Skip to 0 minutes and 7 secondsSo how do we interconnect these billions of things? Ethernet cabling seems like not a good idea, for obvious reasons. Indeed, only wireless offers the flexibility, scalability and cost efficiency needed to ensure viable uptake of the Internet of Things. You save on cable costs, you save on installation and maintenance costs, and your things can move around freely. And we've got a successful wireless system operating on this planet, the cellular mobile phone system. But wait. Can you imagine recharging billions of sensor batteries once a day, as we do with our phones today? Well, clearly not. So we had a problem there.

Skip to 0 minutes and 50 secondsAnd it was the problem of coming up with a wireless connectivity solution which would solve the issue of limited power supply. Ideally you would want a system where a sensor can run on a single AAA battery for years. One of the world's leading technology standardisation bodies, the IEEE, was therefore tasked to design such a low power wireless system, and born was the IEEE 802.15.4 standard, which was adopted by a larger industry alliance aiming to promote this new low power technology. The name of that alliance is Zigbee Alliance, and this is what the system is referred to today. Zigbee, in essence, is a low cost, low power, wireless mesh networking standard targeted at wireless control and monitoring applications.

Skip to 1 minute and 42 secondsThe chips are typically integrated with radios responsible for the wireless transmission, microcontrollers responsible for orchestrating and controlling the data flow in the chip, and memory responsible for storing the data. Now when it comes to wireless technologies, the most important parameters to look at are first, over the air data rates, second, available frequency bands and the regulation of these bands, and third, the power consumptions. In terms of these factors, Zigbee is actually doing a decent job. There's only one small problem-- it doesn't work. It works over short ranges-- in buildings, for instance-- but the technology refuses to work in IoT applications we had dreamt of before, applications which require a highly reliable and highly scalable solution.

Skip to 2 minutes and 32 secondsWe tried with large-scale smart city projects to no avail. We tried with large-scale industrial monitoring projects-- again, no luck. The system would stop functioning. It would lose important data, among other things. Basically, it required babysitting, which is not exactly how we thought of getting the Internet of Things going. And this is when we realised that we may have made a few mistakes on the way, mistakes which arguably held back the IoT uptake by almost a decade. Yes, this is very serious. The first and arguably the biggest mistake of the community in the early days was to believe the pitch on the need for low power technologies. What we actually needed was a high power, low energy solution.

Skip to 3 minutes and 23 secondsThis seems contradictory, but remember-- energy drains your battery, power gives you range. Both are related where energy equates to power multiplied by time. To achieve low energy at a high power, one only needs to design a system which would be able to achieve data transfer in a short time while sleeping in an ultra-low power state with the remainder of the time. With a larger range due to a high power gain, the amount of repeaters and gateways can be drastically diminished, and reliability improved. Now here comes the mistake number two, which maybe is more like a nuisance, really. It relates to the ISM spectrum Zigbee operates in.

Skip to 4 minutes and 9 secondsThe ISM bands are license-exempt bands which are only lightly regulated, and thus allow for quick time to market of IoT products. However, the restrictions on transmission power and the inability to offer guaranteed data delivery, along with some legal complications, proved really, really imposing. Now as for mistake number three, another important decision factor for deployment is cost, where focus was too much on cheap chips and hardware whilst forgetting that what actually matters is the total cost of ownership. And one of the biggest costs in large-scale Zigbee deployments are the running costs-- mainly the salary of field engineers troubleshooting the network. A 50 pence Zigbee connection suddenly costs $50.

Skip to 5 minutes and 0 secondsOverall, while Zigbee is a great solution for short range connectivity, the technology and its variants are surprisingly unfit to power large-scale Internet of Things. Major companies have finally realised this, and some very interesting IoT connectivity alternatives have hence started to emerge. These are system which nobody even dreamt of some 10 years back.

Zigbee: our first steps

In the previous activity, we looked at sensors - the things we need to connect. Now we’ll move onto the process of actually connecting those things up.

In this video, Mischa introduces Zigbee, the first standardised system for connecting the ‘things’ in the Internet of Things, and discusses the challenges posed.

Share this video:

This video is from the free online course:

The Internet of Things

King's College London