Looking at Remote Infrastructure Management

I remember working for an investment bank a decade ago and one of the key priorities of the CTO was cleaning up the infrastructure and improving network support. Nobody from the business was interested in supporting his projects. He even showed them PowerPoint slides of server rooms resembling spaghetti, but when the air conditioning failed one day and it took hours to recover after the servers had all overheated, suddenly the bankers were interested in the infrastructure that kept them trading in the market.

Back then, it was still common to have a big cooled server room within the office of a bank. Now infrastructure as a service, purchased from the cloud, is becoming the latest strategy executives are being persuaded to support.

But remote infrastructure management (RIM) has been talked about for years. IT suppliers have rearranged their support teams and built data centres offshore all chasing the infrastructure dollar as more and more companies consider the risk of having the heart of their network onsite.

The consulting firm McKinsey studied the various types of IT infrastructure in the past and suggested that around three quarters of all infrastructure management could be offshored. That potential market ensures that every major IT supplier now has a busy team all working on RIM services.

And RIM can be applied to many different areas. Network services are probably the best example of what can be remotely managed, but internal helpdesks, server admin, and general maintenance can all be run remotely.

It’s an interesting market development. It was not very long ago that pundits were analysing the cost of offshore outsourcing programmes and suggesting that in locations such as India, the telecoms cost is very high, but the labour cost is very low – so tasks like software development work well. But that model doesn’t work so well for infrastructure.

A huge server farm obviously needs some expertise to keep it ticking over, but the main cost is the hardware, keeping it cool and constantly working, and the telecoms cost to allow these remote servers to behave as if they were in the same building as the client. So, perhaps counterintuitively, it has often made more sense to locate a server farm in Scotland or Sweden than offshore in India or China. The communications are better and cheaper, the cooling is not as intense, and the expertise is available at a reasonable price.

But now companies are moving beyond the RIM promise and into the cloud, so an additional layer of business abstraction is piled on to the infrastructure. A company may well know what application they want to use, but have no insight into the infrastructure needed to make it work – and less insight into the physical location of those servers.

The cloud would suit them just fine because it’s become possible to buy the infrastructure you need, as and when you need it, all because of the cloud concept becoming mainstream.

And it’s not just a theoretical construct. Think of some scenarios where rented infrastructure and flexibility might be important. How about a retailer doing 90 per cent of their business in the month before Christmas? Or a start-up firm not sure if they are going to grow 20 or 200 per cent this year? Or a train timetable service coping with a spike in enquires, as flights are cancelled – familiar to Europeans dodging the Icelandic ash cloud.