This article can also be found in the Premium Editorial Download "Storage magazine: Storage managers give thumbs up to IP storage."
Download it now to read this article plus other related content.
As we expand our use of storage area networks (SANs), we collide with the expense and slow deployment of SAN technology. What if we could simply purchase storage space, as opposed to the storage itself? With ongoing improvements in long-distance SANs, the notion of remote storage facilities for rent becomes more attractive. So, why have IT managers been slow to adopt this approach?
One way to understand is through this analogy. Imagine Public Storage Inc., a real estate investment trust that rents storage space to the public for a small fee. When you rent this kind of space, you expect certain things: secure storage, isolation from other renters' belongings and a staff that gives you good service.
Data storage renters would want similar things. Until you feel assured that your data is safe, secure and readily available, you're unlikely to consider such a utility service.
Yet doing it all on your own isn't exactly a bed of roses, either. Have you ever had a neighbor whose garage was completely full except for a path to his house's entrance (the primary storage area)? Their habits of keeping everything could continue at work, where they treat network drives like a garage ready to be filled up.
Because you can't really follow your employees home to see who will be likely offenders, you may have to resort to implementing quotas across all applications and users. But the low price of disk drives has caused quite a few shops to back away from quotas. Politically,
Additionally, quotas often imply chargebacks, and this results in an accounting system aimed at penalizing offenders. But how can a service-oriented IT department charge additional money for users surpassing their SLAs when the application owners are constantly threatening to outsource their IT needs? The end result is that you continue to just throw hardware at the problem and keep asking for more money.
However, as with most capacity planning fixes done with hardware and human resources, hardware management often becomes a larger problem than the purchase price.
Renting enables HSM
Hierarchical storage management (HSM)--the ability to store data according to value and retrieval characteristics--has been an elusive power application for quite awhile. However, I've seen varying degrees of success with this solution in the field. By definition, HSM implies a tiered approach to storing data. And by tiered, we mean the farther away the data resides from instant access, the more cost-efficient the solution should be for it to provide the maximum ROI.
Having the technological ability to copy data to safe distances only solves part of the problem. The other part is more directly related to economies of scale. An HSM solution provider should be able to show you how its solution will continue to decrease capital expenditures per unit of storage as the demand for storage resources rise. It doesn't matter if you use tape, disk or both throughout the life cycle of your data, as long as the hardware's limitations fall within your requirements. If the vendor can show you how an increase in your data storage requirements will be met with a product that satisfies the economies of scale criteria, you will be more likely to solve your long-term storage requirements than by just buying more hardware.
The more-hardware approach is equivalent to your pack-rat neighbor paying $1,500 a month for a three-bedroom townhouse, accumulating enough things to fill another three-bedroom townhouse, and then paying $3,000 a month for two townhouses. Wouldn't it be better to use a public storage facility and pay $250 a month to house his duplicated furnishings?
This is the same dynamic plaguing storage facilities. The initial investment and ongoing management commitment is too steep to consider. When it comes to storing data you don't use often, it would be much better to rent space. That's even more true for mergers. Data storage utilities can serve as an excellent temporary workspace to test applications and merge data under the resulting partnership. However, during the merger, applications at the individual companies must continue this process until duplicate departments are unified under the same application.
The same logic applies to divestments, when individual companies of a once-joined corporation decide that it's time to go their separate ways. Often, one party has to go looking for a new place in a hurry. And many times, the new place may not be as accommodating in terms of space as their previous office. In these instances, public storage proves valuable because it can be easily acquired and referenced thereafter, with little cost compared to the original storage location.
This was first published in April 2004