Redundant array of inexpensive 'things'

There has been plenty of hype surrounding the internet of things (IoT) and especially super smart things such as the iconic Nest (now Google) thermostat. While many of these devices are interesting, they often come with premium pricing for an expansive set of features doesn’t fit all needs – some things are a bit less smart (& sexy) but could bring significant value.

This is where historically the term ‘Machine to Machine‘ (M2M) has been applied, especially by mobile telcos desperate to seek out new revenue streams from connecting lots of remote low data demand devices. Here, it is often only one or two attributes or states of each relatively dumb ‘thing’ that are of interest – a sensor taking a measurement, a system being switched on. One sensor on its own, is not really interesting, but using the concept en masse with dozens, hundred or thousands of sensors or devices makes things much more interesting.

This does not mean suddenly flooding the internet with masses of data – no matter how much the telcos might like that idea. The term IoT conjures up an image that all of these things, smart and dumb, will be connected to a single network, when in reality most have very little in common, except their ability to converse using a basic universal protocol set, based on IP. An effective IoT application is one that might take advantage of some connectivity to the wider internet, but is also built on internet technologies that exploit the economies of scale of standard components and common protocols.

When this concept of simple, mass connectivity crosses into the physical aspects of the ‘things’ as well, the proposition becomes even more interesting, even in what seems like the simple use of under floor heating.

In the industrial research centre, SPECIFIC, in Swansea, a combination of academic research from Swansea University, with industrial skills from Tata steel, NSG Pilkington and BASF, is leading to some interesting developments of ‘things’. The SPECIFIC consortium has the concept of ‘buildings as power stations’ at its core, and is creating low cost, robust items to capture, store and give off, energy.

One example that is currently being commercialised is a heated floor tile. This is a standard 600mm metal wrapped square of chipboard, designed to be stood on free-standing posts to provide a raised access flooring system, typical of most office, business and educational environments. The only difference is a coating on the top and a power connector, and with power applied, the upper surface of the tile warms up.

Under floor heating is not new, but fine-grained control of a tiny area at very low cost, is. However, being able to individually heat every single floor tile is only of real interest when intelligent controls can be applied. What is the temperature in the office, how many people are currently in there, which tiles are exposed or covered by furniture, which rooms are they in, which tiles is the sun currently shining on etc.?

A single smart system that detects some of this data from sensors, has other elements filled in from information in employees’ calendars or room booking systems and then access to weather sites and other external data sources over the internet can start to be very effective, efficient and comfortable.

This type of application requires some integration and might not look a elegant as the smart Nest thermostat, but could deliver significant benefits such as cost and energy reduction by applying heating when and where required. This would not be a smart IoT object that ‘learns’, but a distributed one that constantly takes into account the current and forthcoming situation and applies those requirements.

Some IoT applications focus too much on putting smarts into every individual ‘thing’, but without fine-grained control of the related physical attributes (in this case heating and energy) being managed, the cost of deploying some IoT applications might be higher than the value of the benefits they deliver.

Smart applications can be built by intelligently assembling and connecting a great many dumb components, not simply by adding expensive smarts into something that was historically dumb. This will be increasingly true for enterprise IoT applications, which have many legacy components and systems to accommodate.

With so much attention on sexy consumer IoT applications there is a danger that the skills necessary for commercial IoT integration to physical things as well as the requirements for IoT applications that deliver real value for businesses, will be overlooked. It would be a shame if the current IoT bandwagon leaves great masses of ‘dumb’, but worthwhile, things behind.