After reading many different discussions on MongoDB vs. Cassandra, it seems to me that there are two different use cases for these products. MongoDB feels like a good fit when database accesses are done in a document-centric manner with little need for transactions and concurrency management, whereas Cassandra is a better fit for high-write, transactional, or element-based updates to data.<p>What are your thoughts?
> MongoDB vs. Cassandra<p>None. We use OrientDB <a href="http://www.orientdb.org/" rel="nofollow">http://www.orientdb.org/</a> (it supports more than one "storage" mode: <a href="http://nosql-database.org/" rel="nofollow">http://nosql-database.org/</a> )<p>On the plus side, it's Java based (without native stuff), so it can run existing production machines where mongo can't be compiled (Sparc, HP, etc).<p>Another nice advantage is the support for SQL syntax, so for existing teams, the DB people won't feel left out :).
First, java is not a plus IMHO. Any time I have used anything java based, it's been quite slow. You also have to deal with all the typical java issues like memory and heap space, but that is another diacussion all together.
NoSQL is just that, no SQL. If you want an SQL database, use one. The creators of the nosql db's did so for a reason, I don't see the need to back port it in. Sorry for the rant.