disaster‑recovery system.

Create two ACK clusters in different AZs within the same region.

Distribute services to both clusters via ACK One’s GitOps capability.

Deploy a fleet and create a multi‑cluster gateway.

Create an Ingress that automatically reroutes traffic to the healthy cluster when a failure occurs.

Gateway features include load‑balancing by total replica count, weighted routing, HTTP‑header based routing for gray releases, and millisecond‑level failover.

Hybrid‑Cloud Same‑City DR

The solution extends to hybrid‑cloud scenarios by registering on‑premise or third‑party public‑cloud Kubernetes clusters as registered clusters and connecting them to the ACK One fleet.

Register IDC or third‑party clusters to the fleet (AZ2) and establish a dedicated line to the cloud network.

Create a fleet in the same VPC/Region and an ACK cluster in AZ1.

Distribute services to both the cloud and IDC clusters via GitOps.

Create a multi‑cluster gateway and an Ingress to enable automatic cross‑AZ failover.

Ensure VPC‑to‑VPC or CEN connectivity for cross‑region traffic.

Inter‑Region DR

For disaster scenarios affecting an entire region, ACK One leverages ALB multi‑cluster gateways together with Global Traffic Manager (GTM) to provide seamless failover.

Deploy an ACK cluster in each region and create an ALB multi‑cluster gateway in Region 1.

Use GTM to bind the ALB gateway in Region 1 with the ALB instance in Region 2.

The multi‑cluster gateway handles 7‑layer traffic forwarding; GTM switches IP only when an entire region fails.

Data synchronization for MySQL/RDS relies on middleware capabilities.

Comparison with DNS‑Based DR

DNS‑based solutions suffer from client‑side cache delays (minutes) and cannot support advanced routing features such as QUIC 0‑RTT.

ACK One’s gateway provides millisecond‑level failover without DNS cache issues.

Unified management of Ingress configurations reduces operational overhead.

Cluster upgrades or rebuilds are transparent because traffic is automatically rerouted to healthy clusters.

Conclusion

ACK One multi‑cluster gateway enables rapid construction of same‑city cross‑AZ active‑active DR, hybrid‑cloud active‑active DR, and inter‑region DR with millisecond‑level failover, centralized traffic management, and lower cost compared to traditional DNS‑based approaches.