Earlier this year, Arqiva, a UK company that owns most of the broadcast masts in the UK announced that it was building a nationwide wireless communications network for the Internet of Things. As we are accustomed to hearing mobile operators tell us how many billions of pounds are invested in their networks, how can it be profitable to build such a network from scratch just to provide low-cost connectivity to connect meters, trackers and the like?
The answer is to apply different rules than to mobile networks. Mobile operators are being driven to provide ever higher data rates to support the rich mobile Internet connectivity that we are hooked on. This is very resource-intensive in terms of the amount of spectrum required, processing power at the base stations, as well as processing power on the mobile phone. While each generation of mobile technology makes more efficient use of the airwaves, they are also orders of magnitude more complex than their predecessors. All this means that mobile phones are expensive from a technology and network perspective.
However the drawbacks that this rush for speed brings do not outweigh the benefits for more prosaic applications such as smart meters or object trackers. Many of these applications can make do with a few kbits/s and require good coverage, particularly indoors and a long battery life, which are not met by traditional cellular technology. This is where a new breed of networks, excitingly-called Low Power Wide Area (LPWA) Networks, come in. In the case of Arqiva, the technology is provided by a company called Sigfox, who estimate that they can provide a 3-4m device network for an annual connection fee of around 2 euros, a fraction of what it would cost on traditional networks. In the meantime, BT have partnered with Neul, a Cambridge-based company recently acquired by Huawei to provide a test network in Milton Keynes.
Strategy Analytics forecasts that such networks may support up to 5 billion connections by 2022, similar in scale to global mobile phone connections. So what makes these new networks so different to existing mobile networks?
Module cost
According to Neul, the cost of their modules should be in the order of $4 by 2016, significantly lower than existing GPRS, 3G and 4G modules. Much of this cost reduction is due to simpler requirements, particularly lower data rates, no voice calls, simpler billing and the fact that they are designed to be used in static, rather than on the move. However it is likely that this price points are only achievable if scale in the order of millions of devices can be achieved.
Network cost
The simplicity mentioned above also brings costs benefits to the network. Additionally, such low-speed networks require significantly less spectrum than cellular networks (around 200kHz compared to 5-20MHz), or in other words between 1-4% of the spectrum required by 4G networks. As a result they can make use of narrow slivers of spectrum that cannot be used for other purposes, thereby drastically reducing network cost. In turn, the lower speeds used allows the devices to operate at much lower power than cellular networks, as well as put much lower loading on the network. This allows for coverage areas between 4-10 times that of cellular networks. Indeed, Sigfox claims that only three base stations are required to cover a city of 1 million inhabitants.
Battery Life
Smartphone battery lifetimes are typically measured in terms of hours, with heavy users lucky to get through a day without recharging. While much of this consumption is driven by the screen and the apps on the device, the radio communications back to the network still represent a significant chunk of the consumption. LPWA networks however operate in a completely different league. According to Neul, battery lifetimes approaching 10 years can be achieved on 2 AA batteries, opening up such networks to a whole new set of applications.
Conclusion
As we have seen, the new networks for the Internet of Things promise radically different economics to what is currently achievable. If they do deliver on expectations on price, battery life and reliability, this will not only allow them to compete in existing applications, such as smart metering, but open up a whole new range of applications, particularly in the areas of smart cities, agriculture, energy transmission and logistics. While this will only happen if significant scale is achieved, the potential applications offer such large numbers of connections that it is looking increasingly likely that these types of networks will become an intrinsic part of the connectivity fabric of the future.