The Zerto architecture for disaster recovery is designed as a scale-out solution that integrates directly into the hypervisor layer. The primary management component is the Zerto Virtual Manager (ZVM), which must be installed at each site (production and recovery) to manage the local resources and coordinate with its peer across the network. Data movement is handled by the Virtual Replication Appliance (VRA), a lightweight virtual machine installed on every hypervisor host where protected VMs reside.

When implementing Extended Journal Copy (formerly known as Long-Term Retention), Zerto leverages its unique Continuous Data Protection (CDP) stream. In a typical disaster recovery scenario, writes are captured at the production site and replicated asynchronously to the DR site. These writes are stored in the DR site journal, which provides a rolling history for short-term recovery. The Extended Journal Copy feature builds upon this by taking data directly from the DR site storage and moving it into a long-term repository. Because the "copies" are derived from the data already present at the recovery location, there is no impact on the production site performance and no requirement for additional storage space at the primary site for backup retention. This "off-host" backup approach eliminates the traditional backup window and ensures that the production environment remains lean while the DR site handles both short-term recovery (seconds to days) and long-term compliance (months to years).