Performance tuning and maintaining a Redshift cluster

Tuning an Amazon Redshift database can make a huge difference to the costs and performance of your application. Amazon Redshift has six main areas you should consider when tuning it:

• Massively parallel processing
• Columnar data storage
• Data compression
• Query optimizer
• Result caching
• Compiled code

We will start by looking at massively parallel processing, which we mentioned briefly in the Overview of Amazon Redshift section.

Massively parallel processing

Parallel processing is when your workload is split across multiple strands to allow for faster querying. In theory, it will be quicker for your work to be done by four different strands simultaneously than by one, even with the overhead of merging the individual strands at the end. With most RDBMSs, the parallelism is contained in a single node, so the maximum speed of the query is limited by the overall resource to that one node, and limits...