For long the high-end server space has been dominated by a handful of players (namely Sun, IBM, HP and Compaq), having proprietary RISC-based hardware coupled with their own OS (usually a proprietary Unix variant).

Intel, the leader in desktop and low-end Windows-based servers, has been trying to get a foothold in this high-end server space, long dominated by the big four. The Itanium (also called IA-64) is the first move in this direction. Co-developed along with HP, the Itanium is the first in its line of high-end processors based on the EPIC (Explicitly Parallel Instruction Computing) architecture. The Itanium is to be followed by McKinley (to be launched in mid-2002), Madison and Deerfield. With the launch of McKinley, the IPF (Itanium Processor family) is bound to play a significant role in the high-end server space. Here are some statements from a recent Gartner report to support this theory.

"IPF will have a significant impact on the high-end Unix server market. By 2006 new Unix servers will be either built on the IBM Power, Sun's Sparc or IPF. Also, the mid-range and high-end Windows server space will be increasingly built on IPF than on IA-32. By 2006, IPF based servers will have a 20 percent market share by revenue."

What does this mean for enterprises?
This raises several questions for enterprises, especially those using servers based on proprietary architecture. Which platform should enterprises place their long-term bets on? Already, three of the four vendors in the proprietary high-end server space have announced (or already launched) servers based on the Itanium architecture. But migrating to Itanium may turn out to be a bumpy ride. It may involve moving to a new OS and porting existing applications onto the new platform to benefit from new IA-64 architecture. It will take some time for Intel to convince enterprises to give up their existing proprietary servers for IPF.

In the short term, say over the next three years, we are bound to see IPF-based servers placed alongside IA-32 and other RISC-based servers. Enterprises are unlikely to dump their existing technology in favor of Itanium. The real key to the success of IPF in the high-end enterprise server space will be the OS and related applications fined-tuned for IA-64.

The low-end
What about the low-end and mid-range server space? Will IA-64 grab IA-32's market share? That seems unlikely. IA-32 processors along with Windows OS are unlikely to be hit soon due to the advent of IA-64. IA-32, with an array of new enhancements like hyperthreading (a technology that makes a single processor look like two processors to the OS), ultra-low power consumption and so on, will continue its dominance in the low-end and mid-range server space. Or at least till Intel decides to phase out these chips and replace them with IA-64 chips.

Sandeep Ajgaonkar, Assistant Editor
sandeepa@networkmagazineindia.com