Friday, 22 January 2010

2009 Database Tuning Predictions

Astrology. Those in the prediction business have a rare ability to come up year after year each December giving predictions on what is going to happen in the next 12 months.

Astrology is always trying to appear insightful, while being as generic as possible in their predictions, just in case anyone remembers their predictions and try to verify any facts along the way.

But I've found a much better way of making sure that my predictions are right. Welcome to the first next year predictions post that is actually made when the year has finished.

Oracle buys MySQL

Well, not quite. Since Oracle is owned by Sun, Oracle has to buy Sun to make this prediction true. Trillions of bits will flow on the Internet discussing this deal, and at first people will speculate on which part of the Sun business Oracle is really buying. Of course, all that will happen amidst a torrent of statements from Oracle about how much they value and respect each and every piece of Sun business, both hardware and software.

The evil Oracle enemies will argue that Oracle is just acquiring market share but does not have plans for their existing product line beyond merely milking existing Sun customers: either convince them to move to one of the existing Oracle products or abandoning both product and customer altogether.

The Oracle enthusiasts will argue that this acquisition will be the nice round of hardware that Oracle needs to become a vertical and horizontal supplier of all things technology in the enterprise world and that the whole Sun product portfolio will be integrated in Oracle, resulting in a more mighty Oracle.

Much of the controversy will be around MySQL, the product that Sun acquired a couple of years ago and nobody understood quite what they were going to do with it. Seems that Sun's idea of offering a complete LAMP stack was not able to gain much market traction.

With that logic in mind, during 2009 people will invest a lot of time around these questions, but MySQL, both the brand and the market share, will not suffer from that.

Cloud computing

The idea of renting or paying only for the resources you use when you use them and get rid of all the complexities and costs of infrastructure management will start to appear irresistible. Many variants of this idea, implementing it at various different levels, from Google App Engine/Azure to Amazon EC2, will either start or consolidate themselves as credible alternatives.

Except for Amazon EC2, really a modern version of a hosting business, none of those variants will conquer a significant portion of the market. Amazon'ś example however, is the living proof that such a model can work and be profitable for both customers and suppliers.

Interminable debates will be held on all the gotchas of the cloud computing model. Issues with data ownership, HIPPA compliance, local law regulations, service availability. None of those debates will have a clear winner.

Still, cloud computing is far from being a mature industry, so expect to see a lot of different offerings coming up from more places than you expect. Even IBM will offer (in beta, maybe) their own version of a cloud, albeit without a lot of success. Microsoft will also interpret cloud computing emergence  as its chance to counter everyone else and push their own tools and frameworks to generate that much valued lock-in that can keep them afloat for years without having to innovate or compete.

It will not be the year of <xxxxx> or <yyyyy>

(replace XXXXX with your product of choice)
It's time for market analysts to stop talking about "the year of". Listen folks, there was no year of the PDP, no year of the 390, no year of Lotus 1-2-3, no year of MS Office, no year of Windows 2000. There was no single event or point in time that, even looking after the fact, signalled a trend towards any of these products becoming dominant or specially relevant. Ditto for the "killer application" myth. That was only true for VisiCalc 30 years ago, but since then no single application has been the sole reason to cause a disruptive change.

Linux and Mac will continue to grow at Windows expense. Postgres will keep its slow but steady growth. But there will be no "year of" anything.


Looking back it turns out that my predictions probably are not just about what happened 2009 but rather a mix of my own opinions and some bits of prediction for 2010. Who cares, I'm still being a lot more accurate than astrologists.

For the next post, I'm considering to make some 2008 predictions, but those would be probably too far fetched to be even remotely accurate, even if I keep them generic enough.

No comments:

Post a Comment