View a markdown version of this page

Cohesity DataPlatform - VMware vSphere Backups to Amazon S3

Cohesity DataPlatform

Cohesity DataPlatform is based upon a hyperconverged architecture. Each appliance in a cluster contributes locally attached storage to a highly available, dynamically optimized virtual storage pool.

The direct-attached storage in each node is either physical HDD/SDD drives, or a combination of Amazon EC2 instance stores and Amazon Elastic Block Store (Amazon EBS) volumes.

DataPlatform offers a variety of storage-related services. This section is focused on how to use it to conduct VADP-style backups of on-premises vSphere clusters to Amazon S3.

There is one supported approach relevant to this use case: Amazon S3 as an archive repository. The most recent backup data (typically 30 days) is kept locally on the cluster. Once it has aged past that, it is then archived to one of the Amazon S3 storage classes for long-term retention.

Note

Archived backup data is fully self-describing. An Amazon S3 bucket containing a complete backup set and a Cohesity cluster that can read it is all a customer needs to conduct restore operations.

Logical architecture

Logical architecture for Cohesity DataPlatform.
Cohesity DataPlatform logical architecture

Illustrated workflows

  • Backup of VMs to local DataPlatform cluster in Site 1

  • Backup of VMs in Site 2 to a DataPlatform Cloud Edition cluster in AWS

    Note

    Automated deployment of a Cloud Edition cluster is initiated through Helios, a monitoring and management service offered by Cohesity.

  • Archival of old backups from DataPlatform Cluster in Site 1 to Amazon S3

  • Archival of old backups from DataPlatform Cloud Edition cluster to Amazon S3

  • Archival by AWS Lifecycle policy from Amazon S3 to Amazon Glacier

  • Recovery of VMs from Amazon S3 to the original DataPlatform Cluster in Site 1

  • Recovery of VMs from Amazon S3 to Site 2 via DataPlatform Cloud Edition

  • Conversion of VM backups in Amazon S3 to Amazon EC2 instances leveraging the CloudSpin feature of the DataPlatform Cloud Edition cluster located in AWS

Note

DataPlatform 6.4 does not support the HotAdd Transport Mode.

Cohesity DataPlatform 6.4 feature support

VADP Backup Features Supported S3 Features Commercial Support GovCloud Support
Application Consistency Yes1 Amazon S3 Standard Yes Yes
Volume Consistency Yes2 Amazon S3 Standard-IA Yes No
Source Side Compression Yes3 Amazon S3 One Zone-IA Yes Yes
Source Side Deduplication Yes3 Amazon S3 Intelligent-Tiering Yes Yes
Global Deduplication Yes Amazon Glacier Yes No
Changed Block Tracking Yes Amazon Glacier Deep Archive Yes No
Single File Restore Yes Amazon S3 Cross-Region Replication Yes Yes

1 via Microsoft-supplied VSS Writers in Windows Guests running VMware Tools; via vmsync for Linux

2 via Microsoft-supplied VSS Provider in Windows Guests running VMware Tools

3 Source-side in this context refers to the Cohesity appliance, not the vSphere hosts.

Further information

For additional information, refer to: