A nationally significant mining and precious metals enterprise was running its primary and disaster recovery data centres in the same metropolitan area. PT Metrodata Electronics Tbk replaced the standby facility with a Multi-AZ pilot light architecture in the AWS Jakarta Region, validated through two full drills.

ABOUT THE CUSTOMER

The customer is one of Indonesia’s largest state-owned enterprises, operating in the mining and precious metals sector with nationwide mining assets and a gold, silver and precious metals retail network serving customers across the country. As a publicly listed state-affiliated enterprise of national strategic significance, the organisation depends on stable, secure and highly available IT systems to support high-volume precious metals trading, ERP-driven financial transactions, inventory management and regulatory reporting to government stakeholders.

CUSTOMER CHALLENGE

As a regulated state-owned entity, the customer was required to maintain a Disaster Recovery Centre with genuine geographic separation from its primary data centre. Its existing arrangement did not deliver that. Both the primary data centre and the designated disaster recovery facility were on-premises within the same metropolitan area, sharing exposure to the same regional hazards — flooding, power grid failure, seismic events. The secondary site provided hardware redundancy but not geographic redundancy, and the distinction is the one that matters in a regional event.

The operational position was equally exposed. Recovery time depended on manual failover onto cold standby hardware and ran to 8–12 hours. Recovery point was governed by daily backup windows and batch replication, leaving 4–6 hours of transactional data at risk. The customer estimated revenue impact at USD 50,000 to USD 100,000 per hour of downtime during high-volume precious metals trading periods. Full disaster recovery drills were expensive and disruptive enough that they were conducted infrequently, which meant the recovery capability was largely unproven. Meanwhile the standby facility consumed substantial capital on servers, storage, networking and facilities that sat idle.

Two further constraints shaped the solution. The procurement terms of reference required that all activities be performed within Indonesian data centre territory, making data sovereignty a contractual obligation rather than a preference. And the terms set a formal service level for the recovery capability itself: a disaster recovery drill or switchover completed within 4 hours, with a recovery point of 30 minutes measured from the declaration of the activity, alongside a 99.9 percent replication service level and a 15-minute response time under 24×7 support.

PARTNER SOLUTION

PT Metrodata Electronics Tbk designed and implemented a cloud-based disaster recovery solution using AWS Elastic Disaster Recovery with a pilot light strategy, deployed in the AWS Asia Pacific (Jakarta) Region. The design replaced the on-premises disaster recovery facility entirely while keeping every workload and every byte of replicated data inside Indonesia, satisfying the contractual sovereignty requirement. Because Jakarta is the only AWS Region in Indonesia, the geographic separation the regulation demanded was delivered through a Multi-AZ architecture spanning three Availability Zones — physically separate facilities with independent power, cooling and networking — rather than through a second Region.

Metrodata began with a discovery and compliance mapping phase across the in-scope estate of 50 servers, comprising 32 Linux and 18 Windows systems totalling 376 vCPUs, 1,661 GB of memory and approximately 26.9 TB of protected storage. The team mapped each compliance obligation to a specific technical control and confirmed target recovery objectives against the contracted service levels. Workloads spanned the SAP landscape — S/4HANA, HANA database, Fiori, Solution Manager and Sybase — alongside non-SAP production systems including ERP application and database servers, the precious metals retail platform, e-procurement, data warehouse and ETL servers, DNS, and Active Directory.

The architecture used AWS Transit Gateway in a hub-and-spoke topology to centralise connectivity across multiple VPCs, with six subnets distributed across the three Availability Zones — three private for the DRS staging and recovery areas, three public for controlled egress. AWS Organizations provided multi-account governance and AWS IAM enforced least-privilege access to recovery operations. Amazon CloudWatch monitored replication health and instance metrics. Access for the customer’s own team during drills was provided through a dedicated VPN gateway, with all recovery servers held in private subnets and no public exposure.

Deployment placed AWS DRS replication agents across the server fleet, replicating continuously over an encrypted channel to Amazon EBS volumes in the staging area. Recovery servers exist as configured launch templates rather than running instances, with instance type, security group and subnet placement pre-defined per server, so that a recovery starts correctly sized and correctly networked without manual configuration under pressure. Metrodata verified per-server launch templates ahead of each drill and confirmed replication at 100 percent with no lag and no backlog before declaring readiness.

Metrodata provided support services across both phases of the engagement. Before implementation, the partner conducted the assessment, designed the architecture, deployed and tuned the replication agents, prepared the isolated drill network, and authored the drill rundowns and recovery runbooks. After implementation, Metrodata provides 24×7 managed support with a 15-minute response commitment, acts as level 2 escalation behind the customer’s own ICT team, monitors replication continuously, and executes disaster recovery drills twice per year with documented results.

RESULTS AND BENEFITS

The capability was validated through two full disaster recovery drills covering the two halves of the estate. The non-SAP drill recovered twelve production servers — ERP application and database, the precious metals retail platform, e-procurement, data warehouse, ETL, DNS and Active Directory. From the declaration of the drill at 10:05, every server was recovered, powered on and passing status checks by 11:30, and full service validation across all twelve completed at 12:22. Total elapsed time was 2 hours 17 minutes, comfortably inside the contracted 4-hour service level.

The SAP drill produced the more instructive result. Five SAP servers — HANA database, two S/4HANA application servers, Fiori and Solution Manager — were recovered sequentially by priority. AWS DRS completed the infrastructure recovery of all five in 27 minutes of elapsed time against a 120-minute plan, with individual servers recovering in 10 to 14 minutes against per-server plans of 25 to 30 minutes. The remaining time was consumed by SAP BASIS work performed by the customer’s own team: parameter configuration, database and application service startup, connection verification and SAPGUI login across HANA, S/4HANA, Fiori, Sybase and Solution Manager, which ran from 10:50 to 12:50. The full drill completed in approximately 3 hours 10 minutes from declaration — inside the contracted 4-hour SLA — and also met its own acceptance criteria of an 18-hour recovery time and a 1-hour recovery point.

The two drills isolate where recovery time actually goes. Infrastructure recovery through AWS DRS took 27 minutes for the full SAP landscape and beat its plan by a factor of four. Application startup for a complex SAP estate took about two hours. For any organisation planning SAP disaster recovery, that ratio is the number worth knowing in advance.

Beyond the measured recovery figures, the engagement changed the customer’s regulatory position. The organisation now holds a disaster recovery capability with genuine geographic separation across three Availability Zones, documented runbooks, and drill evidence it can place in front of a regulator — replacing a same-metro standby facility whose recovery capability had never been fully proven. Because the pilot light pattern holds no running duplicate environment, that improvement was achieved while eliminating the capital expenditure the on-premises replacement would have required. Drills that were previously too expensive and disruptive to run often are now executed twice a year on an isolated network segment with no impact on production.

ABOUT THE PARTNER

PT Metrodata Electronics Tbk (IDX: MTDL) is one of Indonesia’s leading digital solution providers and an AWS Advanced Tier Services Partner, trusted by more than 200 organisations across healthcare, financial services, manufacturing, retail, mining and the public sector in adopting and implementing cloud solutions. Metrodata currently holds the AWS Migration and Modernization Competency and the AWS Glue Service Delivery designation, demonstrating validated technical expertise in cloud migration, data integration, and business continuity. Delivering through its consulting and system integration subsidiaries PT Mitra Integrasi Informatika and PT Sinergi Transformasi Digital, the company maintains a dedicated practice of certified AWS solutions architects and engineers supporting customers throughout Indonesia.