Replica sets establish a hierarchy among the three nodes handling a data set, with one node the primary handler of the data and two others receiving replicated copies. Should the primary node be lost, then according to previously established criteria, a new primary node is chosen out of the two remaining and a new third copy of data is made. Replica sets also help increase the speed of updates to the database.
MongoDB 1.6 has been released and updated with two new features, auto-sharding and replica sets.
MongoDB 1.6 has been released. This open source, NoSQL system from 10gen has been updated with two new features, auto-sharding and replica sets. Sharding is the ability to subdivide and spread a very large data set across many servers but still continue doing high-speed data updates and retrievals, which speeds up response to queries and expands a databases data handling capabilities. MongoDBs automated sharding will set up a routing map in the database based on a key to each shard so that it knows where data is and can respond to a query or update.