Probably the first thing to understand about nosql is that there is no standard, there are multiple paradigms to chose from.

Broad differences in storage models:

* Column Oriented Storage (e.g. HBase)

* Document Store (e.g. CouchDB, MongoDB)

* Key-Value (e.g. BerkeleyDB, TokyoCabinet, REDIS)

* Graph DB (Neo4J)

Furthermore each will have their own consistency, replication & availability models, as well as read/write optimisations.

Before choosing one you need to ensure it is adapted to your usage pattern. Here are some links that go into more detail:

A visual comparison, comparing trade-offs of each type of nosql db: http://blog.nahurst.com/visual-guide-to-nosql-systems

Article on appropriate use-cases for each type of nosql: http://blog.heroku.com/archives/2010/7/20/nosql/

NoSQL - the 'definitive guide' (apparently): http://nosql-berlinbuzzwords2010.heroku.com/#1

Further resources: http://nosql-database.org/

And here's a link on data from the top of HN today: http://www.junauza.com/2010/11/free-data-mining-software.htm...