The Evaporating Computer Room

A long time back when I started my career, one of the activities that I was responsible for was to help my customers create computer rooms. I am referring to the era when mini computers had just made an entry and departmental computing was just about taking off. In India, the minicomputer brought systems within the reach of many mid-sized companies that could ill afford the big iron.

Creating computer rooms was a great exercise with raised flooring to pass a lot of RS232 cables and BNC Ethernet cables, air conditioning in the era of non-precision ACs and the rest of the office being serviced by fans, the IT guys were looked upon as the privileged lot. Security was a great concern and it was a lock-key and a security guard. Fire sensors were rudimentary and extinguishers bulky. The glass house thus became the envy of the rest of the enterprise.

Dust being enemy number one, everyone had to remove their footwear before entering the temple of computing. No exceptions to this rule, some of the larger computer rooms were housed in vantage locations within the office premises. Temperatures of 16 - 18 degrees centigrade ensured that no one could stay inside without warm clothing for too long.

Fast forward to today, the big commercial datacentres run by specialist service providers and telecom companies are under threat of extinction. Almost every software or hardware company has announced their version of cloud services where the enterprise does not have to buy, maintain and manage the hardware and software. Pay-as-you-use is the new mantra and this has raised enough hype in the IT fraternity.

Eventually the cloud is seen as the nemesis of the in-house datacentre as well as the commercial colocation services. In this situation does it make sense for an enterprise to invest in on-premise captive datacentres? Does the math add up to a financially viable model that justifies the creation of a mini facility with all the trappings of a large datacentre even if on a smaller scale?

Faced with this predicament many years back, we decided to collocate to a commercial datacentre; our RoI for the investment was in decades, not years. It was an emotional decision for the infrastructure team like that of giving away a bride. They overcame the state when they saw lower failure rates and stability not experienced thus far in the small computer room that had outgrown its usefulness. What should the CIO faced with such a decision do?

If you are in the IT or related business and there is a possibility of commercially exploiting your IT infrastructure in the future on a large scale, then do go and invest in one. If you belong to a diversified conglomerate or group with a lot of cross charging and allocation of costs, or a hosted software solution provider, it may still make sense to create one; do look at the possibility of getting this as a service from your service provider. You may be pleasantly surprised.

Vendors after vendors will want you to invest with their version of datacentre management or hardware tools. Their sales pitch is always about reducing datacentre costs, pain points, provisioning or management. Some have now started offering their solutions as service too. Look at these with a magnifying glass and always engage in Proof of Concepts before making any commitments.

Consumerisation of IT has already moved applications and some of the data outside the datacentre. If eventually everyone will be in the cloud, years from today do you want to be seen as the one who did not see this coming? Well, it’s a choice to make.