It is no secret that virtualization is changing everything about how organizations think about and manage their data centers. But perhaps one of the more dramatic changes in thinking that still needs to take place is in how they should approach disaster recovery (DR). The odds of this shift in thinking occurring sooner rather than later just got a boost as a result of the announcement of a new replication software solution from a new company called Zerto that emerged from stealth mode this week.
What makes an emerging replication software provider like Zerto noteworthy is that its core team consists of the founders from Kashya (now EMC RecoverPoint) as well as individuals from the team that developed Topio ReplicatorX (acquired by NetApp a few years ago.) So while there are already a large number of providers of replication software on the market, Zerto’s founders already understand what the enterprise demands are for replication and what it takes to succeed in this space.
However what its founders are doing differently in this iteration is that they are no longer taking an appliance-based approach to replication such as they did with Kashya and Topio. Instead Zerto is moving its replication software up the stack into the server and, more specifically, into the hypervisor.
This change in approach is predicated in large part upon the accelerating adoption of server virtualization in all size enterprises. Server virtualization theoretically should make it easier for organizations to implement and manage replication and disaster recovery for a larger number of applications since they are consolidated on a fewer number of physical servers.
The problem is that the traditional ways of doing replication are less than ideal in virtualized environments. Host and application based replication solutions require access to and installation of the software on the VMs while using storage array-based replication requires that all VMs use that storage array. Deploying replication on an appliance (which is the legacy of the founders of Zerto) is also less than ideal since the traffic for the VMs needs to be routed through the appliance.
However what promises to make traditional replication techniques even more untenable is the growing trend of virtualizing more business critical applications such as Microsoft Exchange, SharePoint, Oracle, SQL Server and others. A single application may reside on multiple VMs and unless all of the data across all of the VMs on which this replication resides is periodically synchronized, it becomes almost impossible to create a consistent point of time for recovery at the target site.
So by Zerto introducing its replication software solution into the VMware hypervisor (actually it is a virtual appliance running on the hypervisor,) organizations can replicate the data on any VM. This technique eliminates the need to either deploy agents within individual VMs or use a specific storage array-based replication technology. It also means that users can now manage relevant virtualization elements, such as VMs, virtual disks, virtual networks, and virtual application, instead of less relevant physical elements: physical servers, storage arrays, and LUNs.
Yet what is most impressive about Zerto’s replication solution is not that it is available as a virtual appliance or a software only solution. This type of solution is available from other existing replication software providers such as InMage and FalconStor.
What makes Zerto unique is its high level of integration with vCenter and the flexibility it gives administrators to create Virtual Protection Groups (VPGs) that are application consistent copies of applications across VMs regardless of where these VMs are logically or physically located.
To configure a VPG, administrators can from within vCenter select which VMs make up an application. Then by putting these VMs in a VPG, Zerto replicates the data associated with all of those VMs as it normally does. However as it is replicating the data from within those VMs, it periodically creates an application consistent checkpoints across all of the VMs in that group. In this way, organizations always have a recoverable, application consistent copy of the application at the target site.
As one might expect from a company that is founded by individuals who sold replication software into the enterprise space, Zerto has many of the features that one might expect an enterprise software solution to have. It provides automatic handling of WAN disconnects or degradation, automatic IP re-configuration, bandwidth optimization, compression and throttling.
Yet what contributes to Zerto’s uniqueness in the replication space is its ability to centrally administer and group VMs by application so that companies easily have the ability to recover them at a remote site. Further, since Zerto is deploying its software at the hypervisor level, it opens the door for any organization to replicate data almost anywhere – be it to an internal private cloud, to an external cloud provider or even create some type of hybrid private/public cloud disaster recovery solution.
So while it is premature to give a Zerto a resounding endorsement in terms of how well its technology works, the approach it takes to implementing and managing replication in virtualized environment clearly merits attention. Because by breaking free from application, host and storage array-based means of replication, Zerto gives organizations new freedom to rethink how can protect and recover their virtualized servers in the datacenter of tomorrow.