Round 1—General capabilities
IBM won the coin toss and declined to go first. In the first of two major embarassing moments, Microsoft had hit the wrong button the their laptop and we had to wait several minutes for him to recover. My impression of the first round is that Microsoft is slowly re-inventing Unix within their operating system. .Net server is a bit more scriptable than previous versions of Microsoft Windows and now includes remote administration! Woo hoo! (Of course, that could be due to Microsoft wanting to put Citrix out of business)
Also learned a new acronym: SAN: System Area Network.
And how is that different from Local Area Network?
I was impressed that Microsoft has added a versioning file system to .Net server. The presenter deleted his Power Point presentation and was able to restore from two previous versions. Granted, this isn't new: DEC had this in VMS years ago, so no real innovation there (“but of course Microsoft invented ‘Shadow Copies’”).
IBM? I'm sorry, I fell asleep during his presentation.
It was that bad.
IBM goes first. Highlight of IBM's presentation: a distinction between hackers and crackers. Low point: mentioning the r-commands (like rsh, rcp, rlogin etc.). No one in their bloody minds uses those commands anymore. I never used them when I first started using UNIX back in 1990! Sheesh!
Spring mentioned that Microsoft was using buzz words during their presentations, while IBM was just saying how it worked.
But Microsoft was more polished in its presentation, even if it was empty of real content.
Highlight of Microsoft's presentation: “Relative Attack Surfaces” said with a straight face. Amazing.
He also said that .Net server was secure by
Again, with a straight face.
At the end of this round I got to ask a question: What's the time between an exploit that is found and the time the vendor (Microsoft, any particular Linux distribution) will get a patch out? I knew the answer (Microsoft, if they even acknowledge the exploit, will have a patch out maybe a week or two. Linux: hours). I was quite disappointed in the answers. Microsoft hemmed and hawed and never did give a definite answer. IBM didn't quite know how to answer the question and gave a weak answer, more of a guess, of a week turn around time for RedHat.
Round 3—Scalability and Failover
Microsoft goes first. He tried to create a cluster, but the software crashed on him. He seemed to be running .Net server under VMWare but I'm not sure if it was .Net server that crashed, VMWare that crashed, or he just closed the wrong window. In any case, the presentation failed over to IBM.
This was one of the better rounds for IBM. Or I was less familiar with the material. He mentioned IBM's Blue Gene which is a computer with 65,536 CPUs and some 16 terabytes of RAM (which is 16×240 or 17,592,186,044,416 bytes—a typical book takes up about a megabyte, or 1,048,576 bytes, this thing could hold 16,777,216 books in memory!). And he also mentioned Google, which is now up to 15,000 machines, have indexed some 3 billion pages and handles around 150,000,000 search queries (a day? A month? my notes are a bit illegible at this point).
Round 4—System Administration
Dull dull dull dull dull. IBM just read off the slides and Microsoft was still trying to get the clustering to work from the previous round.
Round 5—Is there a point?
Microsoft finally finished setting up the cluster software (from Round 3) only to shut down the wrong server. I must have fallen asleep at this point since I have no notes at all of what IBM talked about.
End of this debacle
This was thankfully the last round of a rather pointless debate—the Microsoft guy kept claiming to be too technical to answer any questions about pricing or licensing or anything (although he did say he didn't like subscription model of RedHat tech support—this from Microsoft? Who is trying to force a subscription model on software?) and apparently the IBM guy was here in an “unofficial” capacity and did not know Linux all that well (he lost the TCO argument to Microsoft! How sad is that?).