You've moved your disaster recovery operation to the cloud and now it's time to test it. That's essential because it's the only way to guarantee your plan works and that you can recover critical data after a disruption.
Plus, testing allows you to train staff, perform maintenance, engage your business users and suppliers, and make your organization aware of disaster recovery (DR) procedures. Not testing can cause your plan to fail even before an incident. (Still, according to research from the Disaster Recovery and Preparedness Council, nearly one in four companies never test their DR plans.)
Cloud testing differs from traditional assessment. Gone is the need to send tapes and testers or break the link on your replicated disk. With disaster recovery as a service (DRaaS), you can start your cloud test at 8 a.m., begin user testing at noon, and complete the entire process by 4 p.m. All in a day's work.
The Testing Process
As with traditional DR testing, you seek to simulate a disaster as best you can and determine how long it takes to get the application operating in the cloud. A scenario-based approach works well in a test because it lets you use situations that could-prove to be most challenging. Make the test as close to simulating a real-life incident as possible, within reason and your budget, of course.
What you're seeking to do is identify potential weaknesses that can pop up in your DR plan or the people you use and their roles or the technology itself. In fact, consider all the possibilities for trouble. It also pays to have independent and experienced DR reviewers' help with the testing and with follow-up improvements. Be sure to obtain executive management approval.
Test Periodically
Set a time when demand on the system is slow, but that gives you enough time to recover if the systems don't return to life quickly. Many tests are performed outside office hours. That prevents disruption but might not reflect the true working environment. Of course, back up everything. Then turn the systems off, then on again.
Testing your DR plan periodically – at least annually – is essential. You make sure your experiences are consistent. You ensure any changes in the plan don't cause problems. And you give experience to new members of your team.
Similar to cloud computing, DR cloud testing – and post-testing – deliver significant cost benefits. Resources can be employed in minutes, but be sure if you're doing it yourself to have enough servers available for the testing. To avoid this additional CAPEX and OPEX, it often pays off for enterprises to outsource their production and DR environment to handle the testing efficiently and properly.
So remember: the one thing that can wreck your disaster recovery is not testing your procedures before an emergency erupts.
Vishvanath Muthyala (Vishy) is a program director/senior technical architect at Sungard Availability Services (Sungard AS) through Infosys. He has nearly 20 years of IT experience handling large infrastructure and cloud projects for the Fortune 100 companies, and has held key positions in program management, customer relationship management and vendor management. Currently Muthyala serves as program director for the Sungard AS managed recovery program. Muthyala holds a bachelor's degree in information science and technology from BMS College of Engineering. He holds certifications in TOGAF 9, IBM Cloud, ITIL, EMC storage and NetApp storage. He has filed a patent for "Determining, Managing and Developing and Application Topology in a Virtual Environment."
Bob DiLossi serves as director of crisis and test management, North America, for Sungard AS. In this role he manages daily and regional disaster events that cause customers to experience an outage or business interruption. Additionally, DiLossi maintains crisis management centers in Philadelphia, Carlstadt, N.J., and abroad. Since joining the Sungard AS team in 1998, DiLossi has assisted countless companies of all sizes and in a range of industries plan for and recover from hundreds of incidents. DiLossi developed and currently facilitates Sungard AS' internal crisis management training and certifications, and maintains operations and customer service teams with industry standard crisis management policies and processes. DiLossi spearheaded the development of the company's Weather Alert System, which allows his crisis management team to determine the impact of any weather-related event, quickly assess customers' safety and streamline the process for customer outreach.
