Document databases

September 7, 2010 | Comments | Permalink

I have been following MongoDB for over an year. The intention was to use MongoDB to store data related to logging, transmission, reporting etc at Contify, with document database / denormalized data it would be easier to do aggregation, custom views, reporting etc on top of it as compared to RDBMS like Postgres or MySql. Although I never got close to taking it to production ... but did manage to get my hands dirty. At that time I also looked at CouchDB, but preferred MongoDB for the following reasons:

Simple queries
Loved the shell
Collection based (easy to replicate table structures)
Storage / Performance - use of Binary JSON
MongoEngine (Python) that could be easily plugged with Django (my preferred web framework)
Ease of installation and use of JSON

More recently, I have been reading CouchDB and have been really impressed with the following features:

Replication - synchronization of data across multiple CouchDB instances. Enabling offline access of your data
Multi-Version Concurrency Control (MVCC) - Everything is versioned
REST interface - CouchDB as the App layer - You can serve HTML pages right out of your CouchDB
CouchApp - Custom applications to run on top of CouchDB (that are part of CouchDB)
Cloud hosting available at http://www.couch.io/get (at no cost)

As I get my hands dirty with CouchDB, expect to see some apps around data syncing / storage with offline access. I would like to use Couch DB for all my excel related work - be it my issue tracker, task list, number crunching reports etc etc More comparisons:

Great video on CouchDB