Joe Martins, researchBy Baselinemag | Posted 2006-10-19 Email Print
How Real-World Numbers Make the Case for SSDs in the Data Center
In the never-ending quest for storage, three experts in the field tell you how to get more for less.director and managing member, Data Mobility Group">
Joe began his career as a systems software engineer at Lockheed Martin, the world's second largest aerospace and defense firm, and later worked on General Dynamics' e-learning system for military and commercial training. He was also a key contributor to the early stage growth of three successful enterprise content/information management companies in the defense (TEC), health-care (InLight) and publishing (OpenPages) industries. Prior to Data Mobility Group, Joe was business development director for research firm Illuminata. Here are his recommendations:
Tip 1. Aggregate access to your information assets to improve accessibility and minimize redundancy.
Billions of dollars are wasted each year because employees and system resources are not aware of the existence of relevant information assets. The result is greatly reduced productivity and increased risk. Storage professionals tend to recommend consolidation in the physical sense as a means to reduce management and maintenance costs.
But that's just a cost-control Band-Aid. While physical consolidation may cut costs, it is not necessarily the right solution for every organization. Some don't need it and others simply cannot afford to do it. And physical consolidation doesn't necessarily improve access and awareness. Fortunately, aggregating access doesn't necessarily require physical consolidation. Several classes of information and storage management applications exist that can help discover, index, classify, search and manage assets within a customer's existing environment. For those that can afford it, the combination of physical consolidation and aggregated access is a win-win.
Tip 2. Foster closer collaboration between I.T. and its internal clients-and let them share responsibilities.
Much of the cost impact on I.T. originates from external forces--namely those who consume I.T. resources--yet I.T. is left holding the bag. I.T. personnel already have full-time jobs, yet some managers pile on additional responsibilities (e.g., information discovery and culling, and retention policy management) that overtax I.T. and set it up for failure. It's time companies leverage the insight and unused capacity of the rest of their workforce.
In 2005, Salary.com estimated that the average employee spends about two hours of every eight-hour workday on non-work-related activities, and it cost companies an estimated $759 billion. If employees have time to surf eBay, then they have time to lend I.T. a helping hand. How about putting your employees to work finding and removing their non-work-related files? Or perhaps implementing an information management program and the necessary tools to enable those users to properly tag and classify their work so that others may find it and act upon it more efficiently (see first suggestion). I can tell you 759 billion reasons why it makes sense.
Tip 3. Strive for interoperability and transparency up and down the application stack.
Storage admins tend to focus on interoperability across the storage fabric. Don't get me wrong, that's an important goal. Unfortunately, it is one that won't be easily resolved anytime soon. But there's another, more pressing concern. Today's storage applications do not communicate well with business applications, and vice versa--leaving the door wide open to inadvertent conflicts and unanticipated risk.
Application awareness and storage awareness are ripe for improvement, but, like storage fabric interoperability, it's more vision that reality today. The implementation of a global file system/namespace is one way to begin the journey toward integrating and aligning storage and business environments--and it's something that can be accomplished today. Yes, it is another flavor of my first suggestion in that a global namespace would improve accessibility, but the suggestions differ in that a global namespace also provides a common framework upon which to gradually implement application interoperability.