The technologies like HBase, Voldermort, CouchDB, and a huge list of others (each designed for a particular scenario) made me believe; that my favorite RDBMS MySQL has grown old and it can’t handle my data anymore.
Know yourself and know your data. Then just filter out what tools don't fit your needs. Don't blame others if you follow the crowds.
Call me a dog, but fast is not even on my list for reasons to choose a NoSQL solution (namely Riak)
1. Operations. I do not want to spend my time doing this, at all.
2. Linear Scalability. Scalability is not about performance for me, it is about being able to predict costs, and to grow capacity without downtime or high cost surprises.
3. Flexibility. The ability to work without having to fight a schema. No schema means no schema migrations (which goes to 1) but also no inherent assumptions in resource allocation. I've got a new idea that works a different way? Great, just add some new buckets and try it. No more ORM impedance mismatches, no dealing with arbitrary SQL vagueness. IF I don't like the mechanism for doing queries I can create a new solution. Map Reduce should be enough for everybody anyway. (joking!)
The technologies like HBase, Voldermort, CouchDB, and a huge list of others (each designed for a particular scenario) made me believe; that my favorite RDBMS MySQL has grown old and it can’t handle my data anymore.
Know yourself and know your data. Then just filter out what tools don't fit your needs. Don't blame others if you follow the crowds.