New Intel chips look set to clash
Doubts raised over Merced as next-generation McKinley nears release.
Intel plans to pilot the next-generation version of its McKinley Itanium chip this year, raising doubts about the long-term future of its still unreleased debut Itanium chip, Merced. McKinley could be ready for general release as early as the first quarter of next year.
McKinley is Intel's big hope for the server and enterprise range, and analysts say many businesses may now wait for its arrival rather than upgrade to the long-delayed Merced, when it finally ships next quarter.
The chip offers significant advances over Merced, including greater bus bandwidth and an integrated level-three cache, said Paul Otellini, Intel's general manager.
"It represents the convergence of all the major operating systems in the server space. As a result, we have a very strong performing chip," he said, adding that, despite many delays, Merced will ship in the next quarter.
Intel claimed that its Itanium architecture is the most significant development since the 386 processor was introduced in 1985, boasting high-end scalability, reliability and availability.
However, analysts have feared that businesses will see Merced as merely a developmental platform, and won't implement Itanium until McKinley comes to market.
"The question is, will there be a need for Merced or will users wait for McKinley? By the time McKinley comes out, it will be more of a proven solution," said Pia Rieppo, principal analyst at Gartner. "The people on the edge of enterprise computing are most interested in Merced, while the mainstream is waiting for McKinley."
Intel claims to have an 85 per cent server market share with 28 per cent growth year-on-year, but a recent report from researcher IDC said it must implement stronger marketing strategies for its server products to avoid losing market share.
"Intel will face considerable challenges going into 2001," said Thomas Meyer, research manager of IDC's European enterprise server group. "It needs to make a big success of IA64 and must concentrate on the road maps for IA32."