Amazon Redshift is a solution for big data warehousing. Once you’ve created a data warehouse, called a cluster, you can use SQL tools to perform analysis fast. Really fast.
Relational Database and SQL
The data is stored in a relational database and is queried using SQL.
Queries are optimised and enhanced before being distributed among nodes. If you’re familiar with SQL you won’t need to learn anything new.
The interface is based on PostgreSQL and can connect to many BI tools.
Clusters and nodes
When a query is executed, it is distributed among nodes. The nodes get the results in parallel and present a single consolidated result to the leader node.
Distribution keys are used for performing efficient joins between tables. These keys ensure joins are always on the same node, minimising performance losses.
Choosing the right distribution key style will have a huge impact on query performance. Distribution Keys can’t be changed once a table is created.
Data is distributed based on a column you choose. IDs are good candidates for a distribution key as they are often used in WHERE and JOIN clauses.
EVEN aka the Oprah approach:
The rows of a table are distributed in a ’round-robin’ fashion, regardless of the values in the columns. As a result, you will not get good performance when filtering or modifying a large data set.
The entire contents of the table are distributed to each node in the cluster. This is fine for small tables but should be used with caution.
Compressed column storage technology
The data is stored column-wise, as opposed to row-wise, supporting high compression and in-memory operations.
Unlike other database solutions, the query doesn’t have to scan through every row and produces results much faster.
Photo by Bryan Schneider on Pexels