Skip to main content
Uncategorized

Cross-Region Pilot Light Architecture Implementation

By July 3, 2026July 20th, 2026No Comments

Cross-Region Pilot Light Architecture Implementation

 

1. Project Overview

  • AWS Customer Name: SabioPay
  • AWS Partner Name: Techpartner
  • DR Strategy: Pilot Light
  • Industry: FinTech / Payment Acquirer
  • Workload: Mission-Critical Payment Processing & Merchant Onboarding

Customer Overview

Sabio Pay specializes in payment solutions that cater to a wide range of financial needs. Their platform allows you to make instant payments, manage transactions, and monitor your finances—all in one place. With intuitive features and robust security measures, they ensure that your money is always safe and your transactions are always seamless. 

2. Customer Challenge

SabioPay operated its core payment platform in a single AWS region (ap-south-1 Mumbai). As a critical payment acquirer handling digital checkout flows, merchant onboarding, and external bank integrations (specifically Jio Bank), any regional outage posed a catastrophic risk to merchant operations and financial settlement cycles.

Specific Challenges:

  • Single Region Dependency: An outage in Mumbai would halt transaction processing and merchant onboarding, leading to immediate revenue loss and a catastrophic breach of Jio Bank SLAs.
  • Strict Financial RPO: Regulatory mandates and transaction ledger reconciliation loops required an aggressive 15-minute Recovery Point Objective (RPO) to avoid manual bookkeeping discrepancies.
  • Banking Handshake Continuity: External banking partners (Jio Bank) utilize strict IP whitelisting; recovery required a strategy to maintain whitelisted connectivity paths without manual re-configuration delays.
  • Cost Optimization: With a primary infrastructure baseline of approximately $1,350/month, the business required a highly resilient solution but needed to minimize idle compute costs, strictly avoiding a heavily provisioned Warm Standby replica.

3. Partner Solution

Techpartner designed and implemented a Cross-Region Pilot Light architecture. This strategy was selected as the most cost-effective path to meeting SabioPay’s strict 15-minute RPO and 30-minute RTO targets by maintaining a live data layer while keeping the compute resources fully dormant in Hyderabad (ap-south-2).

Technical Architecture & Implementation

The solution maintains an ‘Always-Live’ data sync core while the rest of the application tier and helper resources are fully provisioned via Terraform but kept in a stopped, scaled-down state within a single AWS account.

A. Cross-Region Data Persistence

Techpartner implemented centralized continuous protection for the following resource types:

  • Amazon RDS (MySQL): Configured an active Cross-Region Read Replica in ap-south-2, providing continuous asynchronous transaction log streaming to meet the 15-minute RPO. This is the only resource kept actively running in the DR region.
  • Amazon EBS (SFTP Data Tier): Dynamically isolates stateful file layers by using Amazon Data Lifecycle Manager (DLM) to automate hourly snapshot copies to the secondary region, deliberately avoiding the overhead of AWS Backup.
  • S3 Buckets: Asynchronous Cross-Region Replication (CRR) with Versioning enabled for merchant onboarding documents and secure SFTP file payloads.

B. Automated Regional Recovery via Terraform & CodeDeploy

To guarantee the 45-minute RTO, Techpartner engineered a modular Terraform library that manages the regional environments. All infrastructure resources (VPC, subnets, Security Groups, IAM Roles, ALB, Auto Scaling Groups, and CodePipelines) are pre-created via Terraform and kept in a fully stopped/scaled-down state. If any core AMI changes occur in the primary region, the updated Golden AMIs are systematically copied cross-region to redefine the dormant fleet for future checkouts.

During a disaster failover, the Techpartner DevOps team triggers the activation which:

  • Programmatically promotes the active RDS cross-region read replica to standalone primary status.
  • Executes a Terraform script to instantly start the static helper instances (SFTP, OpenVPN, monitoring) and scale up the application Auto Scaling Groups.
  • Maintains pre-allocated, banking-whitelisted Elastic IPs (EIPs) on NAT Gateways to instantly re-establish secure Jio Bank API and SFTP connections.

Following instance ignition, the independent regional AWS CodePipeline triggers CodeDeploy to execute a fresh codebase deployment, pulling application packages from the replicated S3 buckets to spin up the latest Java and Laravel microservices across the newly active instances.

