Post by Admin on Mar 14, 2014 18:52:20 GMT
Ø Clone of google’s big table.
Ø Distributed (automatic partitioning)
Ø Column-oriented
Ø Semi-Structured (columns can be added just by inserting)
Ø Built-in versioning.
Ø Not an RDBMS
Ø No joins
Ø No SQL
Ø Data usually not normalized.
Ø Transactions & built in secondary indexed available (as contrib.) but immature.
Ø Need to think differently about how you structure data.
Ø Renormalize your data where necessary.
Ø Structure data & row keys around common access.