Alex Hudson

Thoughts on Technology, Product, & Strategy

Month: April 2009

Want to tell UK Govt. to keep their hands off the ‘net?

apComms is an all-party group interested in various technological issues, and they’ve just announced that they’re starting an enquiry effectively into ‘net neutrality. I would link to something useful if I could, but surprisingly(?) their website is well out of date. Paraphrasing the specific questions they’re asking, though:

  1. When should ISPs be filtering/blocking traffic?
  2. Should Govt. intervene over Phorm-like services?
  3. Do we need new initiatives to protect privacy online?
  4. Is the global approach to kiddie porn working?
  5. Who should pay for traffic, and should Net Neutrality be enshrined in law?

If you want to respond, you need to write not more than four pages and submit it via email to the admin user at the domain – you have until the 22nd May 2009. I’d be interested in people blogging on this topic, particularly their responses.

This is an important topic. Just this Monday there were questions from various well-briefed Tories – including a Tory whip who, and I mean no offence to Mrs Watkinson, I’m pretty sure didn’t know about “throttling bandwidth” before she was given that phrase. It’s pretty clear where the wind is blowing on this issue.

Sun vs. Oracle!

So, the news is out that Sun are being bought by Oracle. Personally, I didn’t see that coming – didn’t see Oracle wanting to get into the hardware business, but maybe they will literally just chop those bits out and sell them off. Or maybe they do want to get into hardware.

This has some interesting implications for free software projects, though:

  1. Java. Clearly Oracle are huge fans of Java and will want to continue the development. Will it stay entirely free software? I would imagine so – I don’t see what there is to gain by closing it up again. They pretty much have control of the development process thanks to Sun’s nature, and it’s not really like anyone is going to be forking it at this point. Probably not much danger here.
  2. Solaris. Oracle’s DB software runs best on this platform, so again it’s likely to be continued – at least, in the short term. Longer term, I don’t see Oracle wanting to commit development resources to both Solaris and Linux, and this could be the key time to start to merge the two. Not necessarily great news for Solaris fans.
  3. Erk. Traditionally, Oracle have never been slow to stick it to Microsoft, so at first glance you could see this going great guns under Oracle as Larry tries to sink another Bill battleship. However, it doesn’t really look to me like this would fit terribly well into the Oracle product line-up, and Oracle have traditionally been a bit luke-warm about OOo – for example, their stuff doesn’t really integrate with it at all, whereas it does with Office, understandably. Indeed, search their blogs for talk of OOo and you find basically nothing – and Oracle aren’t even involved in the OASIS technical committee for ODF, which seems to me to betray a complete lack of interest in this area. OOo is potentially in trouble with this news.
  4. VirtualBox. Not sure much will happen with this; I imagine it would continue but I don’t see Oracle being particularly interested in driving it hard. Probably it would merge with Oracle VM, although the latter is Xen-based. Both could continue with the idea of aligning them in the future, which would probably happen naturally.
  5. MySQL. Erk again. Oracle already own the developers of InnoDB, and in fact the BDB developers too, but don’t expect to see MySQL being in a position to compete with Oracle’s database any time soon now. However, much of the interesting MySQL development now appears to be taking place in the community, so maybe this doesn’t make much difference.

Compared to IBM, on the face of it Oracle doesn’t offer a substantively different story with regards free software. They contribute when and where it makes sense for them, and not in ways which could possibly compete with their software. However, their software is also essentially all extremely high-end enterprise process stuff, which is generally relatively bespoke and requires armies of trained monkeys consultants to install. In that scenario, free software offers a much lower threat and doesn’t even come close to touching many of their markets – so it can perhaps feel a little more relaxed about its contributions. The Oracle OSS page is relatively happy reading,

There is an interesting question outstanding too, I feel: never mind Sun’s wares, what’s going to happen to Unbreakable Linux? It doesn’t make much sense to continue with both that and Solaris long term, which is why I think this actually sounds the death knell for Solaris,

All in all, probably a slightly more favourable situation than under IBM, but the big worry is what happens to It wouldn’t surprise me at all to see StarOffice spun out as a separate company again, but maybe I’m wrong about how brave Oracle might be and how much Larry really loves sticking it to Microsoft.

IBM vs. Sun – spoken too soon?

