Sameer Parwani

Today Amazon announced the availability of DynamoDB which is a scaleable NoSQL Database service running on SSD disks in the cloud. It’s highly scalable, highly reliable, and meets the performance level that you request from it (via “provisioned throughput” where you use a dial in their console to select how many reads/writes per second you need).

NoSQL itself is not a game changer. MongoDB has been a pretty awesome NoSQL database and theres also Cassandra and many others. But the largest headache that comes with databases is managing reliability and performance as your database grows. Even when it comes to MongoDB, to get it right you need to first start by considering the right RAID scheme, then create a backup strategy, and eventually implement a replication setup. DynamoDB throws that all out the window. They will manage all the tedious background noise for you. Of course, as a new product we will see if they deliver what they have promised. But AWS generally has a very good reputation and their pre-launch clients have reported good success with DynamoDB.

One concern users may have is whether they will be “stuck” on DynamoDB if they choose to use it. It doesn’t seem like they have too many unique features so you shouldn’t have much trouble. It’s basically a key/value hash so you won’t have much trouble exporting it to the other NoSQL databases if you decide to move on. Of course importing your database from a RDBMS to DynamoDB (or to any NoSQL DB) will require many changes to your application.

Here’s a video explaining more: