Original URL: http://www.theregister.co.uk/2008/01/16/sun_mysql_open_source/

Meet the world's premier open source vendor - Sun

Red Hat, Dell, IBM and HP get SAMPed

By Ashlee Vance

Posted in Servers, 16th January 2008 23:43 GMT

Analysis Almost five years ago, Sun's then CEO Scott McNealy told me his company had little intention of entering the great database fray. "You know, we haven't decided that is a war we want to go fight," he said. "Why not let them all beat each others' brains in?".

By laying down $1bn for MySQL, Sun fattened its arsenal and flashed its brain at Microsoft, Oracle, IBM and Red Hat. Call in the mercenaries. Get ready to start swinging clubs.

It's easy enough to say that Sun overpaid for MySQL. We're told the open source database maker pulled in $50m last year. Some people think the 400-person company may still live in the red. Worst of all, MySQL only makes money off 1 per cent of its customers through services contracts, while the rest eat up the database code for free, according to Trip Chowdhry, a managing director of Global Equities Research.

Judging the value of software companies proves very tough, and judging the value of open source software companies can prove even tougher. With the most popular open source firms, you often find a large gap between a swelling, enthusiastic user/developer base and the much smaller pool of companies actually paying for products or services. How does one put a value on that non-paying enthusiasm and market share, especially in the context of a company such as Sun that has servers, storage, operating systems, middleware and developers to package around a database?

McNealy faced up to that very valuation query in the old interview.

"The real question is how many more database licenses do people really need," he said at the time. "I don't think they need all that much more, and we don't have any revenue model built around that.

"Yeah, I wish we had one. But we do have one - MYSQL, Postgres and the Sun ONE database technology. I think the open source opportunities are non-trivial there."

Rather prophetic on the open source part, no?

Two things of note have changed for Sun since 2003. The first is the arrival of a strong, in-house x86 server business, which has altered some of the company's thinking around that revenue model McNealy discussed. In addition, many of the young companies that once picked Solaris and Oracle as the basis of their technology infrastructure now turn to open source options.

Extra, Extra!

The LAMP stack (Linux, Apache, MySQL, PHP/Perl/Python) ate away at Sun's Unix server sales in the post bubble world. Recently, however, Sun has participated in the LAMP momentum by having a strong x86 server line capable of running the software. As Sun's CEO Jonathan Schwartz pointed out during a conference call today, even those MySQL customers avoiding services contracts need to run the database on something. They also need to buy storage, and they will buy other middleware too.

Sun can sell all of those extras around MySQL today without paying $1bn for the privilege. But, quite frankly, relatively few of the thousands of customers using MySQL are accustomed to thinking of Sun as an x86 option. Now, maybe they will.

Beef service

Like all of the major hardware and software players, Sun wants to see services growth, and it's through services that Sun should enjoy the most obvious short-term gains via the acquisition.

MySQL has suffered from a lack of beef. Large companies hate to bet something as crucial as a database on a small vendor - god forbid a small open source vendor. Now, MySQL appears with the weight of Sun's swollen, according to some, workforce behind it, pimping standalone database services contracts or, even better, deals that stretch across a number of products.

According to Schwartz, this deal should manage to pass without enduring the wrath of Larry Ellison as well.

Sun has enjoyed a very profitable relationship with Oracle over the years. When necessary, the two companies have cut each others' customers special deals. Still, Oracle has partnered with Sun's rivals, and now the database maker will get a better idea about how that feels.

Ultimately, however, Sun bought MySQL to bust into many areas where Oracle doesn't play well. Oracle's databases remain too expensive for webby start-ups or regular businesses looking to get going on the cheap. In the past, these companies often turned to Microsoft. More recently, the most exciting newcomers have turned to open source and MySQL. So Sun receives a piece of fresh action, while taking care of its true enterprise customer base via the fella with the fancy suit and shiny brain.

Did Sun pay too much for MySQL then? Well, yeah - kind of - sort of. That's the nature of the beast, right, especially when you consider that other suitors must have been in play.

The important thing is that Sun has every opportunity to justify the $1bn if it can turn on some of the above businesses cases.

But business is boring. Let's get to war.

The LAMP stack is all the rage, and Sun appears to have found the cheapest way into buying a letter - maybe two.

Sun could well have applied its billions in cash toward grabbing Red Hat. (Let's say $5bn.) Buy why bother?

First off, Sun would need to spend far more to acquire Red Hat. Then, it's battling away with rivals IBM, HP and Dell over the whole proposition, warring with open source zealots afraid of Sun's stewardship and in fighting over Solaris vs. Linux.

As we see it, Sun's only real motivation to buy Red Hat would be to kill it off and pitch the open source Solaris as a replacement, leaving the world with a pair of major server operating systems rather than three.

Canonical stands a cheaper option for Sun if it does want to do the Linux thing, and it may well buy the firm one day.

In the meantime, Sun has busted right into the vibrant gut of the Linux market by grabbing the M. Red Hat can sit on the box. Sun will go ahead and own your data layer though, thank you very much. In addition, Sun gets the attention of thousands of developers, including many of those P folks.

(And, er, take that, Google. Sun will have your development work and loyalty. Many thanks.)

The Open Lock-in

Go big picture, and you find Sun's massive open source claims legitimized. We're talking about the company that funds OpenOffice, that sells the very popular Lustre file system, that will own the most popular open source database and that provides some of the world's most sophisticated open source operating system code. I'm trying to think of a larger, proper open source company and am struggling.

In this context, can those folks out there that hate Solaris because of its proprietary past keep up the anger? Is SAMP any less genuine as an open source stack than LAMP? I don't think so.

(Emotional aside: Dear god, how did Red Hat let this happen? It could have solidified its place as the center of the open source universe and gained permanent leverage over the ever crucial database layer. Instead, it allowed another operating system vendor to grab the database running on most Red Hat servers. Yikes.)

The MySQL buy adds a thick layer of meat to Sun's open source story and must spook Oracle, Microsoft, IBM (DB2 side) and Red Hat.

I wonder, however, how much it spooks HP, Dell and IBM (hardware side). After all, the server realm hosts Sun's real battles.

All of the hardware makers have near equal access to MySQL and MySQL's customers. Now Sun receives the luxury of paying for the development of this software that everyone can use.

But, if you're more likely to buy into MySQL and to buy a support contract now that Sun is behind the company, are you more likely to buy hardware from the company backing the database? I'm just not sure.

Historically, Sun has tried and failed and tried and failed at moving web and application servers on top of its hardware at anywhere near the pace of rivals. In addition, it has been promoting and selling open source databases as an option with its servers for quite a long time.

Given those facts, Sun can only make that $1bn price tag count if it can get past the middleware failures of yesteryear while also improving the database-tied-to-Sun-hardware equation.

Is Sun up to this challenge? It seems to think so.

My, did things just get interesting. ®