To most people, datacenters are a pretty boring concept. Rooms full of computers, processing away, generating data and heat, and maybe a couple of jobs. Not exactly the things with which compelling fiction were ever written. In print and film, the datacenter of the future was populated by giant computers, possibly with potential evil overtones, interacting with humanity through a single point of contact (usually before the computer went mad and began platting to destroy the world).Controlling the environment, however, was rarely the primary concern for these fictional super-computers.
When I was chatting with some techno-geek friends about how they thought the public viewed datacenters, most of them first brought up the film “Colossus: The Forbin Project” (yes, we are a middle-aged group of geeks), then of course, we had the obligatory argument over whether HAL could be considered a datacenter device, and then a spirited discussion over the consideration that the robots Gog and Ma-Gog, from the film “Gog” were actually an example of datacenter automation gone amuck. Skynet, was then brought up as the ultimate in “amuck”, to little arguement. When one of us mentioned the computer in the Hitchhiker’s Guide series, he was soundly thrashed by the person to his right. There are some limits, after all.
I hadn’t given that evening much thought, but it immediately came to mind when I came across a greentechmedia.com article on building automation that talked about Redwood Systems building automation. They’ve introduced a series of devices, ranging from lighting to environmental sensors that connect over their own, independent networking system that provides both power and data transport, and connects back to a central control system that monitors device status and can control their functionality. I immediately flashed on that classic science fiction film imagery of a man walking down a corridor or through a dimly lit room, while the lights came up in front of him and went off behind him.
I also admit to memories of parental admonitions to turn off the lights when I left the room and other tasks that could easily be served by some sort of automation engine. The smart house concept has been around for ages, but we’ve yet to see any successful, large scale adoption of home automation systems, but applying that same concept to offices and datacenters makes a huge amount of sense.
Science fiction has long postulated all sort sorts labor saving devices but there have rarely been practical applications of the concepts, when the idea was that the device would presume what its user needed. Facilities automation systems are one area that should benefit from these concepts, and there are already large systems in place that do power and climate control in many buildings, though almost always on a scheduled basis. Starting with the datacenter, where real-time power and cooling management can make a significant impact and the conditions of the environment are tightly controlled will allow vendors of this type of equipment to fine tune their hardware and applications and develop a body of data that they can begin to apply to moving these technologies into more general purpose applications.
This is a long way from the omniscient technologies often postulated and a very different path from many previous technological accomplishments, which were focused on “labor saving” as their primary goal. I suppose that you could refer to these devices as ‘resource saving” focused, where their goal isn’t making an individual’s life better, but to improve the efficiency of the environment of those that choose to utilize them. There are cost savings to be had in detail management of facilities environmental systems, and concurrent savings when you look at increased lifespan of devices that are properly managed. That smartphone in your pocket has more computing power than the computers that NASA used to get us to the moon; is there any reason to think that the same capabilities can’t be applied to devise scattered throughout your datacenter environmental controls.
