IA-64 Linux From Red Hat
In May Red Hat, Inc. announced the availability of an early release of
the Red Hat source code for the forthcoming Intel Itanium processor platform
targeting high-end servers and workstations. This distribution enables
the development of IA-64 software for the most intensive computing environments,
from high performance design automation to enterprise level databases.
A free download version of the code can be obtained from Red Hat's web
Hat's release of pre-production tools and software represents an important
milestone in the development of the IA-64 Linux platform," said Michael
Pope, Intel's Director of e-Business Solutions Marketing. "This release
will enable the open source Linux community to accelerate their development
of advanced server and workstation applications for Itanium processor
recently announced the availability online of the Itanium Processer Microarchitecture
Reference, a guide for software developers that details the functional
behavior of Intel's Itanium microprocessor. This guide, together with
the Red Hat code will enable developers to create programs that will be
Red Hat/IA-64 ready before the production release of the Itanium microprocessor.
developers know that as the principle source of the GNU development tools
for both IA-32 and IA-64 Linux (as well as for many proprietary operating
systems), Red Hat is positioned to help developers port and optimize their
applications for IA-64 Linux," said Michael Tiemann, CTO, Red Hat, Inc.
"We believe this will be especially valuable to developers who want to
extend their Linux-based offerings into high-end, 64-bit enterprise and
technical computing markets."
pre-production version will include the Red Hat/Cygnus compiler and tool
kit. It is based on the tools and kernel developed by the IA-64 Linux
Project, formerly known as the Trillian project, a group of companies
that include Red Hat, Intel, Hewlett-Packard, IBM, Caldera, CERN, SuSE,
SGI, TurboLinux and VA Linux. The full release of Red Hat's offering will
coincide with the IA-64 hardware availability from Intel.
Although Microsoft's 64-bit version of NT will probably be customer-ready
before Red Hat's version, we expect the "lag time" on the Red Hat version
to be less than a year - probably nine months. The key will be how many
64-bit applications will be available for Red Hat on Itanium. We suspect
that Windows has a larger lead in that arena, although we do not expect
that Redmond has a ton of apps "ready to roll" whenever 64-bit Windows
Hat already has 64-bit experience on Compaq's Alpha processor, and Microsoft
very publicly killed the Alpha-based version of 64-bit Windows (Ref. TEC
News Analysis), so the edge in technology familiarity goes to Red
Hat in this case, but not by much. It would certainly be nice for Red
Hat if they could get some of the existing Alpha applications easily re-compiled
for Itanium, or at least quickly/easily rewritten, but we expect that
falls into the "wouldn't it be nice" category.
battle at this point is for mindshare: Itanium's performance is not expected
to be in the kick-the-doors-down range, the performance version (McKinley,
or whatever they're calling it these days) is not due out until 2001.
However, since applications often drive OS acceptance, it behooves both
Red Hat and Microsoft to have ISVs (Independent Software Vendors) generating
as many as possible. Red Hat should consider leveraging Compaq's/Digital's
software partners - although Compaq is clearly a friend of Microsoft.
the Linux market, Red Hat's initiative will help it consolidate its hold
on market share leadership. Caldera, SuSE, Corel, and TurboLinux will
need to consider responding in kind, if for no other reason than Itanium
support will become a "check box" when customers consider server OSes.
impact of this announcement on Advanced Micro Devices' forthcoming 64-bit
processor, code named Sledgehammer, is presently unclear. AMD says Sledgehammer
will have full x86 32-bit and 64-bit compatibility (unlike Itanium), so
the need to spend a lot of effort to get code rewritten is not as great.
However, we would be more reassured if Red Hat made a similar announcement
There will be no immediate effect on users, except in the cases where
they want to develop their own 64-bit, Linux-based, Itanium-powered applications.
We expect that percentage is currently small, and not likely to grow significantly.
(Few companies still expend internal resources trying to build their own
software apps from scratch.)
the longer term, users will have more alternatives for 64-bit environments.
More alternatives (within reason) generally provides more benefits to
the customers. Although Wintel will certainly be a presence, the Lintel
option will probably garner significant share, thereby providing a more
competitive playing field for customer dollars.
a technical standpoint, it is too early to recommend which OS to use for
which situation(s). Outside of Windows still ruling the desktop (and Linux,
by implication, not being a major presence there), the server market is
too diverse for a "one OS fits all" pronouncement.