Data Centers Without DR Centers

DR has always been a challenge for any business entity dependent on an IT Infrastructure. It’s not only about keeping the services available during a disaster but more about how we efficiently deploy redundant systems so that they are available when needed. Secondly, another more discussed factor is COST. 

DR is always miss understood as Half capable of DC, and due to this school of thought companies going into disaster situations keep trying to bring back Data Center instead of consuming services through their Disaster centers with same service level which is definitely not possible with ½ infra as of DC

With state of art Datacenter facilities where all users are very happily accessing all their day to day applications without a delay would start crying and get disappointed from response of application as soon as Datacenter is not accessible and all services are redirected to DR. If we take a break here and look at the duration of services handover from DC to DR is another point of discussion and will definitely discuss this in more details in a separate paper. DC to DR movement requires an advance level skill to understand each application, their replications and then accessibility.  

First scream everyone will hear is “my application is not working, it was working just fine a second ago” this information reaches to Network Operation Center where no alerts are received, Application engineer is called in to check but this team also complains of no access to server, eventually once server team tells that its Network issue since none of the servers are accessible. Network team starts digging in. Take a pause here…. Why not alerts??? The whole Datacenter is down and this is impacting business a fortune as of now as its cloudy and no one knows what happened.

Off course monitoring application is hosted in DC and either there was no monitoring from DR to save license + hardware cost or no one every bother to check or setup health alerts for monitoring server it self. Now this is another topic how to monitor a monitoring system.

I believe its time to wrap up this discussion and break it down to multiple stages to discuss and come up with a state of art Datacenter with a check list marking it state of Art, and only A Data Center but Data Centers.

One primary conclusion that we can take from here is that there is no DR needed, we need 2 Data Centers with exactly equal infra to handle users requests with same level of greetings. Secondly the method of failing services from one DC to 2^nd DC or to 3^rd DC (Another point of discussion), third but not last a better monitoring system that can raise alarm before people starts screaming.

In next sessions we will come up with more discussions on DC Scalability, Business Continuity, Data Protection, Continuous Monitoring, High Security.