Microsoft NoSQL strategy

Question

Does Microsoft have any NoSQL strategy at all? Any ongoing MS Research projects?

Answer 1

Why would they actively try and destroy one of their key markets? NoSql isn't one size fits all, it works really well in certain scenarios but the very same can be said for SQL which works really well in appropriate scenarios.

To my knowledge Microsoft are actively developing the Velocity project, which is a key value store based cache which could definately be applied in NoSql type scenarios.

Answer 2

IMHO. Microsoft definitely believe in NoSQL for large data storage and its scalability. I think it is beyond just strategy and research now if you look at Azure storage implementation.

See this link for info http://social.msdn.microsoft.com/Forums/en/windowsazuredata/thread/116141cd-6837-46fe-8325-53f16b3d9689

EDITED: Additional discussion about Azure tables (Azure storage) and how it has moved away from traditional relational model

http://social.msdn.microsoft.com/Forums/en-US/windowsazure/thread/b158ac58-fed8-4903-aa29-0930c042c3f7

Answer 3

Azure Tables. Not to be confused with SQL Azure, which is fully relational, 'ordinary' SQL offering.

NoSQL is an umbrella term everybody is rushing to be under, but I would say that originally NoSQL means storage of big data, on commodity hardware, with linear scale-out to +hundreds nodes and partition tolerance (ability to add new nodes on the fly). S3, Cassandra, BigTable. Later the term expanded to cover other non relational models, like schema free document oriented store (MongoDB, which would also qualify as commodity-scale-out), or in-memory processing as Ryak. There are other terms that are associated with NoSQL, like map-reduce processing (Hadoop) but are not necessarily a 'NoSQL database'.

Azure Tables tries to offer a scalable big data store much like S3. Since NoSQL is a such a generic term nowdays, it would be hard to talk about a NoSQL strategy. The terms of the discussion change at such a frantic pace that any announcement that mentions NoSQL would be obsolete by from preparation to release... I would rephrase the question in terms of big data, scale, always-on availability etc etc.