Taking Storage Risks
Over the last decade, there have been various efforts to pigeonhole organizations into one or more adoption category. Using the labels invented by American scholar Everett Rogers in his "diffusion of innovation" theory, the first 2.5 percent of the bell curve are classified as "innovators." The next 13.5 percent are "early adopters," followed by the "early majority" and "late majority" at 34 percent each, and then the last 16 percent are labeled "laggards."
However, such systems fail when they are turned into broad generalities. A company or individual may be an innovator or early adopter on one front and a laggard on another. Walt Disney, for example, ranges from ultra-conservative to ultra-aggressive.
"If technology helps get a movie out faster, we are aggressive," says Jeff Rochlin, manager of server strategies and architectures at Walt Disney. "But for something like SAP, we will really take our time."
Tom Lindblom, CTO of CKE Restaurants, follows a similar strategy. With 3,300 restaurants under his IT jurisdiction, he has to target his aggressive purchasing at a very narrow band of existing operations. Thus, his company can be termed both an early and a late adopter depending on the specific technology being discussed.
"On the regulatory side, we are very conservative for obvious reasons," says Lindblom, who manages a transactional database of 60 TB and a total of 200 TB of stored data overall. "But when it comes to data analysis and marketing we are very aggressive, as that relates directly to immediate revenue."
When To Take a Risk
Perhaps a better way to look at the question is to choose the best way to meet existing needs. If they can be satisfied utilizing proven technology, it makes sense not to change the winning formula. But what about a situation where the technology needed to race ahead, or maintain competitive advantage, simply isn't available from the established players?
CKE, for instance, is always on the lookout for the newer and brighter ways to crunch its sales database to create new revenue opportunities and enhance profitability. That's why it eats up early-day marketing and data mining tools in its operations.
"The only time we recommend 'bleeding edge' technology is when there is a solid requirement that cannot be met by more traditional technology," says Chip Nickolett, a consultant at systems integrator Comprehensive Solutions of Brookfield, Wisc. "Often this relates to performance or capacity requirements that would otherwise go unmet. In those scenarios, it is much easier to justify the risk of new technology and the cost relative to more traditional offerings."
He explains, however, that most technology fell into this category at one time. SANs, high-capacity disk drives, faster disk drives and such have all had to work through the stage of being an innovator, convincing the early adopters of their value and transitioning into the mainstream.
"As an early adopter, you are blazing new trails and are therefore likely to encounter the growing pains associated with new technology," says Nickolett.
That said, he recommends caution even when purchasing products at the edge of the known universe. For anyone going for leading-edge storage technology, he believes it is important to discover if the new gear is capable of component-level upgrades independent of a major replacement.
"There are likely to be problems or advancements that could potentially be implemented, provided that upgrade path has been planned into the hardware," says Nickolett.
Playing It Safe
While Disney and CKE take out their aggression in certain key areas, Stewart Transaction Solutions of Houston stays conservative no matter what. With a massive data center servicing thousands of clients on the continent, there is no room for error or experimentation.
"Everything needs to be highly reliable in order to maintain our competitive edge, so we don't early adopt anything," says Charles Curtis, enterprise storage team leader at Stewart. "We operate a very controlled environment, as we are publicly traded and that demands high reliability."
Similarly, Slumberland of Little Canada, Minn., remains conservative when it comes to managing a centralized architecture spanning 100 stores in 10 states.
"We don't spend time looking for new things," says Seth Mitchell, infrastructure team manager for Slumberland. "We prefer to invest that time in backup administration and extracting more value from what we already have."
Perhaps not as conservative, though far from an early adopter, the Integrated Waste Management Department (IWMD) of Orange County, Calif., is almost done with a move from direct attached storage (DAS) to SAN.
"We prefer products with the smarts built in, so we tend to spend more on hardware and firmware than software," says Patrick Copeland, senior technical systems specialist with IWMD. "We spend 10 percent of our time finding new technology to implement."
That doesn't mean bleeding-edge purchases are the order of the day, however. The department looks at the latest and greatest, but tends to wait for the market to mature before adopting.
"Don't screw up on due diligence," says Copeland. "Show me where a technology is in use and let me go on site before I buy. Operating in a conservative way translates into huge cost savings."
This approach fits in with research done by TheInfoPro (TIP). Robert Stevenson, managing director of the storage practice at TIP, says that although storage budgets are relatively large, the money isn't going to products at the early stages of the bell curve. In fact, the opposite holds true.
"A lot of CIOs and storage managers are early to talk, late to adopt," says Stevenson.
Let Others Do the Bleeding
In general, most users and analysts agree that there is no need for most companies to expose themselves to high risk when it comes to storage purchasing. Yes, there are exceptions, as mentioned above. But with storage technology having achieved a fair degree of maturity in many zones, why take on something unproven when an acceptable alternative exists that has a couple of years of successful implementation?
"Stored data is too valuable to be bleeding edge," say Mike Karp, senior analyst at Enterprise Management Associates. "Let some competitor's IT shop do the bleeding."
But if you insist on stretching the limits of technology, Karp has a few suggestions to minimize potential exposure.
"Test, test, test before you make the new technology operational," says Karp. "Test it off replicated volumes that really represent your IT environment. And always do a pilot project first."
For more storage features, visit Enterprise Storage Forum Special Reports