A strategic model for enterprise storage management: here are the five elements of an integrated storage management plan to help IT become an internal utility dedicated to making information available

Software Magazine, Spring, 2002 by Steve Duplessie

STORAGE MANAGEMENT could be the biggest oxymoron in the entire technology sector. Everyone claims to have a product but no one seems to have anything that works with anyone else's stuff. As users, we seem to have a different management application for every component within our environment--an ad hoc collection of disparate tools with no commonality.

Why do we now need to move to the next level? Simply, we can no longer continue to add a new piece of software every time we add new hardware. Administrators, clearly IT's most critical asset in moving forward, cannot possibly be competent in each and every management element. By attempting to make them do so, we are actually creating a counterproductive environment--if we are attempting to consolidate assets to enable administrators to manage more stuff, complicating the picture with additional software is not logical. Continuing to add software that doesn't integrate together is even more illogical.

I'd like to tell you that after you read this article you'll find the magic bullet, but you won't. Instead, the objective of this article is to educate you to what you should be thinking about in terms of storage management moving forward. Some day all the elements required to effectively manage a chaotic storage environment will be coordinated into a single unified framework, and vendors are furiously attempting to create just that. Instead, we need to understand how the decisions we've made so far will incorporate into our future direction and vice versa. (See Fig. 2, The Building Blocks of Information Availability.)

Helping IT Climb Up the Food Chain

The foundation of the future is predicated on the fact that IT will become an internal utility. If the ultimate aim is complete information availability--anywhere, anytime--then the infrastructure to support that requirement will change. Today, IT organizations are treated as second-class citizens--those guys in the basement who are always late, never prepared, and constantly asking for more money. We don't think that's right, but it's really been the way IT is often viewed. IT has not been able to quantifiably justify its actions--it has been a black art.

In order for enterprises to realize the full asset benefit of IT, IT needs to be able to climb up the food chain and provide true, guaranteed quality of service, secure provisioning, and in a perfect world, automation based on set policies and procedures. When IT can do this, these poor folks won't have to justify incremental funds--the users will. If IT knows exactly what performance and availability levels are required for a specific application under a specific line of business--and can guarantee those levels are met--then the cost associated with providing that service is inarguable by management. Today, since IT can't really guarantee that, the burden is pushed back to the IT side of the fence.

The foundation layer of the information availability model assumes that enterprise storage management (ESM) becomes at least a peer of systems/network and application management. Those frameworks exist today, but ESM does not. In our definition we are concerned with open systems capabilities--Unix and Windows and not so much mainframe. While we love the mainframe, those abilities have existed for years and the players are well entrenched. Open ESM is a green field, where vendors are frantically attempting to bring the level of mainframe management and discipline to the crazy world of distributed systems.

The five core components to an ESM framework include (in no specific order of importance):

STORAGE RESOURCE MANAGEMENT (SRM)--has garnered much attention (but no traction) since Sun acquired pioneer HighGround. It matters because SRM, at a baseline, enables users to ascertain exactly where they are--to visualize what resources are out there, what they are used for, what data is old, tired or replicated, and what users and applications are the causes. Without SRM I don't believe any substantive shop can truly know what stuff is out there. Did you know you were the largest Napster site in Michigan? Did you realize that you were still backing up home directories of employees that haven't worked for us since 1988? Sounds crazy, but such examples are the norm not the exception. We don't believe that you can do any real strategic storage planning until you have solid foundation knowledge of where you are. There are dozens of SRM players--from tiny to huge. Each offers different bells, whistles and operating environments, but what they all have in common is the ability to discover and correlate storage as sets. You can pay as little as $295 and as much as a couple hundred thousand dollars for SRM products. SRM is the one guaranteed payback tool--since almost any vendor will prove their worth before you pay, you can easily put them to the test. We haven't heard of a single customer that couldn't easily and instantly justify the cost of the SRM tool by finding all sorts of new capacity they didn't know they had. This one is a no-brainer.