CloudArchive Direct

15 February 2026

CloudArchive Direct lets you archive data directly to an external target without retaining a copy of the data on the Cohesity cluster. In CloudArchive Direct, Cohesity cluster acts as a local cache to stream the dataset to the external target, eliminating the need to store a full backup copy on the Cohesity cluster first.

This approach reduces the capacity requirements on the local Cohesity cluster to a small cluster footprint, as only the metadata and index, and not the dataset itself are stored on the local cluster. Storing the metadata and index on the local cluster enables much quicker search and recovery down the road. The external target also stores the metadata and index, in addition to the data itself. On the Cohesity cluster, you can register a cloud storage, S3-compatible storage, or NFS-mounted servers as external targets.

For the first archival, depending on the size of data, the Cohesity cluster might experience significant capacity utilization, because the data first gets stored on the local storage (Storage Domain) and then aggressively gets down tiered to the external target. During this time, the capacity utilization will increase till the data is garbage collected upon the successful transfer.

The current version of Cohesity supports:

CloudArchive Direct for VMware VMs data source in addition to the existing support for NAS data sources.
Enabling CloudArchive Direct from the Protection Policy settings in the Cohesity Dashboard.

The option to enable CloudArchive Direct (native mode - no dedupe, no compression) from the NAS Protection Group page is deprecated in the 6.7 version.

How CloudArchive Direct Works

When you run a Protection Group for a registered NAS source with CloudArchive Direct enabled, then according to the policy set for the Protection Group, the file system from NAS is read and stored on the external target as a streaming backup using the Cohesity cluster as a temporary cache.

CloudArchive Direct

CloudArchive Direct Characteristics

Feature	CloudArchive Direct
Supported Sources	VMware NAS sources
Mode of enablement	While creating or editing a Protection Policy.
Deduplication	Supported
Number of Objects per Protection Group	Multiple
Data archival	Incremental forever
Metadata archival	Incremental forever
Granularity level of incremental archival	Block-level—Need to perform only incremental archival of a file if the data or metadata of the file is updated.
Native format as primary	No
Tiering of objects using Cohesity Data Movement	Supported—can choose to tier the archived data to a cheaper storage tier.
CloudRetrieve	Supported
Vault unregistration	Not supported

Supported External Targets

For the list of supported external targets, see Supported Software.

CloudArchive Direct Workflow	VMware			NAS
CloudArchive Direct Workflow	Standard Storage Tier	Archive Storage Tier	Hot Tier	Standard Storage Tier	Archive Storage Tier	Hot Tier
Direct Archival	Yes	Yes	Yes	Yes	Yes	Yes
Restore	Yes	Yes	Yes	Yes	Yes	Yes
File-level restore	Yes	No	Yes	Yes	Yes	Yes
Indexing	Yes	No	Yes	Yes	Yes	Yes
CloudRetrieve	Yes	Yes	Yes	Yes	Yes	Yes

*Indexing for VMware VMs in CloudArchive Direct-enabled Protection Groups may cause data egress from the external target, resulting in charges from the cloud provider.

Differences Between CloudArchive Direct and Cloud Archive

Term	Description
CloudArchive Direct	CloudArchive Direct copies the data from any NAS or VMware directly to any registered external target such as public cloud providers (AWS, Azure, Google Cloud Platform) or any S3-compatible external target to reduce storage costs and eliminate the need to store a full backup copy on-premises.
Cloud Archive	Cloud Archive sends a copy of the backup data (it contains the backup data, metadata, dedupe fingerprint, and index) stored on the Cohesity Cluster to any registered external target. To perform Cloud Archive, at least one full copy of the primary data needs to be stored on the cluster.

Figure X: CloudArchive Direct and Cloud Archive workflow

Considerations

CloudArchive Direct is supported for both physical clusters and virtual clusters. CloudArchive Direct is not supported for DataProtect for Cloud (legacy). DataProtect for Cloud (formerly known as NGCE cluster) uses a CloudArchive Direct architecture, in which backup data is directly down tiered to object storage, and only metadata is stored on the cluster’s cache tier.
If Encryption is enabled at the cluster level, you must also enable encryption while registering the external target for CloudArchive Direct. Archiving directly to an external target may fail if encryption is enabled at the cluster level but not at the external target level.
Review and understand the following considerations if you are performing CloudArchive Direct of NAS sources:
- When you upgrade the Cohesity cluster from 6.6 or below versions to the latest version, the existing Protection Group created with CloudArchive Direct native mode will not automatically switch to the new CloudArchive Direct format. Therefore, you must pause the existing Protection Group and then create a new Protection Group for the objects with CloudArchive Direct new format.
Review and understand the following considerations if you are performing CloudArchive Direct of VMware VMs to Archive Storage tier (AWS S3-Glacier, AWS S3-Glacier-Deep-Archive, AWS S3-Glacier-Deep-Archive, AWS S3-Glacier-Deep-Archive Gov):
- File and folder recovery is not supported.
- Instant recovery of VMware VMs is not supported.
- Instant Volume Mount is not supported.
- Test & Dev (Cloning) is not supported.
- Virtual Disks (VMDK) recovery is not supported.
- Indexing is not supported.

Dataprotect Documentation