Enterprises spend billions every year maintaining (and powering) duplicate racks and even entire datacenters, solely for occasional potential use, in the event of an unforeseen outage or disaster. Required by law in many cases, it may be the largest IT investment with the lowest return on investment. The money invested in disaster recovery isn’t wasted; it simply represents money well spent ensuring that applications will be highly available to users.
In the financial community (and others responsible for handling massive amounts of transactions and critical supply chain data), the cost of downtime has been well documented and more than justifies the DR investment. Outages are often more costly in terms of lost revenue, brand erosion, and employee productivity. So DR is like a kind of insurance policy, except instead of a policy-holder getting compensated for a loss, the policy-holder maintains two (or more) of everything. That is perhaps not the most efficient use of high-cost IT assets nor the energy used to power them.
The cloud’s pay-as-you-go model is a much more efficient model for the occasional-use dynamics of disaster recovery. That has yet to happen, however, at least for traditional multitier enterprise apps.
The public cloud has not evolved sufficiently to support the specialized service integration and controls required for recovery time and recovery point objectives; and the private cloud is simply a way for enterprises to maintain a duplicate app and service infrastructure more efficiently, albeit while paying a virtualization vendor for the right to do so.
That is why I’m convinced that the hybrid cloud is an ideal operating model for disaster recovery -- that is, a hybrid cloud that is a seamless extension of the datacenter. For small and midsized enterprises that simply cannot afford enterprise-class disaster recovery or for larger enterprises that want a more efficient or supplemental level of protection, the hybrid cloud is the optimum operating model. Pay-as-you-go for DR gives them both opportunity to reduce downtime and the ability to invest in greater IT innovation.
Yet as I have discussed here at Cloud Ecosystem, many executives, pros, and vendors are trapped in an outdated notion of a hybrid cloud as two distinct clouds (private and public) with some fundamental app migration capabilities shared across them. That definition would not support the requirements for enterprise-grade disaster recovery. For aggressive RTOs and RPOs the hybrid cloud needs to be a seamless extension of the datacenter, including services like authentication, LDAP, and Active Directory. It needs to support physical and virtual apps and be operating beyond the grip of virtualization or service provider lock-in.
When enterprises truly understand the power of a hybrid cloud operating model and how to deploy it, disaster recovery will be transformed into a more effective, efficient, and powerful way to protect critical applications.