Running a clusters on scaling groups vs as a dedicated cluster - Amazon FinSpace

Running a clusters on scaling groups vs as a dedicated cluster

The original Amazon FinSpace Managed kdb cluster launch configuration is now referred to as a dedicated cluster. In a dedicated cluster, each node or kdb process in the cluster runs on its own dedicated compute host.

A diagram that shows dedicated cluster.

This configuration provides strong workload isolation between clusters and nodes in a single cluster at the expense of needing a fixed amount of compute per node. In contrast, with a cluster on scaling group a single set of compute is shared by multiple workloads (clusters) running on shared compute, allowing you to share a fixed amount of compute.

A diagram that shows shared compute.

Considerations

  • Currently, a kdb scaling group is limited to only one host residing in one Availability Zone.

  • The HDB clusters running on kdb scaling groups must use dataviews instead of cluster-specific disk cache to store database data for high-performance read access.

  • RDB and General Purpose clusters running on scaling groups must use a kdb volume for their savedown storage configuration.