Cohesity (www.cohesity.com), the fast growing leader in secondary storage, delivered a very interesting session during the recent IT Press Tour that helped the press crew to understand all the progress made since our last visit in December 2015.
The company continues on a clear mission to "Redefine Secondary Storage" without any obstacles or I should say without considering any obstacle. As we understood last year and even before, Cohesity designs, builds and develops Data Platform as a pure SDS solution and sells it installed on a physical appliance with a hyper-converged model. The result is a consolidation of data services on same scalable platform able to deliver a gateway to the cloud, a target for backup, archive or tiering, some analytics functions and capabilities to offer a test/dev environment. It replaces various segmented not yet integrated backup, archive and tiering products that got deployed at different times and contributes to the explosion of the enterprise IT budget.
As a fundamental data protection layer, Data Platform now offers 3 key functions: CloudArchive and CloudTier which consider a remote cloud storage target such an object store as a natural extension to the on-premise instance, CloudReplicate, the last option, requires a remote DataPlatform instance to talk with the on-premise one. This capability is available with the Cloud Edition of the DataPlatform.
In addition to the data protection of virtual environments via DataProtect that leverages the VMware vSphere Storage APIs - Data Protection that pulls data without any agent deployed on VMs or Hypervisors, Cohesity is able to protect physical machines via deployed agents. It is now a global enterprise play and we're waiting some application agents to offer consistent backup for specific applications, databases...
To offer all these capabilities, Cohesity has made great efforts to develop a new distributed file system, named OASIS, scale-out by design with distributed KVS, metadata store and lock manager, and data reduction, snapshot, CDM and Hadoop integration. It finally merges 2 worlds and offers a pretty well advanced file system fully masked outside but really central internally. The snapshot technology built by Cohesity, named SnapTree, is unique and is optimized for write performance, again it relies on a distributed redirect-on-write model. All these bring to the DataPlatform an infinite criteria in capacity, in processing, in sites... It was really a very good session, hope we'll be able to measure again the progress in December 2017.