Skip to content

brianmlasky/multi-cloud-dr-platform

Repository files navigation

Multi-Cloud Disaster Recovery (DR) Platform

Workload Tier: Tier 2 (Business Critical)
Topology: Active-Passive (Primary: GCP, Passive: AWS)
RTO Target: 4 Hours | RPO Target: 15 Minutes

Executive Summary

This repository contains the infrastructure-as-code and architectural documentation for a FinOps-optimized, multi-cloud disaster recovery platform. It utilizes a Pilot Light strategy, keeping idle AWS infrastructure costs below 5% of the primary GCP environment while guaranteeing strict recovery SLAs.

This project goes beyond infrastructure provisioning, heavily emphasizing Enterprise Architecture, Fiscal SecOps, and Regulatory Compliance (SOC2/ISO 27001).

🗺️ The Architecture Documentation

Hiring Managers and Auditors are encouraged to review the documentation below, which outlines the system physics, threat models, and management decisions.

  1. The Management Response Matrix: The Principal Architect's formal response to an adversarial security and RTO audit, detailing accepted business risks and applied remediations.
  2. Business Continuity & FinOps Parameters: The mathematical RTO/RPO limits and the CTO/CFO translation constraints.
  3. Security Architecture Spec & Threat Model: STRIDE analysis, Data Classification, and split-brain corruption mitigations.
  4. SEV-1 Disaster Runbook: The step-by-step incident response playbook detailing the human-in-the-loop failover execution.

🏗️ Infrastructure Modules

  • infrastructure/bootstrap/: Multi-region S3 Terraform State and Zero-Trust GitHub Actions OIDC federation.
  • infrastructure/network/: Non-overlapping VPC topology with S3/DynamoDB Gateway Endpoints.
  • infrastructure/compute/: Amazon EKS Pilot Light cluster utilizing Managed Node Groups.
  • infrastructure/database/: Asynchronous Change Data Capture (CDC) replication via AWS DMS to a pre-sized RDS PostgreSQL target.
  • infrastructure/routing/: Amazon Route 53 Global Traffic Management and CloudWatch alerting.

Deployment & Failover

Failover execution is strictly guarded by the Incident Commander to prevent split-brain data corruption. See the Adversarial Game Day Playbook for Chaos Engineering and DiRT validation protocols.

About

Active-Passive DR Platform (GCP to AWS) | Pilot Light FinOps topology (<5% idle cost) | Architected for strict 4-Hour RTO & 15-Min RPO.

Topics

Resources

Stars

1 star

Watchers

0 watching

Forks

Releases

No releases published

Packages

 
 
 

Contributors