Cohesity Data Cloud Platform Fundamentals
Cohesity Data Cloud Platform Fundamentals
This domain tests whether you can place each Cohesity component in the right architectural layer. Questions typically ask what a service does, where data actually resides, and which component fits a stated requirement - not node counts or speeds and feeds.
One cluster, many services
Cohesity Data Cloud is a scale-out, software-defined platform. Certified physical nodes, a virtual edition, or cloud editions are joined into a cluster that presents one distributed filesystem (SpanFS). Snapshots, files, and object data all live on that shared foundation, so platform-level capabilities - inline deduplication and compression, erasure coding, and SnapTree snapshots - apply to every workload on the cluster.
Services are capabilities on that foundation, not separate appliances:
- DataProtect - backup and recovery for VMs, databases, NAS, files, and SaaS workloads.
- SmartFiles - SMB/NFS file services on the same cluster.
- FortKnox - a Cohesity-operated, isolated SaaS vault for immutable copies.
- DataHawk - anomaly and threat detection over protected data.
- CyberScan - scans backup copies for vulnerabilities and indicators of compromise.
- SiteContinuity - orchestrated disaster-recovery runbooks.
Helios is the control plane
Helios is the SaaS management plane: it registers clusters, provides cross-cluster dashboards, alerts, reports, and orchestrates upgrades. It is not in the data path - snapshot, replica, and restore traffic moves directly between clusters, sources, and archive targets.
Core vocabulary
A storage domain groups underlying storage; a view is a logical SMB/NFS share presented from it. A protection source is a registered workload (vCenter, physical host, NAS, M365 tenant); a protection group binds sources to a protection policy and a storage domain; a snapshot is a point-in-time copy; a replica is a copy on a second cluster; an archive is a copy on external target storage.
Worked example
A hospital wants backup copies its own administrators cannot delete, held outside its data centre. Reasoning: retention settings alone do not stop an administrator, and replicating to a second on-premises cluster keeps the same trust boundary. The answer combines DataLock immutability on the local copy with a replicated copy in FortKnox, whose credentials are separate from the production cluster. Helios is not the answer because it stores no backup data.
Traps that cost marks
Assuming each service needs its own cluster; treating Helios as a backup target; confusing a view with a physical volume; assuming archived data is instantly restorable (it requires a retrieval job); assuming a replica is created automatically without being configured in the policy.