11 October 2005

MySQL and money

It looks like MySQL’s office is full of elephants.

For those readers who’ve been living under a rock for the last decade or so, MySQL AB is a company which produces a competent SQL database under an interesting dual licence system. MySQL the company have done some innovative things, a few people I highly respect (such as Arjen) work for/with them, and I’d very much like to see a few of these hybrid business models play out successfully. However, they seem to be under attack from both sides of the hybridity.

Elephant-in-the-room #1: on the FOSS side, a bunch of arrogant and greedy unmentionables calling themselves The SCO Group have wooed MySQL into a partnership. This could be a kiss of death all by itself; it has certainly done MySQL no favours in terms of their precious reputation amongst Free Software advocates.

However, what’s worse is that MySQL’s corporate leadership seem to be having trouble keeping their rationale straight when it comes to their reasons for this partnership.

MySQL didn’t need it to support OpenServer and UnixWare users, since they were already doing that. They didn’t need it to get The SCO Group’s ear since TSG are both unteachable and too deeply committed to their legal circus to back out on the say-so of a partner.

The only visible and concrete motivation remaining is increased sales, and if that’s so, it looks like they’ve swapped birthright for pottage here, big time. And of course, if they withdraw from the partnership now D’ohl MacBride’s traditional response (to just about everything) has been to whip out a lawsuit.

Would that be a bad thing? Fiancially, yes. In terms of “karma”, the answer may not be so straightforward. Being the underdog in such a lawsuit would certainy buy back some goodwill from people disgusted at the idea of supping with the devil, but it may be a bad thing from a business perspective: MySQL would have changed their mind, and effectively admitted in public to being fooled, to not having done their due diligence.

Elephant #1.5 came from Planet MySQL, where I curiously ventured after Arjen mentioned a facelift done on it. Matt Asay opines in a piece unrelated to the above that MySQL’s sales success will earn respect for Open Source. In some ways this is true, but the main recipient of the goodwill will be MySQL, and possibly its hybrid business model. That’s good in and of itself, but IMESHO, it’s not what Matt is claiming. Apache has great repect and negligible sales.

Elephant #2, and this is the mastodon of the herd: much of MySQL’s recent extra goodness has come from backing their system with InnoDB, and Oracle have just bought the owners of InnoDB, Innobase OY. Perhaps ironically, a big swag of the sites reporting this are backed by MySQL and InnoDB.

Now what will Oracle do? I think second-guessing them would be, er, risky. No, I’ll be bold and say “downright foolhardy”. Their official policy is business as usual, but the statement is carefully enough worded to leave lots of wiggle room. Either way, MySQL AB now have some fairly substantial reasons to be polite to Oracle. It’s within Oracle’s reach to handicap MySQL-the-database by making future InnoDB terms impractical. It’s probably also within Oracle’s reach to simply buy out MySQL AB, who must feel about like a dolphin does when an orca enters the pool.

It looks like interesting times ahead. Choppy waters. Does MySQL have a replacement for InnoDB in the wings? Are there other FOSS projects which would fit into that slot? Could MySQL write a replacement? Can they simultaneously convince Oracle that they’re small and harmless while convincing their cusotmer base that they’re big and capable? Are there regulatory shields preventing Oracle from simply cutting of MySQL’s air? Will they work? Is there any other way for MySQL to simplify their situation? It all looks like a testimony to the virtue in having a totally flat supply chain and choosing your other partners carefully.

In case you haven’t figured it out, I’m hoping that Arjen or another respected MySQL luminary reads this and will have a large dose of independent yet consistent reassurance up his sleeve, one he’s willing and able to blog about. Much of it has been answered at least in part here and there (e.g. MySQL 5.1 should have completely pluggable storage backends), but I’d like to see one comprehensive summary.

No comments: