View a markdown version of this page

LSREL02-BP01 Build resilient and highly available research solutions - Life Sciences Lens
Implementation guidanceResources

LSREL02-BP01 Build resilient and highly available research solutions

Design research systems using fault-isolated, redundant deployment patterns that allow continuous access during maintenance or localized outages. Use highly available configurations where near-zero downtime is required (for example, ELNs) and active-passive designs with automated failover for systems with less stringent latency or availability needs (for example, LIMS).

Desired outcome:

  • Research systems remain available during maintenance or outages.

  • Ongoing experiments and data collection continue without interruption.

  • Researchers have a consistent experience across sites and geographies.

Common anti-patterns:

  • Relying on single-instance deployments for critical systems.

  • Maintenance performed without phased rollouts or rollback plans.

  • Highly available architectures implemented without routine failover testing.

Benefits of establishing this best practice:

  • Protects ongoing experiments from interruption, reducing risk of wasted samples or time.

  • Preserves continuity in multi-day or multi-week lab workflows.

  • Provides global research teams access to shared tools without productivity loss.

  • Enhances regulatory readiness by reducing gaps in system availability records.

Level of risk exposed if this best practice is not established: High

Implementation guidance

When designing research workloads for resiliency, use deployment topologies that minimize the scope of impact of maintenance and outage events. Multi-site or multi–AZ deployments allow for uninterrupted operation when a system component is taken offline. Automated health checks and intelligent traffic routing directs users to healthy endpoints, maintaining continuity during upgrades or failover scenarios. Maintenance windows should be orchestrated to minimize user disruption, and resilience should be validated regularly through simulated failovers or planned game days.

Implementation steps

  1. Deploy AWS IoT Greengrass for local buffering of instrument data.

  2. Deploy LIMS, EHR and ELN solutions across multiple Availability Zones using Amazon EC2 Auto Scaling for workload redundancy.

  3. Place applications behind an Elastic Load Balancer (ALB or NLB) to support highly available deployment configurations, or configure Amazon Route 53 DNS failover with health checks for active-passive failover strategies.

  4. Use AWS Systems Manager to coordinate upgrades with minimal disruption. Monitor resilience, health status, and failover events in Amazon CloudWatch, and validate readiness with regular failover exercises.

Resources

Related best practices:

  • Change management and controlled rollout strategies

  • Incident response playbooks for scientific research systems

  • Risk-based classification of R&D applications