Monday
Nov092009
10 NoSQL Systems Reviewed

Jonathan Ellis reviews in the NoSQL Ecosystem the origin of the NoSQL movement and 10 different NoSQL products and how their 1) support for multiple datacenters, 2) the ability to add new machines to a live cluster transparently to the your applications, 3) Data Model, 4) Query API, 5) Persistence Design. The 10 systems reviewed are: Cassandra, CouchDB, HBase, MongoDB, Neo4J, Redis, Riak, Scalaris, Tokyo Cabinet, Voldemort.
A very thorough and thoughtful article on the entire NoSQL space. It's clear from the article that NoSQL is not monolithic, there is a very wide variety of approaches to not being a relational database.
Related Articles
- NOSQL = Not Only SQL?. Google Groups thread on talking about the appropriateness of NoSQL as a label.
- The "NoSQL" Discussion has Nothing to Do With SQL by Michael Stonebraker.
- HBase vs. Cassandra: NoSQL Battle! by Bradford.
- Predictions on the future of NoSQL by Aleksander Kmetec.
Reader Comments (3)
Check out TSnosql on twistedstorage.sourceforge.net - a redis like clone written in Python
I agree with the article in that NoSQL is a bad name. SQL != RDBMS. SQL is just that, Structured Query Language. It _could_ be applied to any db if desired. Amazon's simpleDB is a perfect example of this. It has a query language that is, well, very SQL. In a recent project of mine that was the biggest benefit for me. The trade offs to simpleDB were not an issue for my project.
I would really like to see some of these other column style document DBs put a SQL engine on them. They would be vastly more useful .
I agree with the article in that NoSQL is a bad name. SQL != RDBMS. SQL is just that, Structured Query Language. It _could_ be applied to any db if desired. Amazon's simpleDB is a perfect example of this.