Reality IT: Lessons in Vendor Hype
Proper product selection is just part of the equation. Learning to see through "vendor talk" is also crucial.
September 24, 2004
Waterfront or Swamp?
A few years back, the industry was abuzz about a new type of low-cost client hardware. Some vendors called them network computers; others diskless workstations. The current term is thin-client devices. These machines were all the rage for several years, and they may have worked out just fine for some organizations. But for ACME, they were a major disappointment--and an object lesson in avoiding vendor hype.
In the late 1990s, ACME's call centers were running old 80486 computers, but we were facing a major upgrade because the machines could not be made Y2K-compliant. We considered buying replacement PCs, but one vendor convinced us we would save money and improve control if we used diskless workstations with redundant central servers for the thin clients. We decided to make the leap to the new technology.
The problems started almost at once. The vendor had back-order problems, so we didn't receive all the workstations we had requested. We had a BIOS problem with the initial units, so the vendor had to help us perform an extra update step on each unit. And the headaches were just beginning.
The vendor had insisted on a massive server configuration with huge disk, processor power and memory requirements. But we found that the servers were way oversized and underutilized. And since the diskless workstations turned out to be only slightly less expensive than increasingly cheaper PCs, the high cost of the servers canceled out the cost advantage of the whole solution.Getting the servers to work in a cluster was painful. We also found that some applications worked better on desktop machines. And we discovered that many desktop-management tools had emerged for the PC environment, negating the management advantages of the diskless workstations.
Each time a problem arose, we brought in the vendor's sales, tech-support and professional-services teams to fix it. We complained, threatened and, when necessary, went over people's heads. Some of the problems were honest mistakes, but in other cases--such as the oversizing of the servers--we felt we had been sold a bill of goods. As this column goes to press, we are finally coming to the end of the workstations' and servers' lifetime, and we plan to replace them with regular computers.
Lessons Learned
Our experience with the diskless workstations, among other technologies, has helped us build a more mature acquisition process. Here are some guidelines we use for buying software, hardware and services:
Use trusted sources to evaluate prospective vendors. Talk to your peers, staff members who have experience with related products/vendors and the vendor's customer references. Read analyst research and magazine reviews, such as those from Network Computing (of course).
Get several bids, but not so many that the analysis becomes too complex. Compare strategic or architectural alternatives, such as outsourcing, in addition to doing an apples-to-apples product comparison.
Try out the products. Software vendors often are willing to provide full, working versions for trial.
Do some due diligence on the vendor, especially if it's not an established player. You can do inexpensive business database checks with Dun & Bradstreet and other services.
The ancient Chinese military philosopher Sun Tzu said, "The more sweat on the practice field, the less blood on the battlefield." We in IT should take that to heart. The more time we put into properly selecting our products, the fewer headaches we'll suffer after they go live.
Hunter Metatek is an enterprise IT director with 15 years' experience in network engineering and management. The events chronicled in this column are based in fact--only the names are fiction. Write to the author at [email protected].0
You May Also Like