What is Cirrus Migrate Cloud?

Cirrus Data Cloud uses a distributed computing architecture where each host runs a thin agent called Cirrus Migrate Cloud, for both management and data movement. To ensure the highest level of security, the Cirrus Migrate Cloud agent - deployed and running at each host being migrated - is never exposed to the internet. The agent makes a one-way call to Cirrus Data Cloud to register itself and to establish a secure connection for management. There is never any customer data stored or transmitted to Cirrus Data Cloud. The Cirrus Migrate Cloud agent is deployed by running a single curl line which contains a migration project identifier. This allows for automatic association of the host into the appropriate migration project at the Cirrus Data Cloud portal.

Cirrus Data Cloud automatically discovers these distributed agents and sends migration jobs to them, assesses progress, and executes the cutover process. Each Cirrus Migrate Cloud agent acts independently, performing the migration jobs defined by Cirrus Data Cloud. The migration itself is simply performed by the Cirrus Migrate Cloud agent by executing block-level copying from the source storage (the cloud’s elastic block storage) to the destination Virtual Cloud Block (VCB) storage disk while tracking block-level changes during the copy; repeating the process periodically until workload cutover is executed. A patented process called Transparent Data Intercept (TDI) is employed to ensure end-to-end transparency resulting in zero downtime for deploying, migrating and cutting-over source disks to destination disks.

An innovative and patented technology called “cMotion” allows live workload transition from the source disks to the destination disks without any host changes or disruption, thereby allowing production use of the newly migrated disk without downtime. Once cMotion has been executed, the source disks can be immediately deleted to avoid further usage charges in certified environments such as Linux instances of AWS Cloud, Google Cloud, and VMware Linux RDMs. For Cluster systems, cMotion facilitates end-to-end zero downtime migration for all Windows and Linux cluster nodes in all cloud, physical, or hypervisor environments.

Throughout the deployment, migration, and cMotion cutover processes, no downtime is required. Should the migration need to be reversed, it is not necessary to copy back the entire volume. Cirrus Migrate Cloud tracks block-level changes even after the cMotion cutover thereby ensuring that only the changed blocks are copied back if it becomes necessary to execute the back-out plan.

Given that each Cirrus Migrate Cloud agent runs independently, and no data is ever sent to the Cirrus Data Cloud cloud, the overall migration as a service has no practical limits. The “as-a-Service” nature of Cirrus Data Cloud ensures that features, enhancements, or updates are all implemented with no customer impact. As new host OS platforms are supported, an updated Cirrus Migrate Cloud Agent will be available during the automated deployment process.
Was this article helpful?
Thank you!