Comp150CPA Message 0010.txt from Alva L. Couch(couch)

I neglected to tell you and emphasize that the datastores I mentioned 
are not map/reduce based.  They are all distributed hash tables (DHTs).  
The main hadoop use case is a hadoop job whose results are served via 
a DHT.  DHTs excel at fast return of single objects by key, but do very 
poorly at satisfying regular SQL queries. 

Last changed Thu Mar 10 15:50:32 2011