One theorem to rule them all...

Ok so maybe that is a bit of hyperbole on my part. However, the intent of the headline is real. One of the most important theorems that resonates in almost every non-relational data store is Eric Brewer's CAP Theorem.

The theorem states that in highly available shared data systems, you can have only two of the following three properties: Consistency, Availability and Partitioning.

So how does this relate to NoSQL. Well most NoSQL data stores such as Cassandra, CouchDB and Dynamo pick which two their system can support. Take the following excerpt from Cassandra's Architecture Overview wiki page:

Cassandra values Availability and Partitioning tolerance (AP). Tradeoffs between consistency and latency are tunable in Cassandra. You can get strong consistency with Cassandra (with an increased latency). But, you can't get row locking: that is a definite win for HBase.

Note: Hbase values Consistency and Partitioning tolerance (CP)

Anyway, the presentation is interesting and, as mentioned, a fundamental building block of the NoSQL data stores out there.

Read More: Towards Robust Distributed Systems

Dynamo: Amazon’s Highly Available Key-value Store

Amazon has published a whitepaper about it's highly available Key/Value store Dynamo. Dynamo you may recall has been pointed to as part inspiration for Apache's Cassandra data store.

The whitepaper details many important details such as requirements, design considerations, SLAs and system architecture.

Definitely worth a look.

Read more: Dynamo: Amazon’s Highly Available Key-value Store

Understanding the Cassandra Code Base

I've taken a particular interest in the Cassandra data store. Our next author has as well. Ran Tavory, the creator/developer for Hector a Java based client for Cassandra has written a very informative post about understanding the Cassandra Code Base. In his post he discusses the various internals of the Cassandra architecture. It’s a quick but very informative read.

Read More: Understanding Cassandra Code Base