Month: April 2009

I’d like to think of myself as an informed prognosticator having worked for Sun for almost 9 years; but this is prognostication nonetheless. I missed my chance to say what I thought of an IBM acquisition but I’ll start by saying – I think I preferred it – it probably would’ve been a better outcome for the things I care about.

The things I care about are the people I know who still work for Sun, the Java ecosystem, and many of Sun’s Open Source projects that I directly or indirectly benefit from – specifically MySQL (this post will live in MySQL), OpenJDK and OpenOffice.

It still seems like a strange merger – sure, Sun and Oracle have a huge shared installed based – Sun Servers + Solaris + BEA + Oracle DBMS was the killer enterprise stack for a decade – that alone gives Oracle a vice-like grip on existing customers; but that’s about the past – not the future. I think someone did the analysis and realized it’s a marginally positive move – so I don’t think this is the big technology vision realized that Oracle are trying to promote.

I’m sure a lot of people at Sun and customers of Sun are glad the uncertainty has come to an end but unfortunately it hasn’t. I’m guessing that it will take until the end of the year before Oracle tells the world what they’re keeping and what they’re dropping. FWIW – here’s my informed guess :

Storage / Servers – the press-briefing made the merger sound like it was all about the hardware (servers and storage). I just don’t see Oracle as a hardware company and they have more to loose than gain by pissing HP off. I wouldn’t be at all surprised if some or all the hardware got sold off to HP, Fujitsu, etc. And I think Oracle could probably recover $3.5-4 bn of their $5bn outlay by doing so.

NetBeans – I’m not changing my mind – it’s done for. Oracle has two Java IDEs already – they don’t need a third. Oracle will pick up some great tools developers to write a migration tool but that’s about it.

OpenOffice – Oracle doesn’t like M$ but I doubt they can find a fiscally rational reason to carry the torch for Open Source and at the end of the day Oracle are doing this because it makes financial sense. I think Oracle will expect “the community” to pick up the slack.

Solaris – I think they’ll milk the legacy installed base (which is huge) but the innovation won’t continue. I think there’s an interesting opportunity for Oracle to manage the migration of the last of the Solaris holdouts to Linux. They could do this by GPLing Solaris and moving some of the Solaris features to Linux; or more likely to an Oracle proprietary OS built on Linux.

Middleware – Oracle have everything Sun has – and Oracle are by and large market leaders with large market and revenue share. Sun have one or two products that might survive and certainly some components. Oracle inherit a commitment to continue to produce the Java EE RI (and others) so Glassfish *might* survive – but Oracle have demonstrated that they don’t have appetite to maintain many products in the same market (see how quickly OC4J got killed after the BEA acquisition ?)

MySQL – I think Oracle will continue to do what Sun did somewhat accidentally – namely slowly kill it off.

Java – I’m confident that Oracle won’t fix the JCP and won’t sacrifice control for the good of the ecosystem. I also think the JCP will become a battlefield for IBM and Oracle and we’ll all be the worse for it.

OpenJDK – no-body needs two JDK’s to maintain – I think Oracle would have to move some of the monitoring / diagnostics from JRockit into a proprietary OpenJDK-based platform. JRockit has a small market share but some nice features for enterprise customers.

Cloud – not that there’s much there there but I don’t expect to see Larry eating his words.

Virtualization – I don’t really understand Oracle’s virtualization strategy other than the feeling that Larry Ellison is unlikely to entertain anything that looks like it might actually save customers money on licenses.