Data Center Blues

Many organizations are facing some serious data center blues these days. Aging facilities, overloaded electrical plant, inefficient cooling, and shortage of floor space. Executive management is now asking themselves if now is the time to build a new data center, remodel or look at other options. On top of this, is the political pressure under the “Green IT” label and that we should be good corporate citizens.

Does it make more sense to build your own data center, use a colocation center to house your gear, or lease turn-key space from a wholesale data center provider like IO Data Centers? The size of a requirement has historically been a key decision point in sorting out the economics of data center expansion. But capital, control and speed to market are also important considerations in determining the best approach, more so if we are talking a global infrastructure.

A significant number of enterprise companies still prefer to build their own data centers, usually out of a desire to control all aspects of their operation. Security is often a guiding principle for companies that build their own facilities. Many financial services firms build stand-alone data centers to ensure that their critical IT assets are not sharing space with other companies. But hasn’t security, audit and controls come far enough that this is no longer a road block?

I find that many companies that do this study ask their IT department to conduct the study. Wrong choice. I would not have anyone inside the company conduct this study, but a 3rd party. There has to be an analysis of everything from facilities, property, operations, applications, infrastructure – the list is significant. You want a bias free evaluation that is provided to executive management. Then you bring in your IT management team, business leaders and have a open and honest discussion about best approach. Cost is not always everything; you do have to look at security, disaster recovery, performance, and operations.

I would also look at revitalization of your technology in your data center. It could be enough to buy you some more time to conduct an appropriate study. Virtualize as many of your servers as you can, along with your storage. I would also aggressively go after legacy application retirement, data retention and archive your data. In addition, can you do a hybrid solution around DR infrastructure? Does it make sense to have another Exchange environment running on premise – can that be move to Microsoft? Can you use a baremetal recovery solution for better asset allocation? All good options to pursue.

One more thing – you need a solid disaster recovery plan. You need to develop your plan – a written plan. Test the plan, have a dry run of a disaster. You and your staff need to know what to do in the event of a disaster that has impacted your facility or operations. This has to be a holistic DR plan – including staff, facilities, systems, data and infrastructure. I highly recommend CPSI Inc for your DR planning and design. Start now – to late to think about it the morning after.

Data centers are expensive – and given where many of the data center providers are today, compared to just 5 years ago – it is time to really look at Buy vs Build.