According to Framingham, Massachusetts-based research firm IDC, 15 percent of medium-sized businesses have adopted Linux, but among smaller (and, you’d imagine, nimbler) companies, that figure is just 7 percent. Part of the difference is explained by the number of computer servers at these businesses — an average of 11.26 at medium-sized companies, 2.87 at smaller companies — says IDC vice president Roy Boggs.
A company with only a few servers, says Boggs, would generally want to avoid the risk of experimentation since all the servers are needed for sharing the daily load and for backing up one another. But “if I have more than that,” he adds, “then I can start to explore the potential benefits of Linux.”
Supporting a new and untried system is also an issue. Perhaps 85 to 90 percent of midsized businesses have an information-technology staff of their own — a much higher percentage than at smaller companies. “It’s only when you cross 50 employees that the majority of firms have full-time IT professionals,” says Boggs.
Toronto-based Net Integration Technologies is one company attempting to help smaller clients take advantage of the benefits of open-source software without relying so heavily on dedicated IT staffers. Net Integration’s Linux-based operating system, Nitix, uses principles of “autonomic computing” — a movement whose best-known champion is IBM — to create a computer system that’s similar to the involuntary nervous system. Just as the nervous system allows a biological organism to cope automatically with environmental change, external attack, and internal failure, so autonomic computing can help create a system that is self-tuning, self-healing, even self-aware, and that doesn’t need a full-time Linux expert to keep it running.
Nitix “is transparent to a user and requires a whole lot less maintenance and setup and babysitting,” asserts Net Integration chief executive officer Ozzy Papic. “We managed to take the complete Microsoft SBS [Small Business Server] functionality, which requires over 4 gigabytes of code, and delivered the same thing in a little over 25 megabytes” (less than 1 percent of the original size). “That smaller and tighter architecture gave us a fighting chance to make the system autonomic,” says Papic.
“An autonomic agent is not as capable as a human being by any stretch of the imagination,” he acknowledges, “but by making the system less complex, the autonomic engine that’s in charge of the system has less things to worry about, so it can deal with unforeseen problems and keep everything in check.”
However, Bill Weinberg, a technology analyst with Open Source Development Labs in Beaverton, Oregon, offers a word of caution to those willing to sacrifice complexity on the altar of autonomics. “Reality is complex, and a simplistic view of reality results in an underpowered response,” maintains Weinberg. “The IT environment is complex because user requirements are complex because hardware is complex. It’s not an ailment; it’s a description of the landscape.”
In other words, for autonomic computing to ever become truly “self-aware,” it would need to learn to accommodate complexity, not look for ways around it.
Adds Weinberg: “The goal of autonomic computing is to reduce the need for human interaction in the management and configuration of complex systems. It’s not reducing complexity in the system; it’s reducing the user’s need to comprehend it.” From small-company finance executives, no argument there.