JetStream DR Solution Architecture: Achieving both Performance and Cost Savings

JetStream Software Senior Solutions Architect Dennis Bray presents the unique design of the JetStream DR platform, capturing data immediately as its written to storage, and replicating it as objects to a container in any standard object store, including Microsoft Azure Blob Storage. This enables the most cost- efficient near-zero RPO solution for DR in the cloud. Dennis shows that everything needed to fully recover a set of protected virtual machines (VMs) is in the object store. If desired data can be continuously replicated from the object store to storage in the recovery environment, for a near-zero Recovery Time Objective (RTO).

In this discussion, Dennis Bray from JetStream Software provides an overview of their architecture and proceeds to demonstrate its functionality. He begins by describing the components and technologies involved in their system. The core architecture includes a management server appliance that orchestrates and operates the software, which can be accessed through a vSphere client plugin or APIs for automation and integration. They also utilize IO filters to capture storage traffic and a DR virtual appliance (DRVAs) for processing and replicating data to an object store, such as an Azure storage account. The protected virtual machines, along with their configurations and status information, are stored in the object store, allowing for recovery in the event of a disaster. Dennis explains two recovery options: a standard failover with a longer recovery time objective (RTO) and a near-zero RTO option. The standard failover involves setting up a recovery site, deploying the software, configuring DRVA, and transferring ownership of the protected domain to the recovery site. Once completed, the failover process can be initiated, enabling the recovery of protected virtual machines. The near-zero RTO option requires a preconfigured and running recovery site, where virtual machines can be quickly recovered using the stored data and configurations from the object store. The discussion also addresses some questions from the audience, clarifying aspects such as the need for a live vCenter environment, the responsibility of customers in preparing the recovery site, and the compatibility of different object storage targets.