Problem solve Get help with specific problems with your technologies, process and projects.

Consolidation eases disaster recovery and saves money

See how one company saved money and eased disaster recovery by consolidating and switching to virtualization infrastructure.

High electricity and cooling costs convinced Prudential Fox & Roach Realtors, based in Devon, Pa., to consolidate and virtualize half its server infrastructure and, in the process, simplify its disaster recovery operation.

The real estate firm replaced nearly a hundred HP servers with an IBM BladeCenter system running VMware ESX virtualization software. Web servers, intranet servers, custom real-estate applications, file and print servers and monitoring systems now run on virtual machines (VMs). With substantially fewer physical servers, the company said it expects to save $60,000 a year on electricity and cooling costs.

Prudential Fox & Roach Realtors runs four BladeCenter H chassis in its central data center as well as one BladeCenter S chassis in a separate location to support all of the company's remote offices. This secondary site acts as a DR facility.

Introduction to Virtualization e-book
This article is excerpted from Chapter 1 of the Introduction to Virtualization e-book, which covers the basics of server virtualization technology. Learn about server consolidation, disaster recovery, high availability and more.

The company replaced its Sun Microsystems tape storage equipment with IBM N-Series NAS to securely and centrally manage company assets and client data. Prudential takes snapshots of all its VMs and stores them on the N-Series at the secondary site.

"By bringing backup online and getting rid of tape, we can rebuild machines in minutes instead of days," said Tory Skyers, IT consultant for Prudential. Previously, the firm would have to rebuild the OS manually. It also required that someone in the secondary data center install the backup client, perform the restore from tape or disk and eventually reconnect the new server to the shared storage.

Now, recovery at the DR site is a matter of creating a new VM and redeploying the virtual machine disk format (VMDK) over the network to this machine, while restoring the data from the snapshot. The VMDK file contains the entire OS environment, which facilitates simple loading and saves the VM. Not only does this file contain the image of the OS and application code, it also describes the required VM configuration, including virtual processors, memory and devices.

This serves as a simple and portable single file that contains everything needed to describe the server environment and the actual code and data that make up that server. Launching the VM from a VMDK automatically configures all parameters on the fly.

The new DR strategy was put to the test when a virus brought down one of the firm's servers. A snapshot of the server existed at the secondary site; it was brought back online within two hours. "We can conduct business if one data center goes down without breaking the bank," Skyers said.

The next phase of the DR strategy is to automate the entire process. Currently, the IT team at Prudential creates the new VM, redeploys the VMDK and updates IP and DNS changes manually. But there are tools that can automate those steps. "We still have to sell that upstairs," Skyers said.

Skyers advises IT shops to consider bandwidth between data centers. Without decent bandwidth, it will take a while to get the VMDK from site A to site B. He also suggests replicating only line-of-business data that will cost the company money if it's not running.

Dig Deeper on Green data center: Reducing power consumption with virtualization

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.