Presentation: Tweet"Why distributed databases suck, and what to do about it - regaining consistency"
The CAP theorem has opened our minds to the advantages of sacrificing consistency on the altar of availability. But by abandoning the straight path of transactions, you risk losing the users' trust in the correctness of your data. This is why choosing, knowing and communicating your consistency model is more important than ever.
This talk explains consistency concepts of distributed systems ranging from extreme write availability to 2PC. The goal is, that you walk away with an understanding of how to choose a consistency model, and understand the necessity of different consistency concepts for different business problems. We'll also touch upon how some popular distributed NoSQL databases like Riak, Cassandra, CouchDB/Base and MongoDB support consistency.
Download slides