Newsletter

TCO Should Include Value as Well as Cost: indirect vs direct costs

Part 2: Disk Management

Disk management is a broad-based term that encompasses quite a few applications and disciplines. Some of the key components of disk management are disaster recovery, data backup and disk organization.

One way to significantly reduce TCO of your storage investment is to optimize your applications so that one action can accomplish multiple tasks. Such is the case with disaster recovery and data backup. A well-devised disaster recovery plan will not only secure your system, but also protect and back up your data. There are two popular backup strategies — disk imaging and file-based backups — that take very different approaches to disaster recovery. We'll look at these issues later. For now, however, let's look at a more global issue: how a company implements its disaster recovery plan.

Plans will vary based on the size of a company. The disaster recovery plan of a home business will differ significantly from the 50-person architectural firm or the four-branch credit union with 100 people. Likewise, the credit union's plan will differ greatly from the disaster recovery plan for a Fortune 500 company. For the small to mid-size business, real money can be saved by developing and implementing a plan that takes advantage of the company's size and flexibility.

It's not enough to have a plan written down in a binder no one reads; you have to practice the plan, making sure that everyone knows what to do and how to do it. Potentially one of the most significant losses an enterprise faces is downtime. If your computing environment fails, chances are you aren't making any money. Here's how one company made sure it wasn't caught off guard.

At Hudson Valley Federal Credit Union in Poughkeepsie, NY, a program is employed that includes a variety of components ranging from working with a business continuity partner like IBM to offsite storage of backup tape with Iron Mountain to internal disaster recovery and disk imaging software from Acronis Inc.

Every year, the credit union's IT staff conducts a full-scale, off-site, disaster test that entails restoring systems that are identified as critical to the institution's operations. These systems include Internet banking services, key file and print servers, Windows domain controller functionality, and, of course, the core processing system. Remote credit union offices connect to the recovery site and then test their ability to perform "normal" transactions.

Testing a disaster recovery plan plays directly into TCO. Should a plan fail because equipment was not tested properly, people didn't know what to do, and the expectations of systems recovery could not be met, that increases the cost of storage ownership.

For example, let's assume that a company had to do a bare-metal restoration of a critical system that was damaged in a flood or fire. Assuming that all of the requisite software to restore the system was kept off site and not damaged in the disaster, what would it take to get the system back up and running? Just finding all of the applications, including all the patches and upgrades, literally could take hours, if not days. (Do you know where all of your installation disks, activation codes and registration numbers are?) Reinstalling all of the software and reconfiguring all of the system and application files also is time consuming. In such a case, the quickest part of the recovery would be restoring individual data files that had been backed up; restoring the system would be a nightmare.

Now let's assume that the IT department made an exact image of the hard disk while the system was working properly, along with nightly incremental images. Restoring the damaged system could now take a matter of minutes. This restoration would include all system and configuration files, application files, data, and anything else that was on the disk.

The time savings alone for restoring an image versus doing a bare-metal installation can be counted in hours if not days. From a financial standpoint, that means that not only is the computer more productive, but so is the IT staff that administers the system and all the employees who use the system.