|OW! My FUD - Sun and Mysql
||[Jan. 17th, 2008|02:48 pm]
Sun just bought mysql. I'm not going to say this is only suitable for techies because it has less to do with the technology, and more to do with the bloody politics of it all. And to be honest, I probably shouldn't be writing about the following, but nonetheless I have to say I'm worried.
First off, I love the Sun. I love a lot of their sparc hardware and Solaris, but I don't always like how they operate corporately. Translation: They make some wonderful products, I loved working there, but I don't always agree with higher strategy.
Now, when I say I don't agree with higher strategy, I don't mean how they work on the Open Source front, not right now, because I'll get to that later. Basically Sun has hit lean times since the Dot Com Bubble burst, and they had to get rid of a lot of people. That's fair enough, a lot of companies had to do that, but since then other cutbacks had to happen. Again, fine. But, they made "strategic cutbacks", meaning that they near killed certain sections, leaving some untouched. It's a bit like how the Irish government are screwing over the public sector (YET AGAIN, but that's an entirely different story), the low hanging fruit to cut are the groups that don't look important... like installation ("Hey, we already have an installer, and it's pretty mature, we don't need those guys"). This group turned out some truly marvellous technologies like Wanboot (you can install Solaris on sparc systems as remote as halfway across the world) when when I logged some bugs in the direction of the install/boot teams, it turns out that nobody on the install team is ever free to fix it because they never have the headcount.
So, I disagree with policy there. Now, before I go on, I should probably mention the theory of dieting: Most diets work on 2 stages, starvation to reach your target weight, and then a balanced diet after to maintain it. The balanced diet bit is important, because when you starve your body, it means that it doesn't get everything it needs and has to burn fat reserves to get it, so you eventually have to return to a state where you provide everything it needs to run. Keep that image in your mind, it'll make sense later.
Ok, now let's talk about how Sun does business. There are 5 big organisations in Sun: Software, Systems, Services, Support and Sales. Now, under Jonathon Schwartz's care, Sun is gradually open-sourcing all of their products. This includes products that were previously generating revenue, like Sun's Cluster products - they were a revenue-generator for Sun, and now it's suddenly put out to open source pasture. While the open source hippy in me says yay, it doesn't make business sense to me to suddenly drop something that will earn revenue. At any rate, it seems that the ideal underlying principle to making money from Sun's open source products is to sell the hardware, and also sell services and support around the software products. That's where the Systems group come in. They're the ones who make the money. It's also why you'll see a steady trickle of different platforms emerging, because they can't afford to fall behind on the hardware front. As a result, you'll be seeing some very shiny products from them, and if you can afford them, they're sweet. (Speaking of which, I really want to see a Rock system, they're going to be fun)
Note that I said the ideal. Sun typically makes a profit through headcount reduction and cost restriction; It's not really so much about increasing revenue. So every year, every department's budget is trimmed, and their headcount revised. Very rarely do new positions get created, but some places have a stable head-count budget; that means that if someone leaves, their position is refilled. In worse cases, there's the attrition budget, meaning that if someone leaves, their position goes with them; it's a lot less dramatic than firing a lot of people, but just as effective.
So, every time Sun gets revenue, more heads are cut, on top of the attrition budgets (golden parachutes cost money). The Microsoft settlement paid for layoffs, the Intel deal paid for a few new places and more layoffs. Every deal, more heads will roll out the door. At the same time, Sun will also "buy in revenue", which is what it does when it buys other companies, like StorageTek; the idea is to buy their way into a new market share... it rarely works.
So, back to the diet. The ideal way for Sun to survive is to get to a balanced diet based on the "systems, service and support for free software" method. But until that kicks in (if it does), the investors are appeased by claiming a "profit" based on the crash method of reducing headcount and limiting budgets for groups. Why do I liken it to the starvation part of a diet? Headcount is what generates revenue, they are the people who make, support, sell, and provide services for your products. It puts pressure on the people that remain, and somehow the higher-ups expect to provide the same quality of service when the team's been slashed. That's how the install/boot teams got slashed.
Now, the really funny thing is, have a look at some of the job opening in Sun: The Software group is hiring. That indicates either a stable headcount budget, or better. But the software group aren't the ones that turn profits, and the other organisations that do turn a profit are on an attrition budget.
Here's the thing, they seem to be getting by that way. I'd still go back there. However, spending one third of their revenue from this quarter to bring in MySQL? That doesn't sound so good. Looking at some other analyses, it's really putting forward opensolaris, zfs allowing them to sell more hardware, but I don't want to see MySQL suffer. I don't know if the company who owned it previously were doing well, although the fact that a cool billion changed hands gives me some hints, but it strikes the cynical side of me as being a very weird combination of buying in revenue (which doesn't work) and trying to buy a good open source reputation (which never works). The more trusting side of me is saying that it's just Sun being honest and true to its open source strategy. If it was any other company, I'd see them as wanting to do bad things, but I trust that they won't screw with the fact that it's dual-licenced and being given away. In fact, it fits perfectly well with their higher ethos, so I trust Sun to try and do right by it.
You know, I do have some sympathy for Schwartz. He's facing opposition to open source inside and out. I mean, look at Solaris, there's an absolutely kickass filesystem called ZFS - it's so good that Linus Torvalds has famously said he'd consider switching Linux to GPL3 just so that they can snarf it[*]. People could pay money for that as a storage solution, but they give it away, free. People inside don't see why, wondering why they aren't milking what is a Killer Storage Solution (at least in the case of zfs). And outside, people don't trust Sun, even though they've generally been sharing new shininess from the beginning; For example, look at NFS and RPC - both had implementations and standards handed off so everyone could work with them. Yet, Open Source people don't trust Sun. That said, the OpenOffice fiasco probably had something to do with it.
What I'm ultimately concerned about is that Sun has a habit of doing the starvation thing, and I don't want MySQL to suddenly become subject to this. I'm also concerned that if the serious FOSSers distrust Sun, then the open source contributors to MySQL might fall off, leaving Sun/MySQL to do all the dev-work. At which point, Sun might start putting an attrition budget on the MySQL folk, especially after just blowing a third of their quarterly revenue on them.
I know a lot of companies are applauding it as a smart move, and there's a lot of optimism, but I think the applause is because most of the other companies would have wanted to do the same.
To be honest, I want both of them to do well together, but having been within Sun, I'm pessimistic. But, if it doesn't go well, I'll still be with the optimists in saying, "It really is a pity, there was enormous potential".
Let's see how it goes. To Sun, MySQL and the future.