So, probably as I was posting my little blog piece yesterday on IBM taking over Sun, it seems that the IBM and Sun deal was falling apart – seemingly a quabble over the pricing, but I suspect a little more must have been to it than that.

Again, I’m reminded somewhat of Microsoft – when Yahoo! refused their take-over offer, which at $31 represented an extremely generous premium over their then ticker-price of about 62%, with a total deal worth $44.6 billions. Such a rich deal that even Microsoft would have been forced into debt (though doubtless they’re thanking themselves for walking away now – the timing would have been awful). When we look today, it’s around $13 and has been as low as $9. Shareholders were rightly steaming.

At $9.50 for JAVA stock, IBM would have been paying an almost 90% premium over the sub-$5 price pre-takeover talks. What’s going to happen to Sun in trading today? I suspect the market will punish them, hard – they appear to be the ones walking away from the deal, not IBM, and people will be well aware of that before trading starts later. They’re now going to talk to HP and Cisco about a merger – I’m not sure either of those deals makes sense, particularly Cisco, but HP haven’t long ago swallowed Compaq and are still making sense of that.

In many ways this is Sun all over, but the management will have a lot of explaining to do at this point if another deal doesn’t come together quite quickly. If it falls apart and the stock value starts sinking, we may see IBM come back in a bit later and pick them up even more cheaply…

Sun vs. IBM

If we’re to believe what we’re told in the press, sometime tomorrow – or perhaps later in the week – IBM and Sun will announce some kind of merger. I’m not sure anyone is under any illusion that this would effectively mean the end of Sun in time, being absorbed into IBM, although there is a lot of speculation over what would happen to various projects. Some, like NetBeans, seem pretty certainly done for, and the amount of life left in the SPARC architecture post-merger seems limited.

It doesn’t seem to me that MySQL will be affected much. IBM may or may not continue interest in it – I suggest they would – but Sun appear to have made such a complete balls-up of that acquisition that it doesn’t seem like it would matter any more. Sun’s argument was that they would make MySQL enterprise-ready – but then they released 5.0, which scared the horses and that $1 billion value looks to be rapidly eroding. Similarly, OpenSolaris would probably continue in some half-hearted effort, but again there appears to be a touch of “Well, why bother?” about it.

No-one much seems to have talked about The community side of OOo, in terms of code, is really pretty limited and most interesting stuff happens in Go-OO. Sun are basically OOo, and the ODF TC at OASIS is heavily Sun/IBM. Presumably a merger would weaken their current combined grip on ODF by some small amount, but what of OOo? IBM are currently off in the weeds with Lotus Symphony, which has an outstanding user interface (compared to OOo anyway) but is based on some relatively ancient version of OOo. It would be nice to think that some grand merger of OOo and Symphony would happen, but any such scheme would likely completely alienate what OOo developer community there is because it would take time and be an extremely exclusive process.

I’ve always thought Sun had significant problems with their attitude to free software / open source. They talk about it a lot, it seems like their heart is in the right place, but fundamentally they don’t appear to get it. The difference between “controlling what we do” and “controlling what other people do” appears to be lost on them, and when they giveth with one hand opening up previously proprietary code-bases, they take away with another with licensing restrictions or limitations on contributions. However, IBM are an entirely different beast: they fundamentally do get it. Sadly, though, they get it and decide not to participate: hence Lotus Symphony, hence AIX, hence WebSphere, hence DB2, etc. etc. They contribute widely, but selectively and judiciously. Their participation in stuff like ODF is as much tactical anti-Microsoft activity as anything else.

Based on that approach to free software: very smart, very thoughtful application of development resource to projects which directly benefit IBM reminds me a lot of Microsoft. Fundamentally, their approach is identical – we’ll put effort into a few things where it suits us and we’re not going to be competed with, but everything else we’ll do proprietary thank you. Even where the pain of that proprietary cost is high – e.g., Symphony – they have the resource and the ability to do it.

It will be interesting to see what happens post-IBM. The job cuts are surely going to be the start, rather than the end, and some significant projects will stop – possibly abruptly. Things that will continue will quickly become IBM-branded. In many ways, it will be entirely illustrative of where Sun has gone wrong: IBM will not be putting effort into things which don’t generate return. They’re not going to be wowed by buzzwords about “cloud computing”. IBM are sharp, and the culture clash will be huge.