Billy MacInnes interviewed our Service Delivery Director, Ciaran Connolly, for the August edition of Tech Pro. Check out the transcript of the interview below:

Building a plan is very strategic, Connoly says, but most smaller companies don’t generally think about it because they are “so focused on getting the service out to the customer and they haven’t been burnt. When you’re brought in to talk about a project with the company, they have a functional requirement, but if they’re not mature enough, they might not think about non-functional requirements.”

The bigger the company, the more necessary BC and DR is. As the service becomes more important, companies need to think more about non-functional requirements: security, data protection and being able to recover from unscheduled outages. “Larger companies are already aware of it,” Connolly remarks. “They nearly demand you do a security review, make sure the data is protected in the right way and have a minimum unscheduled outages for applications. They nearly give you a deployment scenario.”

One of the big trends identified by various market research companies is DRaaS (disaster recovery-as-a-service) but Connolly says it really depends on the client’s requirements. “The more important the application, such as billing application with all a company’s primary data, the more it will want to make sure it’s able to recover it in any kind of scenario. That’s why mature companies have geographical redundancy for many of their applications.”

For large solutions with various levels of stacks, a BC plan needs to get many people involved and if one aspect of the solution doesn’t work, there’s a chain effect. “It can be quite easy to recover from or quite difficult, but you have to invest time and effort,” he warns. “If you’re outsourcing infrastructure or an app or service to a cloud vendor, you’re giving that responsibility to someone else, but you still have to set up and manage it and all the processes around it have to be understood. When there is a breakdown somebody has to manage that aspect.”

The process around BC and DR is templated and quite formulated, so it’s not overly complex but organisations still need to formulate a plan and test the recovery. They also have to invest time in testing security, data protection and failover. He accepts that “using AES is great, but deploying on your own infrastructure works as well. There’s not a size fits all but it’s nice to have options and to walk through those options with clients. We might recommend putting parts into the cloud because it might be cheaper in the short term to use cloud. A lot of decisions come down to cost and suitability.”