More retrospection on Sun Microsystems

I just read about a speech that former Sun Microsystems Sales Chief Masood Jabbar gave at the Silicon Valley Historical Museum the other day. You can read it for yourself at: http://www.theregister.co.uk/2009/06/15/jabbar_sun_regret/

So many regrets like many of us former Sun employees but there was something interesting that took me back. Understanding that the Silicon Valley Historical Museum is now located in the former SGI Building 20, Mr. Jabbar waxes nostalgic in recalling how focused Sun was about putting SGI out of business. Here we are in 2009 and both companies are (or have already) disappeared (although it now appears that Rackable which purchased all remaining SGI assets for $25M is going to change its name to SGI). Funny thing about history is that Masood, Scott and the rest of the executive management at Sun may have been focusing on the wrong competitor to put out of business. Now I know that SGI was an initial and very successful competitor and a likely target for Sun but now with 20/20 hindsight keeping SGI viable and successful may have helped Sun far more than they realized at the time. The demise of the focused and successful Unix vendors has not only hurt IT but I think will hurt consumers as well. Those who are familiar with operating systems will say that Unix will always be around and will always have a place in the technology market (and it remains at the core of Linux and Mac OS) but the death of focused Unix vendors will only weaken the case for Unix versus Windows. And, let’s face it, that is the real battle for the hearts and minds...

As a long time Solaris and Irix user I chuckle when I read the numerous threads comparing and contrasting Windows vs. Mac OS. I laugh loudly once a quarter when I see IBM’s earning statement and the amount of revenue they steal from their customers in order to maintain and integrate everything and anything in a datacenter (IBM is the farthest from technological focus as there has ever been in our industry). The remaining vendors maintain a “we don’t care” attitude about recommending operating environments in a datacenter or (as we used to do it) by workload. Windows to so many is the same as any other operating system. While this logic and rationale makes perfect sense for SMB and many mid-market customers, any business above $500M in revenue should really think long and hard about implementing a Windows-only solution in the datacenter. Of course, VMware, recognizing this, makes bundles of cash supporting Linux and Solaris along with Windows in their hypervisor implementation.

The long and the short of this rant is that we seem to have lost a bit of the workload to OS implementation methodology of the late ‘90’s. This relegates the OS to a tactical (or licensing) role versus the strategic role it should play in a datacenter. (Remember the good old days of “-ilities?” As in reliability, scalability, availability... Whatever happened to that marketing? Nowadays I’m not sure anyone understands that just because a vendor claims these capabilities doesn’t mean that their OS can provide them.)

Funniest thing is that I’m sure as we load and grow our datacenters once again (knock on wood that this recession ends sooner rather than later) these very same issues will start to crop up again. (But don’t get me started on how the Cloud Computing promise may simply move the problem from corporate IT out to the service provider...)