4. High-Level and Technical Detail of AWS Services Used

  • Amazon RDS: Utilized for continuous cross-region read replication and rapid database promotion to standalone primary.
  • Amazon Data Lifecycle Manager (DLM): Orchestrates automated hourly cross-region EBS snapshots for the stateful SFTP tier without overarching backup tools.
  • Terraform (IaC): Primary tool for managing deployment environment configuration, maintaining stopped-state environment parity, and executing scaling state changes.
  • AWS CodePipeline & CodeDeploy: Fully decoupled CI/CD deployment mechanisms running independently in the recovery region to automate codebase deployment across scaled EC2 fleets upon recovery activation.
  • Amazon CloudWatch & EventBridge: Monitors real-time replication lag metrics and triggers automated alerts if database sync time exceeds acceptable thresholds.

5. Outcomes and Quantitative Results

The Pilot Light implementation was validated through a formal, full-scale DR Game Day simulation on June 15, 2026.

Quantitative Metrics:

  • RPO Achievement: Verified data freshness at a consistent 35 seconds (well within the 15-minute target) via active RDS replication logs.
  • RTO Achievement: Full infrastructure ignition, instance start/scale-up, and live transaction processing achieved in 30 minutes (Target: 45 minutes).
  • Cost Efficiency: With a primary monthly bill of ~$1,350, SabioPay secured approximately $800/month infrastructure cost savings by holding all EC2 compute capacities strictly in stopped/scaled-down states via Terraform until disaster activation.
  • Operational Confidence: Successfully restored external Jio Bank handshakes immediately through pre-allocated EIP configurations without manual configuration lag.

6. Financial Impact and Risk Mitigation Performance

  • Baseline Exposure: Prior to remediation, SabioPay’s lack of a warm failover capability left its payment processing platform highly vulnerable to primary region disruptions. Relying on a traditional backup and restore path during an outage would result in a minimum of 4 hours of unmitigated platform downtime in ap-south-1 (exacerbated by the 30-to-45-minute recovery timeline required for the 306 GB RDS database alone). Operational modeling indicated this baseline posture incurred a total risk exposure of ₹2,75,000 per 4-hour window. This financial impact was driven by steady-state merchant transaction revenue leakage (₹25,000/hr), contractual Jio Bank SLA penalties for settlement cycle misses (₹50,000 flat rate), central banking regulatory compliance risk (₹75,000), and merchant trust degradation leading to customer churn (₹50,000).
     
  • Target Objectives: To safeguard SabioPay’s high-velocity Jio Bank processing pipeline against these compounding financial and regulatory exposures, the engineering group targeted a highly responsive business-critical architecture. The agreed-upon strategy mandated an aggressive target Recovery Time Objective (RTO) of less than 45 minutes to prevent triggering contractual SLA breach notifications, and a near-zero Recovery Point Objective (RPO) to eliminate transaction data gaps. This required a cross-region Pilot Light disaster recovery framework designed for deterministic, low-latency cross-region failovers. 
     
  • Actual Results Achieved: SabioPay successfully deployed an AWS-compliant Pilot Light disaster recovery model in ap-south-2 (Hyderabad), anchoring the architecture with an always-live, real-time Amazon RDS cross-region Read Replica paired with dormant compute layers. The actual results demonstrated a highly cost-optimized infrastructure insurance posture, keeping total DR spend to a lean $550.00 per month ($6,600 annually) by completely eliminating idle, steady-state compute overhead. This strategic investment successfully neutralizes the ₹2,75,000 downtime risk exposure per incident, ensuring immediate failover readiness while fully securing the platform against existential financial, regulatory, and reputational liabilities during a regional AWS outage 

7. Conclusion

By combining AWS’s native continuous replication mechanisms with Terraform-driven zero-scale computing, Techpartner successfully transitioned SabioPay to an audit-ready, highly resilient active-passive state. The Pilot Light strategy provides an elegant architectural balance—eliminating heavy idle infrastructure costs to save $800/month while guaranteeing that SabioPay can recover from a total Mumbai regional failure within minutes, protecting both merchant onboarding revenue and Jio Bank integration compliance.