Virtual SAN and Cohesity DataPlatform

By Gaetan Castlelein • July 11, 2016

Today, VMware and Cohesity are publishing a joint white paper on “Simplifying Data Protection for VMware Virtual SAN with Cohesity DataPlatform.” You can download a copy of the paper here. The paper is jointly authored by our good friend Rawlinson Rivera of VMware (@PunchingClouds), Sai Mukundan of Cohesity, and me. In addition to the paper, we are also running a webcast with Rawlinson and Johnny Chen of Cohesity tomorrow. You can sign up for the webcast here.

So – why do we think this paper and webcast are important? In a nutshell, we’ve experienced the benefits of Virtual SAN (VSAN) and Storage Policy-Based Management (SPBM) to make primary storage much better. But secondary storage and data protection haven’t improved much over the past decade or so. Cohesity DataPlatform provides the perfect hyperconverged secondary storage to complement Virtual SAN and bring the benefits of hyperconverged to both primary and secondary.

vSphere primary storage has gotten a lot better

VMware has made tremendous strides over the past few years to improve primary storage in vSphere environments. VMware Storage Policy-Based Management shifts storage management from being static and infrastructure-centric, to a dynamic, policy-driven, application-centric management model. VMware Virtual SAN provides the VMware hyperconverged storage offering. Virtual SAN has seen tremendous traction over the past couple of years, because it’s much simpler and cost-effective than provisioning traditional SAN and NAS.

Secondary storage is a new bottleneck

While we’ve all been very busy making primary storage better, we haven’t paid much attention to secondary storage. In fact, there has been a surprising gap of innovation in this space. Typically, vSphere data protection is dealing with the following issues:

  • Unnecessary complexity and costs: Data protection for vSphere consists of target storage (dedupe appliances such as Data Domain), backup software (Commvault, Symantec Netbackup), backup infrastructure (master servers, media servers), replication solutions, cloud gateways, etc.
  • Dark, unproductive data: The protected data is unproductive. It’s just sitting on a dedupe appliance until a disaster or data loss event happens. You can’t run test/dev, analytics, or any other kind of workflow from this data.
  • Static service levels and long RTO/RPO: The typical backup strategy is daily incrementals, weekly fulls. We usually don’t adjust these based on the application, business requirements, etc. They’re static service levels. They don’t interoperate with the new Storage Policy-Based Management way of doing things. We can’t provide faster RPO or RTO for more important applications.
  • No native cloud integration: Leveraging the public cloud requires complicated and inefficient bolt-on gateways.

Cohesity provides a fundamentally better approach to data protection for vSphere

What makes Cohesity unique? It boils down to a few key differentiators:

  • Web-scale platform with global dedupe: Cohesity provides much better target storage for backups than dedupe appliances like Data Domain. It’s a scale-out platform with global dedupe and compression. It eliminates all the issues of dealing with individual ‘point appliances’. There are no single points of failures, no forklift upgrades, and capacity planning and utilization are a lot more efficient. Cohesity can be used as better target storage with your existing backup software like Commvault, Symantec or Veeam.
  • Integrated backup, recovery, and replication: Cohesity eliminates all the clutter of separate backup software and infrastructure. It provides an integrated backup suite that runs directly on the Cohesity appliances without all the clutter of conventional backup solutions.
  • Making the protected data productive: Traditional backup solutions store the data as an unproductlive ‘blob’ on dedupe appliances. Cohesity takes a different approach. Backing up consists of moving a VM snapshot from primary vSphere storage and mapping that snapshot to a native snapshot on the Cohesity platform. With our SnapTree technology, we can provide unlimited, zero-cost snapshots and clones. By keeping each image as a fully hydrated, instantly available snapshot, we can provide immediate access to the data for a number of workflows. For example, we can create fast clones for test/dev purposes or for analytics workflows. With its hyperconverged architecture, Cohesity provides plenty of compute capacity to run applications and VMs directly on the Cohesity platform to support secondary workflows without having to copy the data to yet another storage device.
  • Policy-based management and fast RPO and RTO: Cohesity can provide RPOs of 15 minutes and instantaneous RTOs. The data protection service levels (retention policies, RPOs, etc) are defined in protection policies. Each application or VM can be associated with a specific policy to match service levels to individual application requirements.
  • Native cloud integration: Cohesity provides native integration with Amazon, Google and Microsoft to support data tiering, archival, and replication to the cloud. This integration is provided out of the box without the need for bolt-on cloud gateways.

For more information on the joint VMware Virtual SAN + Cohesity solution, you can download the paper here and sign up for the webcast here. And look for more great joint content to be published in the near future as we build out more demos and solutions!