Oracle and Sun: winners and losers

Written By:
Content Copyright © 2009 Bloor. All Rights Reserved.

Sun is a hardware company that has, historically, been rubbish at marketing and selling software. Oracle, on the other hand, is very good at selling software but has no experience as a hardware vendor. This suggests that once the acquisition is complete then the software bits of Sun will be rolled into Oracle while the hardware will be established as a separate division and, possibly, sold off. But for the time being let’s assume Oracle keeps the hardware. Who wins and who loses from a data management perspective?

The first winner will be MySQL users (unless they are Oraclephobes). Who would you rather have developing and supporting your database: a company that has more than 20 years experience in doing just that or one that doesn’t have any? Conversely, the first loser will be IBM’s marketing department. The Information Management division has been fixated for years upon proving that DB2 was outselling Oracle but this has been difficult because Oracle bundles associated revenues into its figures. Nevertheless, IBM could at least make some sort of case – but this won’t be true once the MySQL revenues are rolled in the general database figures. Which makes one wonder: if IBM had known that Oracle would buy Sun for just 10c more per share, would it have felt that was worth it in order to be able to claim that it was the world’s leading database supplier, which it probably would have been with MySQL added to DB2, Informix et al?

The next loser will be Oracle data warehousing. The company has been making a big pitch about its partnership with HP in the jointly developed Database Machine and Exadata Server. But if Oracle now owns Sun won’t it want these to run on Sun hardware rather than HP’s? And if it says it doesn’t, do you believe it? At the very least it will create fear, uncertainty and doubt and put off potential buyers. Which is also, of course, bad news for HP (on the hardware side) but is good news for anyone (Netezza, Kognitio, IBM, Teradata and so forth) who competes with Oracle in the data warehousing space which, naturally, also includes HP—so they win some, they lose some.

While confusion over Oracle’s data warehousing plans will help everybody it will help some more than others, notably ParAccel and Dataupia, whose offerings can be viewed as upgrades to an Oracle environment in a similar (and probably better and, in any case, not just applying to 11g) way to Exadata.

Actually, there is one vendor that may suffer and that is Greenplum. A significant proportion of its larger deals are based on Sun’s Thumper platform and this must be similarly cast in doubt. As an aside, the same also applies to CopperEye, whose software is embedded in Sun’s SDRS solution. Conversely, that will benefit Sensage and other companies focusing on the EU Data Retention Directive, as well as HP (again), because it embeds Sensage’s software in its Compliance Log Warehouse.

Then, there is the issue of Kickfire, whose product was just launched last week (it was pre-announced last year). This is, effectively, MySQL for data warehousing—the sort of performance you would expect from an appliance-based warehouse with a focus on analytics but with the ability to run operational MySQL schemas without change—so you can just load and go. In general I expect Kickfire to do well and this announcement shouldn’t much affect its success in the short to medium term. However, there is the question of what Oracle will do with MySQL’s native capabilities for data warehousing. The company has historically plugged the message that you can have the same database for both operational and analytic purposes, though I don’t agree with them. Assuming they maintain that stance then they won’t do a lot for MySQL in warehousing terms, which means that Kickfire is likely to become the de facto choice for MySQL data warehousing. This, in turn, could make Kickfire a takeover target for Oracle in due course. If that is the case then Kickfire could be the biggest winner of them all.

Finally, let’s consider what happens if Oracle does sell off Sun’s hardware. HP is clearly a possible buyer, which would mean that many of the effects just discussed would become invalid, though there would be cause for concern until this was announced, which would not be some for months. But is HP the only contender? One can imagine a number of other companies being interested, not least Cisco.

Until the future of Sun’s hardware is resolved there will be doubt and doubt plays into the hands of everyone else. Only once that doubt has gone away will we see who the real winners are.