How Stretched Clusters Maintain High Availability across Multiple Sites

Gartner estimates the average cost of downtime is about $5,600 per minute, which translates to more than $330,000 per hour. Several other studies also put the hourly cost in the six-figure range. Of course, the hard costs are just the beginning. The damage to an organization’s reputation in the eyes of its customers, vendors, employees and investors can be difficult to repair. Depending on the industry, downtime can also create serious safety, security and regulatory compliance issues.

Virtualization helps to improve business continuity by making it possible to move virtual machines (VMs) from one host to another in a high availability (HA) cluster. In a HA cluster, multiple hosts perform like a single system to eliminate downtime. HA clusters are commonly used for load balancing, backup and failover. All hosts in the HA cluster have access to the same storage environments, which allows VMs to move to a different host in the event of failure to provide continuous uptime.

A stretched cluster is a deployment model that adds a geographic diversity component to HA and load balancing. In a stretched cluster, multiple virtualization host servers in different locations are included in the same cluster and act as a single system. VMs can easily be migrated to servers in other geographic locations and still maintain network connections with servers in the stretched cluster. For example, if a server in one location goes down for any reason (weather event, fire, security breach, etc.), critical VMs can be moved to servers in an unaffected location.

VMware vSphere Metro Storage Cluster is an example of a stretched cluster. Introduced in vSphere 5, it enables live migrations between active data centers for proactive load balancing and failover in mission-critical environments.

However, this should not be an organization’s primary disaster recovery strategy. Because VMware vSphere Metro Storage Cluster requires a high-bandwidth, low-latency connection, it’s typically used when active site balancing and extremely high availability are required.

Traditionally, active/active stretch clusters have been complex and expensive to implement. Synchronous replication and stretch cluster solutions require extensive planning, expert services, license fees and external gateways. Setup is costly and time-consuming and ongoing maintenance is difficult.

Pure Storage ActiveCluster in an active/active synchronous replication and stretch cluster solution capable of achieving zero recovery point objective and recovery time objective. Implementation of ActiveCluster with an existing Pure Storage FlashArray configuration is as simple as performing a non-disruptive software update, with four steps, one new command, and no third-party assistance required. Setup takes just minutes with Pure1 Cloud Mediator, while managing ActiveCluster is like managing a single array.

Built into every FlashArray, ActiveCluster is part of the extended collaboration between Pure Storage and VMware. Organizations using solutions from both companies will be able to seamlessly integrate them and achieve always-on operations at a lower cost with simpler deployment. Pure Storage ActiveCluster support with VMware vSphere Metro Storage Cluster creates a highly available virtualized environment for Tier 1, mission-critical applications.

Pure Storage ActiveCluster provides the high availability you need in an active/active metro stretch cluster without the cost and complexity of traditional tools. Let us show you how to avoid downtime by combining Pure Storage and VMware solutions.