Greenplum Data Hypervisor™

At the heart of Greenplum Chorus is the Greenplum Data Hypervisor-a distributed cloud-scale execution framework that manages complex cross-database state, data movement, and orchestration in a fault-tolerant and self-healing manner.

Greenplum Data Hypervisor is responsible for all "outside the database" activities and state management, and must reliably orchestrate complex, long-running activities such as cross-data center data movement and multistep database instance provisioning and expansion. It must also respond to events and failures anywhere within Chorus with compensating actions and escalation.

Based on these requirements, Greenplum Data Hypervisor was architected using cloud-scale distributed fault-tolerance principles, with an underlying distributed consensus and replication model similar to that used in large Internet services such as Google and Amazon AWS.

As a result, it can handle and recover from failures in mid-operation-even within complex multistep flows. Greenplum Data Hypervisor is scalable to tens of thousands of nodes across geographic boundaries and WAN links, and runs unnoticed within every Chorus server, and elsewhere as needed.