Operating Hybrid Cloud with Amazon EVS - DR Site Construction and Burst Capacity
Learn how to operate a hybrid cloud with Amazon EVS, including DR site construction, on-demand capacity bursting, and integration patterns with AWS services.
Hybrid Cloud Design Patterns with EVS
Amazon EVS functions as an extension of on-premises VMware environments, making it well-suited for building hybrid clouds. There are three representative design patterns. (1) DR Site Pattern: Use on-premises as the primary site and EVS as the secondary, configuring automatic failover with VMware SRM. When an on-premises failure occurs, VMs are started on EVS, then failed back after recovery. (2) Burst Pattern: Operate on-premises during normal times and expand workloads to EVS during peak periods or capacity shortages. Use HCX vMotion to live-migrate VMs, then move them back on-premises after the peak. (3) Dev/Test Pattern: Keep production on-premises, deploy dev/test environments on-demand on EVS, and delete entire clusters when no longer needed to optimize costs.
DR Site Construction and RTO/RPO Design
When using EVS as a DR site, combine VMware SRM (Site Recovery Manager) with vSphere Replication. vSphere Replication asynchronously replicates on-premises VMs to EVS, with RPO (Recovery Point Objective) configurable from as low as 5 minutes. SRM defines recovery plans and automates VM startup order, IP address remapping, and DNS updates during failover. RTO (Recovery Time Objective) depends on the number of VMs and startup order, but for environments with several dozen VMs, 15-30 minutes is a typical target. Compared to on-premises DR sites, EVS offers the advantage of not requiring hardware procurement in advance and the ability to conduct DR tests (failover drills) without impacting production. If the DR site doesn't need to run continuously, a pilot light configuration (standing by with minimal resources and scaling out during a disaster) can reduce costs.
Enhancing Operations Through AWS Service Integration
Since EVS runs within a VPC, it can integrate directly with AWS managed services. For backup, AWS Backup supports VMs on EVS, enabling policy-based automation of VM-level backup and restore. Backup data is stored in S3 and can be automatically transitioned to Glacier via lifecycle policies to reduce long-term storage costs. For monitoring, EVS host metrics (CPU, memory, storage utilization) can be sent to CloudWatch and monitored alongside vCenter alerts in a unified view. For security, GuardDuty analyzes VPC flow logs to detect suspicious access to the EVS environment, and Security Hub provides centralized security posture management. These integrations enable efficient operations that combine VMware operational tools with AWS managed services. For a systematic understanding of disaster recovery from basics to advanced topics, check out books (Amazon).
EVS Hybrid Operation Costs
In the DR site pattern with EVS, you stand by with a minimum 3-host configuration (approximately $23,544/month) during normal times and add hosts to accept workloads during a disaster. In the burst pattern, you add EVS hosts only during peak times and delete them afterward to convert costs into variable expenses. HCX live migration (vMotion) enables VM movement between on-premises and EVS, allowing flexible workload placement. Compare the annual operating costs of on-premises VMware environments (hardware refresh, data center expenses, personnel costs) with EVS costs to evaluate the economic viability of a hybrid configuration.
Summary - EVS Hybrid Operation Guidelines
Amazon EVS enables diverse hybrid cloud patterns including DR sites, capacity bursting, and on-demand dev/test environment deployment. The pillars of operation are VPC integration with AWS services, automatic failover with SRM, and VM-level protection with AWS Backup. We recommend a step-by-step approach: start with DR site construction, then expand to bursting and dev environment deployment as you gain operational experience.