Slashdot Mirror
IBM Merging with Sequent
Ober was the first
to write in with the news that IBM is merging with Sequent. IBM says it
plans to begin selling Sequent products immediatley after
the merger, and it looks like they even want to improve
some of Sequents NUMA machines. IBM and Sequent are
both involved in Project Monterey.
Sequent mananged to clear out some bottlenecks with their NUMA architecture, but their systems are still bus based. Even though the bus is large, they will not be able to push as much data as say a SGI Origin class machine (with cache coherent NUMA and a Hypercube topology instead of a bus topology)
More information is available on SGI's NUMA page, still more info available here
Disclaimer: I work for SGI, but I believe in everything I say.
I read the internet for the articles.
You're referring to "pooling of interests" accounting. FASB wants to eliminate or severely restrict this and require companies to account for these as purchases. The tech companies are screaming bloody murder about it because under purchase accounting the excess payment over book value would have to be recorded as goodwill and amortized over many years, which would deflate future earnings. Many of these acquisitions would be something like 99% goodwill since the companies are virtually worthless on paper.
I've used Sequents for a large clustered Oracle implementation and as application servers going back to the old Symmetry 2000 line. The idea of Sequent was to build an SMP box based on stock Intel chips. They apparently were one of the first people to solve the SMP problem in Unix with their Dynix/ptx operating system. The 2000 series was horrible junk, IMO. We ended up upgrading to their Symmetry 5000 line (the based on Pentiums) which was much better. The last time I worked with them (about two years ago), we were playing with the NUMA line.
Sequent had two problems in my opinion. First, they were charging an outrageous amount of money for their equipment. Despite claims to use off the shelf components such as the Intel chip, virtually everything they had was proprietary. They even used non-standard SCSI connectors. The machines were supposed to be "mainframe class" and were priced like it. The second problem was their proprietary Unix OS. Sequent was always a niche player and it was problematic getting support tools to run on them. Everybody's first platform is Sun, then HP, IBM, etc. Sequent was way down the list if supported at all. Plus being a minor platform app vendors always tried to blame their problems on OS bugs.
Having said that, the app I worked on is now running on all NUMA's and is doing quite well. We were able to run some serious transaction volume through these machines. Though I'm not sure how they stack up versus the competition today.
This buy out has two interesting ramifications. Everyone reading the press release saw how the sequent boxes 'compliment' the IBM line. But they do directly compete for data center space as well. The sequent boxes may not be long for this world. IBM, historically has proven to go either way on these sorts of things and with sequent constantly lumped in the 'other' piece of the market share pie it may make no sense for IBM to keep the company around.
Also, project monterrey has three members, SCO, Sequent and IBM. Now it has two members. Is SCO in the merger kill zone too? Yes they are. What would project monterrey be if only IBM was the only member.
(flame bait)Not that monterrey will be able to compete with linux and solaris.
Seriously Sequent adds up to almost nothing when compared to any of the Unix big boys this is barely news worthy, only because sequent has shown up in all the project monterrey press releases next to big blue.
Good luck to all those sequent employees. I can hear the resumes being edited from my office.
If you have any insider knowledge of IBM, you know that IBM is rapidly trying to restructure itself to adapt to the changing industries. The Old IBM was too slow and uncreative; in 1994 IBM suffered the greatest corporate loss in history.
Whatever decisions IBM makes at this crucial time are going to be for long term interests. Recently, IBM has announced a lot of support for Linux, has really pushed its RS/6000 archetecture and AIX operating system and has now merged with Sequent, a company whose products compliment the RS/6000. And a little off topic, an IBM executive recently made a testimony against Microsoft regarding OS/2 and Windows, in court.
In the past couple years IBM has partnered with Sun Microsystems and Netscape to try to destroy the Microsoft monoply. The three companies allied not only in technical inovation but also in court.
There is an obvious pattern here. IBM is trying to push away from the industry's dependance on Intel and Microsoft systems. And although it's going to take a long time for this to really happen, any attempts made by IBM will definetly have an impact on the industry, consumers and even Linux.
Dunno 'bout you, but the fact that they bothered to differentiate says worlds to me about the current state of the corporate mindset. As far as things have come for Linux (and open software in general) in the last 18 months or so, every now and then I have to wonder if I really know the half of it.
oops.. further review of the Sequent site (and some delving into the old stuff, I think) led me to find this, which might help some. It seems to look technical, and I'll bet it's what I was looking for... :)
Insert mind here.