News Review: Compaq drops Alpha in favour of Intel.
Compaq drops Alpha in favour of Intel. Appears to be a serious cost-cutting exercise. The end of VMS ?
Compaq have announced that they will, in future, standardise on the Intel Itanium chip, instead of their in-house Alpha chip, for future development of TRU64 UNIX, VMS and Himalaya.
The announcement of 25th June was a surprise to some people. Even people who had NDA-disclosure were told that the Alpha EV-8 series of chips were being developed. Indeed, the announcement two months earlier of a tie-up between Compaq and Oracle strongly suggested that Compaq were attempting to re-position Alpha/TRU64 as a competitor to Sparc/Solaris in the UNIX world (see Compaq License Tru64 Clustering Technology to Oracle. In return for an undisclosed sum, Oracle gain ready-made clustering for Oracle 9i. But does this signal a change of direction for Compaq's Unix Strategy?).
A few days earlier, the Compaq chairman had announced a "180-day turn-around" target. This announcement therefore appears to be primarily a cost-cutting exercise.
Effectively, Compaq have decided to withdraw Alpha from the market within 3 years, and standardise on the Intel Processor Family for all Operating Systems. They will therefore need to port TRU64 and VMS and the Himalaya non-stop architecture to Intel. The TRU64 port is expected within a year. VMS porting may be more difficult, and could take 18 months.
Assumptions, Assumptions
This decision was based on the assumption that future Intel chips will out-perform Alpha within four years.
This is no minor assumption.
At the moment, Compaq have a clear technical lead in 64-bit computing. Neither Sun, IBM, nor HP have really got a viable, fully functional 64-bit Unix solution. So why throw away market advantage? IT is a game of chance. Throw in a good hand in order to win with a better one ? The "good" is the enemy of the "best" ? Maybe.
But I am not convinced. Yes, if Intel do manage the develop a strong 64-bit solution, they are likely to wipe the marketplace. But this is an "IF" - and, in my opinion - it's a big "IF".
Just because Intel made a good 8-bit chip, a very good 16-bit one, and a not-too-bad 32-bit one, does not by any means ensure that they will be able to pull off the 64-bit solution. Look at what Digital Equipment struggled with. There were at least two failed attempts before the Alpha came along. How many millions of $$$s of R&D were wasted before the solution came along. Can Intel do it ? There is at least one observer who is not 100% convinced.
Reactions - mine and others
Terry Shannon, well-known analyst of Digital Equipment and it's subsequent incorporation into Compaq, has stated that this is "good news for Compaq, Intel and IT Customers". However, in my opinion, he does not build a very powerful case for this assertion.
According to an IDC report published before the announcement, Compaq continued to lead the UNIX Divisional Segment with a 34.1 percent share, was number 2 in the Enterprise Server market (15.2 precent revenue share), and in the Department Server market it ranked third at 21.3 percent. IDC commented that "Over the last three years, Compaq has made a determined effort to extend and solidify it's position in the technical server market and has been rewarded with a leadership position.... We believe that Compaq is positioning itself as a technical leader in the industry - that is, the company is actively working to advance the field and gain the support and confidence of the overall technical computing market". Compaq risk throwing away the crown jewels.
Personally, I think this announcement signals the end of VMS, at the very least. To port one Operating System from chipset to chipset is difficult enough (although don't forget the Digital have done it with VMS from VAX to Alpha). But to port three successfully and not lost market share ? I don't think so.
TRU64 has already had problems getting ISV commitment. If I was an Independant Software Vendor, I would not be pouring all my energies into an Operating System that has a dubious future.
Compaq have thrown away a technical certainty, in favour of a market uncertainty. That's a big gamble.
Reviewed by Dennis Adams in July 2001