Disaster Recovery is Definitely a "Start Small but Think Big" Application when it comes to Virtualization

By: Jerome M Wendt and James Koopmann

Wednesday, December 9, 2009

Even though Gartner Research says that server virtualization is not yet widely implemented (only 16 percent of workloads currently run on virtual machines according to Gartner), Gartner does point to a more virtualized environment in the very near future. It expects that fully 50% of workloads will run inside virtual machines by 2012 and represent nearly 58 million deployed machines. But as this transition from physical to virtual occurs within data centers, traditional disaster recovery (DR) software, procedures and techniques are not positioned to migrate so cleanly into this newly virtualized environment.

To date it has been large companies that have been the quickest to deploy virtualization to reduce server sprawl, data center floor space, and power consumption. However smaller companies (100-199 employees) are now getting the server virtualization message and are expected to rapidly close the gap. Tom Bittman, a vice president and analyst at Gartner who covers servers and storage, says, "By 2010, smaller companies will have a higher penetration of virtual machines deployed than the Global 500."

To get started with server virtualization, he advises small as well as larger companies, to "start small but think big". Companies should first start small with a specific project and then broaden the virtualization plan to encompass more projects, existing systems, and operational processes.

Using this methodology companies can reduce risk while building a foundation that creates a sustainable environment and delivers a targeted reduction in TCO. However, to roll the 'think big' component into the plan requires that a company identify projects where they can safely "start small" but then use their successes in that area as a starting point to expand and improve 'big' data center processes.

Disaster recovery software is a good example of an application that organizations can pick to start. More often than not, each application server has its own tweaked and siloed DR scenario that will encounter issues when migrated to a virtualized environment as it got this condition because each application server had its own dedicated CPU, memory, storage and networked storage resources and maybe even its own system administrator.

Server virtualization changes all of that. The days of having full access to a physical server's CPU, memory, storage, and network resources are giving way to the more controlled virtual environment that share resources in an effort to maximize computing resources.

Because this is occurring, users must let go of tightly coupled DR implementations, services and processes that today define DR for many applications on physical servers so they can move into the shared resources of a virtual environment.
 
This gets dicey. Virtualizing existing applications means that they may now run on shared network, server and storage hardware resources that are different than what they are used to operating on. Further, it is no guarantee that they will all be virtualized on the same server virtualization OS (for example, they could be virtualized on VMware vSphere, Citrix XenServer, or Microsoft Windows Servers 2008 Hyper-V) as an article that appeared this past Monday on SearchServerVirtualization.com confirmed.
 
Performing DR in virtual environments further complicates the situation. DR as it stands now in physical environments is frequently a fragmented mix of software products, including backup software, replication software, high availability software, and failover software. Each has its own recovery processes that call for individual applications to run on their own dedicated physical machine(s).

This mix of DR products coupled with the new challenges that trying to manage DR in a virtualized environment create makes DR a logical candidate as an application that organizations should look to standardize on and simplify in their virtualized environment. However, because physical servers will likely never go away in many environments, organizations should identify software that provides DR for both virtual and physical environments.

To accomplish this, DR software should possess the following characteristics:

  • A central recovery platform that supports both physical and virtual environments.
  • Is not tied to any single virtual OS
  • Supports heterogeneous applications, servers and storage
  • An integrated platform that collects data once but can make it available to multiple different sources for a variety of reasons (recovery, test, development, reporting, data migration, etc.)
  • Can move to a differential model that can then make data available wherever and in whatever form it is needed (locally, remotely, block level, file level, or system level)
  • Handles application or data recovery locally or remotely by leveraging disk-based recovery and recovery automation to reduce risk and improve recovery reliability
  • Supports application-consistent recovery with minimal overhead
  • Meets strict RPO and RTO requirements (in the order of minutes)
InMage is one such software-based solution that meets these complex new requirements that the recovery of physical and virtual environments imposes on organizations without becoming complex to manage itself. InMage includes a collection of next generation data protection technologies such that it delivers a "4 in 1" disaster recovery solution (backup, high availability, fail over, and replication) that recovers data and applications either locally or remotely for both physical and virtual machines.

The increasing penetration of virtual server technology into today's data centers is changing the computing landscape and driving a new set of recovery requirements for all size companies - enterprise as well as small business. Yet a mistake organizations can easily make is to implement server virtualization and just move existing recovery processes into it. If they do, they will quickly discover this is less than an optimal solution as it can generate too much overhead for virtual servers that are pre-configured to run at or near capacity.
 
Organizations need a comprehensive DR solution that runs the gamut - it must support physical and virtual machines; protect both data and applications; provide local and remote recovery and can "start small" while enabling them to "think big". What they will find is a very short list of solutions that satisfy these broad criteria for recovery for both physical and virtual environments and almost none that can do it simply, offer a centralized recovery management platform and can do it cost-effectively. InMage stands out as notable exception to this current DR reality that all size organizations must eventually confront as they go virtual.

Blog Services by DCIGInc.com