Many developers are discovering that traditional relational databases make it hard to scale to the large data volumes and user traffic required by Internet-scale applications. MongoDB is sailing up as one of the leading contenders in the NoSQL space. We at Comoyo are using MongoDB on Amazon Elastic Cloud Compute (EC2) for our payment subsystems, with a target of some hundred millions users regularly accessing the system. As we are handling financial transactions with high demands on reliability, we needed to make sure that MongoDB did not sacrifice our customers' safety on the altar of performance. This talk presents how we use MongoDB to gain higher availability and scalability than traditional databases, with simpler development and administration, without losing the required reliability and durability.
The talk describes how we configured replication, and our approach to work around the lack of transactions in MongoDB. There will be pointers on tuning MongoDB for availability and reliability. Also, the talk will describe how we used MongoDB to implement once-and-only-once messaging semantics on top of Amazon Simple Queuing Service (SQS).