february 14

Addicted to data
In his weekly Computerworld column, Jim Damoulakis, GlassHouse CTO, suggests that we might just be addicted to data.

http://www.computerworld.com/hardwaretopics/storage/story/0,10801,108646,00.html

In his recent State of the Union speech, the president asserted that we're "addicted to oil." At the risk of stepping on a land mine, I'm guessing that this is a statement that most people would agree with regardless of their political persuasion. Well, the same can be said for data. We love all its shapes and forms and want to keep it all forever -- we're addicted!

Unlike oil, the problem isn't a shrinking supply -- quite the opposite. We are smothering in its abundance. Data storage continues double-digit growth rates, and while per-unit costs continue to fall, our appetite outpaces the decline. Even more significant are the ongoing multiyear costs to manage, support and protect the data that consumes this storage.

It isn't the creation of new data that is at the heart of the data management problem, but the mountains of data we retain and never eliminate that is overwhelming our capacity and putting enormous strain on tasks like backup and disaster recovery. Data archiving and expiration policies and processes are missing elements of data management in most organizations. As a consequence, huge quantities of the data sitting on expensive storage systems and consuming thousands of tapes is old, infrequently accessed information and may very well have outlived its value to the organization.

There are at least two reasons for this. First, organizations often lack an authoritative voice to say when a particular piece of data can be expired. Second, the actual process of deleting or even archiving data is usually very difficult. There are often application-specific concerns that make it difficult or impossible to identify and separate the data wheat from the chaff.

Most current data-archiving activities focus on e-mail, a well-understood application with existing archiving tools that can be implemented relatively easily. Other applications are not as straightforward. A basic prerequisite is some degree of application data classification -- a daunting notion when one considers the thousands of applications found in large organizations. But some organizations are finding that by starting with a few select applications and setting practical goals, significant savings can be realized. Don't develop an elaborate classification scheme. Establish a few basics rules, identify where they apply, and quantify the potential cost savings. The approach is to focus on finding the big nuggets rather than the flakes of gold.

Jim Damoulakis is chief technology officer at GlassHouse Technologies Inc., a leading provider of independent storage services. He can be reached at jimd@glasshouse.com.