Scaling SQL Server
December 10, 2008
It is official. SQL Server does not scale too well. Sure, you can whip about some data, but when it comes to petascale data management, SQL Server is like the kids in the lower quartile of the SAT–good enough to compete but probably not cut out for computational chemistry in the first year of college. Microsoft, according to Eric Lai, has hired, David DeWitt, a database wizard, to fix what Microsofties have not been able to do in the last six or seven years. Microsoft marketing asserts that SQL Server is a super-charged database management system. Anyone who has tried to get this puppy to handle petabytes of XML knows that SQL Server is a wee thing. You must read the ComputerWorld story here about this surprising admission of SQL Server’s weaknesses.
For me, the most interesting comment in this article was this statement:
DeWitt concedes today that [Google’s] MapReduce “does scale pretty well.” He hails its ability to continue queries without interruption if a particular server fails, which most clustered databases cannot do. But he stands by his argument, which is that true relational databases “give you a lot more leverage and good features.” And DeWitt said he will soon release research to back that up.
The article contains quite a few jibes and juicy items about Google, Microsoft, and assorted DBMS gurus. But the fact remains that Google had its MapReduce data management system in place and working pretty well before the firm’s initial public offering in 2004. Since that time, Google has invested heavily in dataspace technology, which in my opinion is the problem Microsoft and Dr. DeWitt have to address–Google’s head start. I think Microsoft’s new database lab is a good idea. But trying to narrow a five or six year gap is a big job. Microsoft may pull even with MapReduce and find that Google has disappeared into dataspace. Will Microsoft be able to leap frog Google? Possibly. But Microsoft has a problem now and may be facing the unpleasant situation of catching up to Google only to find the Google has managed to extend dataspace management lead. Google is in “as is” mode. Microsoft is in “to be” mode. I think there is a significant difference in the two firm’s positions in DBMS.
Stephen Arnold, December 9, 2008
Comments
One Response to “Scaling SQL Server”
-
- Subscribe to Beyond Search
Feature archive
News archive
-
[…] Scaling SQL Server (arnoldit.com) Brak komentarzy, Zostaw komentarz lub trackback […]