I’ve recently been looking into how exactly nosql document databases work. Not just existing solution APIs like Amazon DynamoDB and Azure CosmosDB, I mean how they work inside, with a view to creating my own for fun. Amazon DynamoDB’s wikipedia page is a good starting place but I wanted more technical detail!

Luckily this talk goes into much more detail than I could find anywhere else, explaining secondary indexes, throttling, burst capacity, accumulating read capacity and adaptive capacity amongst other areas.