At SOSP 2007, Amazon presented Dynamo: Amazon’s Highly Available Key-value Store. It’s the thing that “launched a thousand NoSQL databases” – along with Google’s Bigtable, it kicked off the NoSQL movement of the late 2000s and early 2010s. Last time, we looked at Google’s Chubby, which provides a key-value store replicated using Paxos for strong consistency. In many ways, Dynamo is the opposite of Chubby. Dynamo – not to be confused with DynamoDB, which was inspired by Dynamo’s design – is a key-value store designed to be scalable and highly-available.
I think it’s a generally agreed-upon sentiment that papers are Too Damn Hard to read. Which is really very unfortunate, because I think that reading peer-reviewed papers is basically the only way to really learn about a lot of things. I’m not really at a place to be able to change that. But I can try and do the second-best thing: make it easy for others to learn from them. So that’s what I’ll do.