I wrote a piece about Sun’s short term future based on two pieces of optimism. The first was a third quarter of revenue growth, and a first of profitability for a while, the second was the hope that the systems market would permit competition through differentiation. I said, “At Sun’ we’ve just returned to profitability with our third quarter of revenue growth in a row and as some very famous economist said, three data points are a trend. One of the insights underpinning our strategy is that Sun innovates and monetises intellectual property. We are also one of the last technology companies to own the design and engineering of both CPU and operating system.”
Solaris/SPARC is and will remain a key driver of innovation and competitive advantage in the data centre, because we can design both sides of the coin! To us it is obvious…., but what do others think?
It wasn’t that obvious since it had taken us five years to get there. I pointed at some collateral, all gone now, from Sun and Fujitsu and to the announcement of a engineering agreement with Intel. I said, that one of the best pro-Solaris endorsements occurred last month with the announcement of the Sun Intel alliance. Paul Otellini, president and CEO of Intel says
“We’re thrilled to be working with Sun to make Solaris on Intel Xeon processors a great solution for our enterprise customers worldwide… Bringing together the best technologies from both Sun and Intel will result in innovative products for years to come.”
In the press conference to announce the alliance, Otellini said that they’d like Solaris on Itanium and I wondered if it’d happen.
ooOOOoo
It’s interesting looking back at this post. To some extent I knew it was a trade puff, and that the Solaris and the Labs were not enough. Sometimes the market chooses sub optimal winners, distorted by deep pockets of legacy players. One could argue today that VMware’s virtualisation was a blind alley and thus a mistake and that for capital productivity we should be looking at PAAS not IAAS; the key virtualisation technologies were global file systems and the distributed scheduler, not another CPU multiplexer. Solaris, apart from being all things to all people, couldn’t act as a general purpose O/S, it was never going to get onto a phone. However this gets perilously close to arguing about why Sun failed and I am better now.
Sun got virtualisation wrong partly because of its desire to monetise IP as a systems vendor, rather than as a designer (ARM) or through the production line (Intel); it was destroyed by the operational efficiency of their competitors, much of it enabled by Linux.
Despite this, some of the things Sun did with Solaris, the Fault Management Architecture springs to mind couldn’t have been done without this engineering ambition to develop a computer system.