NoSQL is for the Birds

Benjamin Black has an incredibly insightful article on GigaOM. Normally on posts like these I might try to break them down or extract a few points, however, I feel it's better in this scenario to do two things. One give you a taste of the article and then simply give you the link so you can read it in its entirety itself. Trust me it's worth the read.

Scale breaks everything. Scale even breaks your assumptions about how best to store and query data. Scale does not care about your personal engineering preferences, or about SQL vs. NoSQL. The demands of rapid growth and ever-higher expectations for availability, performance, and cost efficiency force you to re-evaluate and re-imagine what you need, what is possible, and how to best achieve your business goals. This is the context in which non-relational databases like Dynamo, BigTable, Memcache, and Membase were conceived and built. However, even when relational databases are used to build large-scale services, they are unrecognizable as relational. Instead, they look almost exactly like a NoSQL database.

Read more: NoSQL is for the Birds

Cassandra Basics: Indexing

This presentation, by Benjamin Black, discusses indexing with Cassandra. Specifically, Black starts out discussing the differences with indexes and query creation between traditional relational data stores and column stores.

Relational stores are schema driven. Start with your schema and work forward

Column stores are query driven. Start from your queries and work backwards

In addition to these differences he discusses how things change when you are "at scale".

Denormalization is the norm

Everything depends on the indices

Black, proceeds to discuss two types of indices, super column and composite key, and provides some very good examples.

From a deployment perspective, there are two common strategies, that I know of, to deal with your indices. First and most simple is simply deploying them along with your application data in the same keyspace. A second strategy is to deploy a separate physical cluster of machines that handle the indices. Complicates the deployment and the code, but I'm sure there are some good reasons to do so.

Cassandra: Replication and Consistency

In this presentation, by Benjamin Black, we get an introduction to Cassandra's replication and consistency model. Black provides us with a very concise and easily understood definition of both replication and consistency.

Replication. How many copies of the data do we want in the system?

Consistency Level. How many replicas must respond to declare success?

In addition, to providing us with good definitions Black proceeds to discuss a number of topics. Including partitioning, snitching and placement. Essentially, it provides a good high level intro to some very low level concepts of Cassandra.