The McNealy and Schwartz Sun legacy
How dot com survivor became a beached whale
Comment Sun's long revivalist saga is over. Soon to be ex-CEO Jonathan 'The Pony Tail' Schwartz will be remembered for what he didn't do and how he turned Sun into a corporate anorexic.
He, chairman and co-founder Scott McNealy and Sun's board have sold Sun to Oracle so Larry Ellison can do what they couldn't or wouldn't do - gut the ramshackle structure they built up and make business sense out of it.
Sun exploded into a $200bn company in the dot com era. It said it was the dot in dot com, but this was hubris.
The company is an extraordinary collection of hardware and software products and technologies that are full of internal contradictions. The biggest one of all was trying to transition the company from being a supplier of quite good margin Unix-based systems, based on its own SPARC chips, to a supplier of open source software, freely downloadable, and industry-standard servers.
This free software/industry standard hardware combination was being used for Open Storage while, at the same time, Sun was selling traditional proprietary storage arrays and its own hybrid server/storage units like the X4500 Thumper. Just what kind of drive array storage vendor did it want to be? You can't carry on selling proprietary storage arrays whilst simultaneously developing and shipping Open Storage 7000 machines that are designed to undercut them.
Sun even went so far as to say, through systems business head John Fowler and others, that Sun would do to storage arrays what industry standard X86 architecture servers had done to proprietary servers. Were they bonkers? Pity the poor Sun sales man having this sort of conversation with customers:
Sun guy: "Mr customer, you need an affordable Sun StorageTek 2530 array which delivers reliable RAID functionality and the highest availability in its class. The flexible, high-density SAS array is easy to deploy and optimized for rack-intensive environments. Try it free."
Customer: "Sounds good. But what about the Open Storage 7000?"
Sun guy: "Good point. Our Sun Storage 7000 Unified Storage System represent a radically simpler and more powerful way to manage enterprise storage at a fraction of the cost of competing systems and delivers the best $/IOP, $/GB and watts/IOP in the enterprise storage market. Based on industry-leading server and operating system technology, Sun opens the closed world of storage systems to innovation and market choice. By delivering industry-standard hardware and releasing a comprehensive suite of storage software to open source via the OpenSolaris Storage community, Sun has broken the final barriers to building state-of-the-art storage systems with truly open architectures. Try it free."
Customer: "You mean it opens the closed world of storage systems, closed storage systems like the StorageTek 2530?"
Sun guy: "Er, exactly!"
You can imagine the same conversation about servers. Yes I know SPARC is an open hardware design and Fujitsu makes them too, but it isn't an industry-standard server and it is, in essence, proprietary. Try buying one from HP, or IBM, or Dell, or whoever.
With quad-core Xeons in 5400 and Nehalem 5500 form, not to mention the quad-core Opterons Sun was using, there was no long-term justification for SPARC to exist. Why was Sun, the espouser, developer and proselytising evangelist for SPARC, also being a vocal Intel cheerleader, dedicated to destroying the proprietary server business?
You can't have it both ways. Either SPARC is not industry-standard, in which case farm the closed base like IBM does with its proprietary chips, or industry-standard chips are the future, in which case join in or exit the show. Don't stand with one foot in the industry standard server camp shooting your mouth off about their virtues, and have the other foot in the SPARC camp blowing froth out of your lips as you shout out its - competing - virtues.
That's Sun for you; it could never decide what kind of company it was. Its loyal customers bought the products while watching bemused as Sun eviscerated its own value propositions and tried to become a Red Hat-like software supplier, that also sold hardware to the grateful hordes of developers and would farm huge service revenues, whilst simultaneously selling SPARC hardware and getting Solaris to run on all of its competitors' servers? Excuse me, is Solaris a Linux competitor? Run that one past me again.
There were also the grandstanding acquisitions: StorageTek for $4bn in 2005 with Sun trying and retrying to put StorageTek somewhere sensible in its organisation. Ultimately it got lumped in with John Fowler's systems group, and execs did a revolving door act as they came in, tried to get a grip of it, and gave up or passed on. Remember David Yen? He's a CEO at Juniper now and I bet he had a smile on his face when he heard about Oracle buying the Sun shambles.
There was MySQL, bought for $1bn in 2008, with more revolving door antics by execs as a result.
The thing is that the top Sun people are simply not business people. They exhibited behaviour which showed they didn't have the amount of grip required to take hold of Sun's wildly disorganised set of products, some of which were impossible to value, and reduce them down to ones that it paid a dollar to produce and sold for $2, with realistic prospects of continuing to do so. That's all they had to do, and they blew it big time.
Okay, you can point to successes but the two turkeys at the top are selling the bag of goods they created to Larry Ellison because they've simply run the company into a cul-de-sac and don't have a reverse gear. Oracle does. HP with Mark Hurd does. IBM with Sam Palmisano does. NetApp with Dan Warmenhoven does. EMC with Joe Tucci does. What is it about Sun co-founder Scott McNealy and Jonathan Schwartz? Schwartz was a management consultant for God's sake. Business school students will be picking their grotesque mistakes apart for years to come.
Sun is like a great beached whale that headed deliberately and bone-headedly for the shore. McNealy and Schwartz took a recoverable business after the dot com boom and ran it so badly they've had to sell it off after years of seeing it essentially go nowhere in share value terms, while other dot com fall out sufferers prospered. Remember the biblical parable of the talents? That's it. That's Sun's recent history all over.
It's about time someone states the obvious. McNealy and Schwartz failed with Sun, failed big time, and should have given up a long time ago. ®
Seems Fair to Me(Not)
I think it is only fair that Oracle buys Sun. Oracle is the one that has been charging main-frame sized licensing fees for Oracle to run on large Unix SMP Servers.
who writes this stuff?
David Yen is "one of the CEOs" at Juniper?
Juniper's just got the one. Yen is a VP there.
re: Matt B.
I would like to be the first to say that Solaris will be stronger two to three years after this purchase. I will not make as bold a statement about SPARC, but I will say that it will have at least a 5-8 year life span (better than you could say for Itanium).
These are my opinions. Not statements of fact. I would not deign to be as arrogant as Matt the Bratt by saying that I know everything.
I'll be a little more Matt (read Troll) like in this statement that this purchase officially kills HP-UX and Itanium. Why would Oracle continue to give laughable multipliers to Itanium now that they hold the keys to some of the fastest single socket CPU's on the planet? Once you put SPARC and Itanium on an even playing field, Itanium falls down very fast and with it HP-UX. Again, Mr Bryant, this is opinion, so don't get your panties in a bunch.