Architecture of an Amazon Redshift cluster

The image above shows the basic architecture of an Amazon Redshift cluster, it is summarized below:

1. The total number of nodes in the redshift cluster is equal to the number of "EC2 instances" used in the cluster.

2. Each slice in a redshift cluster is at least 1 CPU with dedicated memory and storage.

3. The image above shows a cluster with 4 nodes, each one contains 4 slices, the maximum number of partitions per table is 16 partitions.

4. The leader node (Leader Node), is responsible for coordinating lower level nodes, manages external communications and optimizes queries.

5. The lower level nodes, slave nodes (Compute nodes), as mentioned above, each slave node has its own CPU, memory, and disk, depending on the type of EC2 instance selected, this architecture has the ability to "scale out (add more nodes to the cluster)" or "scale up (add more resources to a specific node)".