Yesterday I saw a very informative Pycon talk about alternatives to Relational Database systems , which were the de facto choice of database Systems of the past. But now there are problems which demands a whole new approach towards how we used to look at the database systems.
While the traditional databases do make sure that database is in correct state at all times , it comes at the price of slow retreival , insertion and various other scalability limitations. That’s where new Ideas like Document oriented databases , Key value datastores come up.
In short The talk discusses use cases and comparison of
- Distributes Database Systems like Google’s Big table , Facebook’s Cassandra.
- Document Oriented Database Systems like CouchDB.
- Key Value datastores like Tokyo Cabinet.
- Column and Row based Data systems.
- Using Bloom filter and other such interesting stuff.