IoT Connectivity

Geoff Varrall
July 14, 2015


Internet of Things connectivity is often discussed in the context of the 4G to 5G transition including the spectrum and standards requirements needed to achieve an order of magnitude reduction in connectivity cost and an order of magnitude decrease in connectivity energy consumption. There is a persuasive argument that this could be accomplished by repurposing GSM spectrum at 850 and 900 MHz, subdividing existing GSM 200 kHz channels into narrow band carriers that could support low data rate long range low cost connectivity

The counter argument is that the Internet of Things also includes devices that require high data rate and low latency connectivity. This implies a wider range of spectrum and physical layer options including LTE and LTE Advanced and whatever comes next, coupled with regulatory innovation – the licensed versus lightly licensed versus unlicensed debate.

At this point the discussion becomes alarmingly unfocused, failing to comprehend the cost of band and standards complexity.

Band complexity

The number of mobile broadband cellular bands has increased from three core bands in 1990 to thirty bands in 2010. Currently the count is 44 LTE bands including TDD options. By 2020 there will be at least three hundred band combinations assuming carrier aggregation is adopted.

That’s before WiFi at 2.4GHZ, 5 GHZ and 60 GHz is added to the equation plus other country specific ISM bands such as the US ISM band at 902 to 928 MHz.

Each new band and band combination increases design and test cost. Each additional filter and switch path introduces additional component cost and performance loss making any of these connectivity options progressively less efficient for IoT.

Standards complexity

The relentless LTE Release process is increasing R&D and test cost at an alarming rate compounded by an IEEE standards process that remains at best loosely coupled to 3GPP/ 5GPP work items.

The cost targets of IoT connectivity can only plausibly be met by realising a global harmonised spectrum allocation and physical layer implementation. Getting consensus agreement on the optimum choice becomes harder as the options multiply.

A possible solution – think wavelength not frequency

A possible solution is to think in terms of wavelength rather than frequency

For the first fifty years of radio, wavelength was used as the default way of describing spectrum. This was simply because it was easier to measure wavelength and harder to measure frequency. Many of these descriptors are still used today including long wave, medium wave and short wave.

Accurate measurement of frequency required highly stable quartz crystal oscillators. As these became more readily available through the 1930s there was a shift towards describing radio in terms of frequency – VHF, UHF or other arbitrary naming systems – C,X and K bands for radar for example.

The introduction of cellular radio from 1980 onwards marked a shift to describing radio systems with a band number, the 800 MHz AMPS networks became Band 5, the 900 MHz TACS/ETACS networks became Band 8, and the 1800 MHz networks became Band 3.

But many of the design and performance challenges of IoT connectivity revolve around form factor and RF efficiency. These are a function of wavelength rather than frequency or band number.

For example an IoT device at 300 MHz is working at a metre wavelength (wavelength being the speed of light, 300 million metres per second divided by frequency).

Given that the optimum theoretic length for an antenna is one quarter or one half of the wavelength to be received or transmitted it is clear that a 300 MHz IoT device is either going to have a big antenna or an inefficient antenna, neither of which is desirable.

In terms of IoT device size and performance it is therefore helpful to think about three rather than 300 bands, the Metre Band, (300 MHz to 3 GHz, 1 metre to 0.1 metre), the Centimetre Band (3 GHz to 30 GHz, 10 to 1 centimetre) and the Millimetre Band (30 to 300 GHz, 10 to 1 millimetre).

The Metre band (300 MHz to 3 GHz) captures more or less everything we have in present mobile broadband wide area connectivity and the US 900 MHz ISM band.

The Centimetre band (3GHz to 30 GHz) captures 5 GHz WiFi with 802.11p for the connected car and satellites at Ku band (12-14 GHz, 2.49 centimetres to 2.14 centimetres) and Ka band (27-30 GHz, 1.1 to 0.99 centimetres). New launch and propulsion /orbit station keeping techniques, spot beam antennas, phased array antennas and solar panel efficiency gains are together transforming the link budgets and economics of these wide area connectivity options.

The Millimetre band (30 to 300 GHz) captures WiFi at 60 GHz but placed in the context of the options for 5G connectivity including the 66-71 GHz band, 71-76 GHz band, the 81-86 GHz band and the 92-93 GHz band – effectively repurposed lightly licensed point to point radio for mobile broadband IoT connectivity. It also comprehends present and likely future sub space radio system innovation including Google balloons and Facebook Drones for low cost internet connectivity.

The Millimetre bands are not generally discussed in the context of IoT connectivity but these are the bands which DARPA and the DOD are proposing to use to implement high bandwidth low latency military radio and telemetry systems supported from sub space base stations on board unmanned aerial vehicles. These are not local area networks but mobile wide area rapidly deployable systems with a free space line of sight range of up to 60 km.

We tend to think about the Internet of Things in terms of familiar everyday objects that become connected – the pot that calls the kettle back. This ignores the market for Small Devices that do things that we haven’t thought about yet. These small devices need short wavelength transceivers. The Internet of large things will need supporting as well and includes those dastardly drones. This wide dynamic range of devices requires a wide but selective and optimised range of spectrum and technologies.

This could be achieved by having not more than two IoT 'micro bands' within each wavelength 'macro band', one for local area, one for wide area connectivity. In the Metre band this would be a 900 MHz repurposed GSM carrier for wide area with 2.4 GHz for WiFi/ Bluetooth local area, in the Centimetre band, WiFi at 5 GHz including 802.11p for car connectivity and Ku or Ka band satellite for wide area, in the Millimetre band, 60 GHz WiFi for local area and one of the four sub-100 GHz bands for wide area sub space. Each option would use simple two or four level frequency or phase modulation. This combination would provide a wide range of data rates, coverage from a few metres to many kilometres, energy efficiency and global market and design scale economy.

The Internet of Things is promoted as a new market paradigm but we are still thinking about connecting the IoT in an old fashioned way.

Think wavelength not frequency. It makes the debate about IoT technology and spectrum much easier to comprehend.



Geoff VarrallGeoff Varrall joined RTT in 1985 as an executive director and shareholder to develop RTT's international business as a provider of technology and business services to the wireless industry. He co-developed RTT's original series of design and facilitation workshops including 'RF Technology', 'Data Over Radio', 'Introduction to Mobile Radio', and 'Private Mobile Radio Systems' and developed 'The Oxford Programme', a five day strategic technology and market programme presented annually over a period of twenty years.

Geoff is a co-author of the Mobile Radio Servicing Handbook, Data Over Radio, and 3G Handset and Network Design. Geoff's fourth book, Making Telecoms Work – from technical innovation to commercial success was published in early 2012. He is presently writing a book on 5G spectrum and standards. He also writes regularly for a number of European trade journals and chairs a broad cross section of industry conference and trade events.

As a past Director of Cambridge Wireless, Geoff is actively involved in a number of wireless heritage initiatives that aim to capture and record past technology and engineering experience.