Amazon releases DynamoDB: A NoSQL for Elastic Computing
Werner Vogels, Chief Technology Officer of Amazon, has posted an article on his blog detailing the release of DynamoDB, a NoSQL database solution built from the ground up to scale in the cloud.
This is exciting for a number of reasons, not least of all because it provides a solid NoSQL platform that developers can use for elastic computing. The availability of DynamoDB means that many more developers can start building scaleable apps in the cloud without having to worry about separate database scaling and performance. Amazon have been working with NoSQL and non-relational databases for a long time, and this release represents a major step forward for developers hosting with Amazon.
According to Vogels, DynamoDB is “[A] fully managed NoSQL database service that provides fast performance at any scale.”. So now not only do we have a viable, fully supported NoSQL platform, the database layer will also be automatically scaled and managed, so latency and availability on the data end (traditionally one of the major choke points of any application’s performance) are addressed from the beginning.
Check out the original article here.
If you are planning on working with DynamoDB or have done some work with other database solutions on Amazon, share your experiences with us in the comments.
DynamoDB vs SimpleDB
- DynamoDB saves you time by automatically distributing data so you can scale, rather than forcing you to provision SimpleDB domains of 10GB a piece and then manually partitioning the data.
- DynamoDB indexes only the primary key compared to SimpleDB which indexes all colums. You gain speed but loose some capabilities.
- DynamoDB makes SSD an option
- The pricing is different. DynamoDB pricing is based on allowed reads/writes per second. SimpleDB is priced according to computational expense of performing queries on your servers – measured in Machine hours. Put simply, DynamoDB is likely to be cheaper for most use cases
- DynamoDB has an SLA guarantee, SimpleDB does not because of disk I/O and complex request types