Sun and Google: What's the Big Deal?

by Timothy Prickett Morgan

One of the hot stories this week was a vague partnership between formerly high-flying server maker Sun Microsystems and nearby golden child Google, which could just be the two bookends of the dot-com boom, if you think about it for a second. The partnership between Sun and Google doesn't amount to much--at least as far as anyone can tell, at least not yet.

But the deal does put a current IT thought leader and a rival of Microsoft--that would be Google--in the same camp as a company that used to hold the same position and wants to again--that would be Sun.

Sun and Google are right down the road from each other in Silicon Valley, Calif., and they have strong indirect ties. First, they were both spawned out of Stanford University, one of Silicon Valley's two major IT centers (the University of California at Berkeley across the bay is also a big IT force). Larry Page and Sergey Brin are not Stanford graduates, however. They never finished getting their PhDs. Sun founders and former roommates Scott McNealy (its current chairman and CEO) and Vinod Khosla (who is now a powerful venture capitalist), were getting MBA degrees when they ran into Andy Bechtolsheim, a Stanford PhD student who designed the company's first Unix workstation. Bill Joy, the fourth Sun founder, came from Berkeley, where he helped create the Berkeley Systems Design variant of Unix and then commercialized it. When Page and Brin started Google after dropping out of Stanford, Bechtolsheim was very wealthy and gave them a check for $100,000 to get started, which turned out to be a substantial portion of the $1 million they raised when they first passed their hat around to family, friends, and venture capitalists. Both Sun and Google share board members, and Eric Schmidt, who was Sun's chief technology officer after Bechtolsheim left Sun and who helped make Java happen and who had a brief stint as Novell's top executive, ricocheted back to the Valley to become Google's first CEO as Page and Brin realized they needed a professional business manager to deal with the boring details, like taking Google public. Bechtolsheim is back at Sun, and was the designer of Sun's just-announced "Galaxy" line of Opteron-based servers, which have power conservation as one of their defining features.

There is so much hopeful as well as dreadful speculation about what Google will and won't do next out there, particularly as it relates to Microsoft, that it is hard to make a big deal about the relatively few things that Google and Sun said they would do together. In my opinion, this story was cooked up and promoted like crazy for two reasons. First, to give Sun's executives a chance to look cool and hip again, as they did two decades ago when they were founded and one decade ago when they partnered with the former Netscape Communications on Java. And second, to try to steal some public relations thunder from IBM, which was announcing its new Power5+ processor for RISC Unix and Linux servers (which compete against Sun's own Sparc and Opteron iron rather aggressively) on the same day.

What Sun and Google have agreed to do is quite simple, and Google's Schmidt and Sun's McNealy just let the press go wild with speculation about what else the two companies might and might not do. (Hey, this is a sport in the IT media, and it is a lot of fun, I might add.) Brace yourself. First, Sun has distributed some 700 million Java Runtime Environments worldwide and sees about 20 million downloads of the JRE every day, and now that runtime distribution will include a copy of the Google Toolbar, a plug-in for Web browsers that embeds Google search into the browser. Sun is apparently getting paid for this. Second, Google has agreed to help Sun promote the JRE and the open source OpenOffice.org office application suite that is pretty much the only viable alternative to Microsoft's Office, which is a much better collection of programs, by the way. Schmidt made some vague comments about being a Sun server customer already and that it would significantly increase its spending on Sun gear. Google did not offer to distribute or use any Sun software in its vast network of computers. Sun will also buy a bunch of ads on Google's site, too.

For many of you in the IT business, this movie has something of a feel of a flashback. It was Netscape, more than any other company, that lit the wildfire that burned client/server computing to the ground and started the commercialized Internet revolution a decade ago. Netscape, you will remember, was crushed by the full force of Microsoft (which was a tropical storm then compared to the Category 5 hurricane it is in the IT industry today). I can remember championing Netscape and the whole idea of browser-based applications. I can remember playing with Java-based office applications and thinking that they stunk. I recall thinking that this was the right idea, but the wrong time.

