It is highly unlikely that such a response to an order enquiry will ever come from a Plant Hire company, but it could be a good bit closer to reality than any of us might think possible at the moment. 

What prompted the thought is the news that Dell has come up with its own version of the `datacentre in a container’ model pioneered by the likes of Rackable Systems and Sun Microsystems last year. With IBM also planning a move into the market it has to be assumed that these companies see some future in it, or that the marketing suits are at last running the asylum.  

Is there a market for a new line of business for containerised datacentres? At one level it does seem unlikely. I can see an argument as a rented-in piece of kit as a short-cut to a resource overload, but it is only theoretical. Even a totally committed Sun, Dell or IBM shop would struggle to quickly configure such a system for any existing production environment – and that is what an overload solution would require. And given the pace of change and development in datacentre equipment, it would probably spend most of its time back at base being updated. Also, which company would be happy about the risk of leaving valuable data on disk drives for the next user to find?  

This, I suspect,  is a model that would only suit committed legacy systems environments – for example a bunch of old Digital Equipment Vax minis in a container. 

There is an argument for using the approach to solve the real-estate problem. Yes, it could work if there is already a pile of containers outside the building. But what company would risk serious data to a system where any light-fingered `entrepreneur’ could come along with a flatbed truck, a small crane and a set of cable cutters and have it away in half an hour? 

But there may be one market where the containerised datacentre could work. The number of corporate references to `The Cloud’ has started to grow. Indeed, it is likely to be this year’s SOA, the big buzzphrase which every vendor in the world suddenly claims to be offering and thus confusing everyone as to what any of it might mean.  

While The Cloud has huge potential to offer, it also provides significant risk for many companies. Some of these are genuine, while some are based more on perception and myth, but the containerised datacentre could find a niche as way for businesses to test out the idea and potential without putting the existing production environment at risk. And there is a genuine argument that The Cloud will not fit all applications requirements. There will be situations where companies need their own systems.  

One such situation is the issue of transaction speed, in business environments such as banking where there are very high transaction rates. Here, brute physics can get in the way. For example, it takes some 50m/sec to get a signal from coast to coast in the
US, not much slower than the speed of light. But that is still too slow for high transaction volume applications. This invalidates every other argument in favour of operating in The Cloud. 

And when such a company has worked out how and where The Cloud and the infrastructure integrate and interoperate, it can then decide what type of `cloud’ is required. For some a totally internal Cloud may prove to be the best approach, for example. But for many, they will then have learned where The Cloud makes sense and how to use it, in their own terms, `safely’. Then they may not require the containerised datacentre and it can be reused with another experimenter.