Guest Column | September 5, 2012

Avoiding A Data Disaster: Part 1

Jim Tessier

By Jim Tessier, Eaton Power Quality Division

Strategies for maintaining business continuity during power outages in virtualized environments

A data center’s top priorities when utility power is disrupted (or when other events jeopardize power availability) are keeping mission critical workloads operational and preventing data corruption. To achieve these goals, data centers have long relied on the combination of uninterruptible power supplies (UPSs) and power protection software to complete a graceful server shutdown. With the rise of virtualization, however, the process brings new significant opportunities to keep mission critical workloads running.

Disaster recovery products provided by virtualization vendors allow administrators to initiate recovery sites manually once alerted of a power crash, but unless data is synced real-time, critical information will be lost upon initiation of the recovery site.

Data centers also have the ability to perform planned migrations, or movement of running virtual machines (VMs) or applications between physical machines at different locations (or cloud sites) without disconnecting the client or application, rather than a disaster recovery. But despite their advantages, no migration products include built-in functionality for responding to power outages. This means when a virtualized data center loses power, technicians must manually initiate the recovery process in a specific sequence, often in the face of intense time pressure, hindering operational efficiency and potentially delaying response time.

By integrating the latest power management solutions directly into the dashboard of their virtualization platforms, companies can simplify their disaster recovery and planned migration processes in order to fully take advantage of server virtualization’s numerous benefits.

In fact, one of the most recent software developments allows virtual machines to be moved between sites (co-located sites, public clouds, hybrid clouds, etc.) based on early notification of power loss to the primary protected site. The software communicates with the UPS to identify the failure before sending an alarm to the virtualization management dashboard. As long as the data center manager provides sufficient battery run time to go through the steps of the recovery plan, the software can then initiate data synchronization and startup of a backup site without data loss.

Additionally, the IT manager can proactively set the prioritization of virtual machines, ensuring that the most critical machines are restarted at the recovery site and the less critical machines can be shut down.

The benefits include:

  1. Starts recovery process on several different events: Software will initiate the execution of a recovery plan upon several different events. Some of the events one would want to initiate a planned migration include utility failure, redundancy loss and virtual power source loss. A standard disaster recovery action would be initiated if a UPS has an internal failure, or over heats.
  2. Less downtime for end users: VMs will be down only for the amount of time required to transfer the latest snapshot and will restart once the transfer is complete. The lesser priority VMs will continue to run on the primary site and drop when the primary site runs out of battery.
  3. Customization for end-users: Users can customize the script included in the power management package to fit their needs. For instance, when users want to configure site recovery capabilities and power management software for low battery and protection loss features, they’re able to make such customizations. This allows users to best align their needs with the integration.
  4. Unattended execution of recovery plan before server crash: The latest solutions provide recovery even before the entire site crashes. With the use of site recovery features, users will have the backup ready even before the crash, which keeps them secured all the time.

Within the span of a few years, server virtualization has progressed from promising new technology to data center mainstay. Virtualization arms IT and facilities managers with potent new tools for meeting service level agreements with fewer physical servers, which in turn results in dramatic savings in energy and footprint. But in order to successfully implement a virtualization project, IT teams need to account for some realities, including solutions for preserving data integrity, business continuity and brand reputation during power outages.

The good news is the latest in power management solutions position companies to take full advantage of server virtualization’s rewards while mitigating its power-related risks. Companies should therefore view advanced power management software as an essential component of doing business within any server virtualization environment.

To learn more about power monitoring and management for your organization’s virtual infrastructure, please visit