I also remember another big deal, which is the flashback part. AOL bought Netscape in November 1998 for $4.2 billion. As part of that deal, Sun Microsystems and AOL agreed to sell and support the Netscape server products, which included a Web server and various other pieces of middleware. Sun paid AOL $500 million for systems and services to run the Netscape products on its own site, and Sun bought $350 million in advertising. That meant that the net money from Sun to AOL was $150 million. Sun and AOL sold the Netscape servers as part of its iPlanet partnership, which Sun took over in October 2001 nearly two years after AOL paid $182 billion in stock to acquire Time Warner. Sun has used the Netscape servers as the basis of its Java Enterprise System middleware suite for Solaris, Linux, HP-UX, and Windows, but Time Warner kept the rights to some of that software and last year sold them off to Red Hat.

If any company on this planet embodies the Sun slogan of "the network is the computer," it is not Sun or its Sun Grid utility. It is Google, which has quite happily built multiple data centers with reportedly well over 100,000 homemade servers (that count is from April 2004) and its own global file system (called Google Filesystem) to create the Google search engine and its Ad Sense ad serving engine. Google's servers don't even have cases, just a motherboard, two disk drives, mounted on a cookie sheet. If anything, Sun could learn a thing or two by visiting a Google data center, which Bechtolsheim undoubtedly has done more than once. If Sun wanted to do something more valuable with its money than buying StorageTek, it should have either tried to buy Google well before it went public and was worth a lot less than the $87 billion market capitalization it has this week, or decided that it wanted to sell 100,000 or more servers really quickly and sold itself for $15 billion to Google.

Sun buying Google would have made sense, but would be financially impossible, because as the AOL-Time Warner deal so aptly demonstrated, you just can't eat anything that is a lot larger than your own head and live. But Google buying Sun? Now there is an interesting idea, and one that the IT press didn't pick up on as they were talking about how Google would build its own alternative, ad-supported, free Internet (I like that idea, until I think about Google knowing every thing I do online) or taking on Microsoft by creating its own operating system and office suite (presumably based on Linux or OpenSolaris and OpenOffice). Sun has some of the smartest engineers on the planet, and it is trying to build the kind of low-powered, grid infrastructures that Google undoubtedly needs. Then again, the obvious question after pondering this possibility is: What would Google need Sun for?

Why, to engineer and mass manufacture the hardware it needs to make its mythical Googlenet and its mythical Google Desktop that plugs into it, of course. Wouldn't it be ironic if Google actually finishes the work that Sun started so long ago? Bechtolsheim took a bunch of spare parts from the compsci department at Stanford and mixed them with some other parts he bought from electronics suppliers to make the first Unix workstation and proclaimed that in the future, the network would be the computer. This is something that Google clearly believes and Microsoft clearly fears.

But then again, Microsoft throws off more cash than any company in history, and Google's measly $3 billion plus in cash is not worth a hill of beans when it comes to fighting against Microsoft, which has three monopolies--Windows, Office, and half for Windows as a server--and a cash pile that is just under $38 billion and growing. Google is going to need a lot more ammo than that, and quite a brain trust, too, particularly if Microsoft buys AOL from Time Warner to beef up its MSN online service and to take away a revenue stream from Google. (AOL is a big customer of Google's.) If Google is going to take a run at the IT market at an oblique angle (as many are rooting for it to do) and is indeed girding its loins to do battle with Microsoft--and if it is not, then the managers at Google are fooling themselves about the threat they pose to Microsoft and its looming and harsh reaction to that threat--then maybe merging with Sun is not a stupid idea after all. It's certainly less stupid than the AOL-Netscape or AOL-Time Warner deals, anyway. But it is also very unlikely that Google would want to do this--particularly when it can make server and software makers jump through hoops to get its business.