Low-end and midrange servers could become the building blocks of high-end data centers, thanks to some new R&D effort from the major vendors.
The big cheeses of servers -- including Dell Inc. (Nasdaq: DELL), IBM Corp. (NYSE: IBM), Hewlett-Packard Co. (NYSE: HPQ), Intel Corp. (Nasdaq: INTC), and Sun Microsystems Inc. (Nasdaq: SUNW) -- all say they recognize that customers are starting to use the less expensive servers for higher-end applications. To this end, they're trying to pack more serious features onto the cheapest machines (see Get Set for Server Savings).
An example comes from Sun, which is improving its own low/mid-range servers next summer by updating the Solaris operating system with a feature called Continuous Systems Telemetry Harness [ed. note: sounds painful!]. Although something of a mouthful, the software is currently used to measure voltage, current, temperatures, and other metrics across Sun's higher-end servers.
The software, explains physicist Kenny Gross, of Sun's Physical Sciences Research Center, can warn administrators a few weeks before a server's system board fails.
Gross's team is already working on the next step, which is to scale the technology across many servers at once, using pattern recognition, to help administrators identify troublespots. "Humans can see the anomalies in the [voltage] signals very easily, but the human eyeball doesn't scale," he notes. Enough said.