If there's one problem that most companies want to fix about their IT systems, it's the unreliability of their Intel servers. Every year, unscheduled server outages cost companies billions of dollars in reduced employee productivity and lost revenues, not to mention expenditures for getting the systems running again. That's why IBM's recent claims that it can reduce Intel server downtime could gain the attention of corporate executives, including the ones that run your company.

Earlier this year, IBM started claiming that its Intel-based xSeries servers not only are more reliable than those of its top competitors, but also cost less to own and operate because of their reliability advantage. To back up the later claim, the computer giant worked with Gartner Group to develop a new Return on Availability (ROA) software tool that helps companies calculate the cost to their business of unscheduled downtime. The tool, which IBM is providing to its sales force and some Business Partners, lets account executives help customers determine how much an unscheduled server outage can cost them. The tool then lets customers configure and price two Intel servers with different availability features and determine the annual downtime that each configuration will likely experience. These downtime figures are translated into a cost of operations that the tool adds to the sticker price of each configuration.

As you might guess, the ROA tool demonstrates that downtime costs dwarf the cost of the original server. The tool also demonstrates that while IBM's xSeries servers may cost more up front than the competition's servers, their higher availability allows them to cost far less over their operational life. Indeed, by including downtime costs in its calculations of a server's total cost of ownership, IBM can claim that its xSeries servers may cost 30 to 35 percent less than competitors' servers over a five-year period.

Since there's a good chance that you'll encounter IBM's ROA tool in the near future, you may soon be asked by your management team to help them put IBM's availability claims in the proper perspective. If you receive such a request, feel free to tell them that I made the following statements (after all, it always helps to blame a consultant).

• From a hardware and systems management software point of view, IBM's xSeries is an availability leader. Compare any xSeries server to its competition, and you'll likely find availability features that the competition doesn't have. These features include dedicated management processors, real-time diagnostic monitors, and software that performs predictive failure analyses on a broad range of system components. Another feature, Software Rejuvenation, detects conditions that may lead to software failures and schedules the server for a reboot at a convenient time. For more information about the xSeries' availability features, click here.
• Server downtime is just one component of overall downtime. IBM's availability campaign, as well as its ROA tool, focuses on technologies over which IBM has control--the hardware and systems management tools. However, these technologies only cause about 20 to 25 percent of all unscheduled downtime. The remaining outages have their roots in other problems: what release of Windows you're running, how you configured the operating system, what system patches you've applied and how you applied them, your memory utilization rates, and a host of other considerations. IBM's ROA tool does not consider these and many other critical factors of overall reliability.
• The best solutions to downtime are holistic ones. If you buy xSeries servers and use IBM's systems management tools in a thorough and consistent manner, you'll more than likely experience less downtime than you would on a comparable server from another vendor. However, as my observation above implies, you may realize similar or even greater downtime reductions if you migrate from Windows NT 3.5 to Windows 2000 or implement stringent server change management policies. As such, you should consider the xSeries as just one element of a broader initiative to reduce unscheduled downtime.

In short, IBM is packing some incredible technologies into its xSeries servers that make them a "must consider" for IT shops. However, creating a highly reliable Intel server infrastructure is not a goal that you can achieve overnight, nor one that you can realize with a single product. It requires an extended, disciplined focus on the issue that includes identifying and monitoring reliability problems, then devising long-range strategies for addressing those problems. These strategies should include the development of standards and policies for purchasing hardware and software, deploying those products, then managing them using consistent and proven IT practices. This makes the road to reliability a tedious and often highly political process. However, if you stay the course, you can realize the return on availability that IBM is promising with its xSeries servers...and then some.

Lee Kroon is a Senior Industry Analyst for Andrews Consulting Group, a firm that helps mid-sized companies manage business transformation through technology. You can reach him at This email address is being protected from spambots. You need JavaScript enabled to view it..