2012 Prediction: Nimble Storage
Virtualization Drives Democratization of Business Continuity
Mid-sized enterprises (MSEs) present a unique set of challenges to technology vendors. They have mission-critical applications and uptime needs like larger, global corporations, but cannot afford large capital investments or teams of experts to deploy complex solutions. The good news is that the aggregate spend by the hundreds of thousands of mid-sized enterprises has been a principal driver of breakthrough technology innovation.
One area where we are on the verge of such a breakthrough is in facilitating disaster recovery and business continuity – hitherto only the realm of large enterprises.
Three factors have made disaster recovery difficult for MSEs in the past:
- The capital cost of replicating servers, networking and storage at the second site.
- The bandwidth costs of replicating data between sites.
- Most significantly, the implementation complexity of replicating data, duplicating network and server set-up, and individually tailoring how the failover of each and every application should work.
As a result, the vast majority of MSEs have had to be content in the past with having an off-site copy of their data, often on tapes. In fact, even large enterprises have had to compromise disaster recovery for less-critical applications, putting them at risk to lose a day or more of data and suffer hours or days of downtime in the event of a disaster. However, each of these barriers has been crumbling over the last couple of years.
The biggest change has been that applications are now being deployed within virtual machines as the underlying containers, and the really complex problem of how to individually manage failover for each and every application has been simplified to one of how to manage failover of virtual machines from one site to another. Further simplifying this problem, VMware has deployed software (Site Recovery Manager) such that much of the set-up and configuration steps needed to fail over a virtual machine from one site to another can be automated - ready to be invoked at the push of a button in the event of a disaster. The same software solutions even allow customers to simulate a disaster and test how well the failover solution would work in the event of a disaster.
The second big change has been the significant reduction of the capital needed to set up an alternate site. Most customers now do not need to invest in their own facilities, networks and data centers to have a secondary site. There are several cloud and hosting service providers that provide infrastructure at a lower cost and higher resiliency than an MSE’s own secondary location might provide. Further, since virtualization facilitates sharing of a small number of servers between many applications, the capital cost of deploying physical servers that are only called into duty during a failure has shrunk substantially.
The third change has been in replication. Replication used to be inefficient in the amount of data being moved, and in many instances replication traffic was incompatible with traditional IP networks and required proprietary network equipment. That has changed substantially as modern storage solutions have become very efficient in tracking exactly what data has changed at a very granular level and only sending that changed data over traditional IP networks.
The surest sign of this transformation for me is when I look at our own business over the last year. Over 80 percent of our storage solutions are deployed in virtual server environments, and we have a replication attach rate of over 70 percent - astonishing by the standards of the storage industry, but understandable given the larger context of business continuity becoming simpler.
