Wednesday, January 30, 2008

The End of an Architectural Era?

Stonebraker et al have an interesting paper pointing out the antiquity and consequent limitations of the classic relational database architecture in today's world of massive disk/cycles/core.

If they are correct (and in spite of the recent MapReduce blunder Stonebraker has made a lot of great calls in the DB world), the world of data management is going to get awfully interesting in the coming years. The DBA's & DA's of this world have been living a relatively comfortable existence compared to those who are wandering across the stormy badlands of the middle tier. But Stonebraker postulates at least 5 radically variant database architectures to address specific use cases of data management. This would seem to lead to a much more complex world for data architects. But maybe a windfall for the consultants who find their niche?

3 comments:

mentaer said...
This comment has been removed by the author.
mentaer said...

wow.. you are right: this Stonebraker has a bunch of interesting papers/works (INGRES, POSTGRES, O-R-DBMS, Stream-Monitoring)

And if more surprising: I have met one of the co-authors N. Tatbul (small world)

Michael A. said...

Stonebraker has been preaching this one for a couple of years now. He has some very good points, but I'm not convinced by every point in the argument.

For example, the thesis, seems to overlook the reason (IMO) why Relational Databases are ubiquitous today. Relational databases are never the "best" solution. They've never been - if the very best performance is required, it has always been possible to develop a custom-made solution that outperforms a database by an order of magnitude. That's one reason, for instance, why my workplace has lots of old flat-file implementations.

What the "traditional" database excels at is providing a general solution that is flexible enough to do a "good enough" job. It is a major advantage, not having to deal with 20 different databases/systems, just because you have 20 different applications. Metal (bigger, faster servers) is always cheaper than meat (people having to maintain the systems).

The interesting part that I found in his talks, is his belief that free databases like PostgreSQL and MySQL will take over the "general" database market, with the "Pay-for-Software" databases (Oracle, DB2, MSSQL) needing to look more toward the specialized markets to compete.