Singapore insurer: cross-cloud DR architecture with local operations

Context

The client is a Singapore-licensed insurer. Their core business systems run on AWS EKS, backed by message queues, EFS and S3 and other AWS-native services. When things are healthy, everything hums. But MAS and internal audit both require:

• If an AWS region goes down, business has to be recoverable on-premises
• The recovery process has to be rehearsable, observable and produce an audit report
• The recovery process can’t stay on paper — someone has to actually execute it

Their previous DR plan was “backup + document” — when something breaks, someone works through it step by step. In MAS’s view, that’s no longer enough.

What we did

TWO TWO built a cross-cloud, cross-environment DR architecture from AWS to the local datacenter using Jumborca SupDRC + SupKube:

1. Kubernetes layer: EKS ↔ on-prem K8s

• SupKube structurally syncs AWS EKS workloads (Deployment / StatefulSet / Service / ConfigMap / Secret) to the on-prem Kubernetes cluster
• The local K8s cluster keeps headroom sized to the client’s traffic profile — ready to take production on switchover

2. Data layer: EFS / S3 → local storage

• EFS file data syncs to the local distributed file system
• S3 object storage replicates to the local object store via SupDRC’s object sync
• Message queues (core to insurance event flow) sync with policies ensuring no message loss on switchover

3. Orchestration: SupDRC runbook

• SupDRC provides a runbook script engine that encodes the full flow — “detect fault → check state → stop writes on AWS → bring up local K8s → cut DNS/ingress → notify business” — into an executable script
• One-button failover — initiated from the console, checkpoints automated, no reliance on human memory
• Rehearsals run on a schedule, producing structured audit reports

4. 24×7 local operations

TWO TWO’s local on-call team monitors cross-cloud sync state, supports rehearsals and handles incidents. Client IT participates in critical decisions; daily operations sit with us.

Results

• Cross-cloud DR is actually executable — AWS EKS ↔ on-prem K8s structurally consistent; business can recover locally
• One-button switchover — via SupDRC runbook, “paper plan” becomes “console button”
• Rehearsable — regular drills produce audit reports; a credible posture in MAS’s view
• Local operations — 24×7 coverage; client IT team is no longer the single point in the recovery path

Full case is NDA-bound. Reach out to discuss the architecture, delivery timeline and operating model in a private conversation.