Good Reading – The End of an Architectural Era

Continuing on with the whitepapers I’ve been (re)reading, another very interesting read is entitled The End of an Architectural Era (It’s Time for a Complete Rewrite) that can be found here and which was written by a whole bunch of people, including (again) Pat Helland from Microsoft.  A very good overview of the article is here and is better than anything I’m going to add now (though that won’t stop me).

I’m sure that many people who are part of the NoSQL movement haven’t read this, but I’m also sure that many people have, as it hits on many of the points that will resonate.  It’s a fairly technical article, but not overwhelming so.  It argues for an elimination of RDBMS in favor of newer technologies that (supposedly) overcome the technical limitations of RDBMS.

If you are a fan of RDBMS, then a lot of the more ‘heated’ rhetoric will probably strike you as a bit over the top, but that shouldn’t stop you from finding a lot of value in it.  The fact that the proposed H-Store implementation achieves an 80x improvement on TPC-C benchmark tests should peak your interest (the fact that there doesn’t seem to be a lot of movement with H-Store itself is a mitigating factor, but one could argue that the flourishing projects that attempt to accomplish the same things makes this irrelevant).  It is interesting that the approach they suggest advocates stored procedures, which goes against a lot of what people seem to think a rejection of RDBMS entails.

RDBMS aren’t going anywhere, for a multitude of reasons.  One could argue that Microsoft’s Azure SQL stuff attempts to have things both ways, but unless and until Microsoft and Oracle introduce systems that embrace NoSQL ideas, they won’t take off in the market.  And the simple fact of the matter is that the vast majority of applications are served well by ‘antiquated’ systems like SQL Server and Oracle.

Nevertheless, anyone with an open (yet still skeptical and critical) mind really should take a look at this.  Certainly, anyone who is interested in CQRS and the ways in which you can implement CQRS-inspired systems that don’t rely (solely or at all) on RDBMS should be interested in the concepts.

Take a look.

posted on Wednesday, April 07, 2010 10:59 PM Print
No comments posted yet.

Post Comment

Title *
Name *
Comment *  
Please add 1 and 1 and type the answer here: