No matter the size of your business, in today’s economic climate, having an IT infrastructure that includes cost-effective disaster recovery (DR) and business continuity (BC) solutions is a necessity. Virtualization now offers an affordable solution for small and midsize businesses (SMBs) to not only save on energy costs and space requirements, but to perform effective disaster recovery.
 
Traditionally, disaster recovery has relied on data backed up and stored on physical formats, such as tape, Blu-ray disc or DVD. Periodic backups are then archived for long-term storage or regulatory compliance. After a disaster, data can be restored to either the repaired hardware or to hardware located within an entirely new datacenter.
 
DR in a virtualized environment provides a range of options for recovering data and performing rollbacks to earlier, damage-free iterations of a primary virtual disk. On the other hand, BC comprises a more active, agile protection mechanism. It enables IT to recover files while simultaneously maintaining application processes with minimal downtime. Then, when disaster recovery is performed due to a malfunction or system failure, the process is generally transparent to users, who can continue to work without interruption.
 
Protecting business critical systems and data is crucial for providing SMBs with a competitive edge.

Data Retrieval 

Generally, once a decision is made to retrieve data, the expense for an SMB is based on the tools implemented. The type of disaster, from small application failures to large data center malfunctions (see sidebar: Disaster Types and Virtualization Solutions), often determines the resources that need to be added to an SMB’s existing set of software applications and hardware. They can include:
 
·         Hardware for prepared host servers or hot site (i.e., computing resource, network infrastructure and storage)
·         Hardware for a cluster (i.e., computing resource, network infrastructure and storage)
·         Cluster licenses
·         Additional software licenses per node
·         Operating System licenses
·         New security solutions across applications and data centers
 
In addition, an SMB should calculate the number of IT resource hours and consulting expenses required for installation, configuration, maintenance and testing of the new solutions, adding these to the overall expense of disaster recovery.
 
Finally, part of any DR test should be the adequate restoration of data. This process in itself can be quite time consuming. For example, with a standard Windows Server, a number of steps are needed, from re-installing Windows, properly mounting tape media revisions and final reconfiguration to thoroughly testing applications and returning them to the production server.
 
The problems related to quick data recovery after a disaster can be compounded by the added pressure on an IT administrator to perform tasks correctly and efficiently while trying to function within a disaster situation.

Virtualization-based Back-ups

In addition to the protection of storage subsystems, the ability to seamlessly rollback changes to a point before a disaster occurred as well as to create dependable backups are two of the most important aspects that virtualization offers.

• With virtualization, disks are abstracted so only a few files are needed for comprehensive backups. A virtual machine can then be recreated using one or two large virtual disk files and meta-data files.

 

• Snapshots of disks made before the installation of patches are useful for displaying where changed blocks have been written. Administrators can then rollback changes to the primary virtual disk, still in its pristine and unchanged condition.

 

• Virtualization, which consists of simultaneously running multiple operating systems on the same server, offers a series of scripts at installation for each guest OS. These scripts have the means to ensure that data on the virtual disk is quiescent, or at rest, which leads to better backups. These tools also allow for file level restore, in case the data file becomes inadvertently damaged.