Ericsson has rather dramatically slashed its forecast for the number of connected devices which will be in play in 2020, from its famous 50bn – one of the most quoted figures in the history of wireless – to 26bn. The reduction may reflect more sober assumptions as the internet of things (IoT) starts to evolve for real, or the almost inevitable fact that it will not explode as quickly as many forecasters have hoped – partly, according to one IBM executive, because there is no convincing IoT business model yet.
Ericsson’s latest Mobility Report represents the first time the Swedish vendor has wavered from its 50bn prediction since it first made it in 2009, but it now claims that was a “vision”, not a specific forecast, and that its new estimate is its “first connected devices forecast based on thorough analysis”. Today, there are about 13.5bn connected devices in use, says Ericsson, including cellphones, PCs, tablets and M2M devices.
However, it hedged its bets this time, adding that “the area is still new and much can change”, while insisting the market would get to 50bn, but “some time after 2020”. It is certainly right about the uncertainty of the space, as reflected in the wildly varying estimates offered by its rivals – 100bn by 2020 being the most optimistic, from ZTE – and by research firms. Gartner echoes Ericsson’s figure of 26bn by 2020, Machine cites the same number of devices but with a 2023 timescale, while IDC is looking for 28.1bn by 2020.
Growth will be below all those forecasts if significant changes are not made, warned IBM’s global telecoms industry leader, Rob van den Dam, who told this week’s CommunicAsia2015 conference that “the internet of things needs a reboot”. He said that the IoT would not scale to hundreds of billions of gadgets without major changes, because “the business models are broken”. Key barriers to growth are high cost and lack of privacy, he said, and the IoT is “not future proof”, given that a car may have a lifespan of a decade, an aircraft of 40 years and a house of even more. He believes the industry is thinking too much in terms of the rapid replacement cycles for handsets, but “you’re not going to change your doorknob every three years”, he remarked.
Another major barrier is the “lack of functional value” for the end user in many connected devices, using everybody’s favorite example, the smart toaster, to make his point. He said he could not identify any revenue model for the toaster – a user would not pay a subscription for the connectivity for for updates; they would not want to receive advertising “burnt into their breakfast”; and they certainly would not want their insurance company being notified of “heavy consumption of Pop Tarts”, he joked.
But he was making a serious point that, especially in the consumer IoT, the business models are unclear and the cost of devices and services are too high to drive significant uptake. So far, only 10% of smart TVs are actually being used for internet viewing, he claimed, while only 30% of industrial equipment is networked.
To get over these problems, the second wave of the IoT will need to adopt three architectural principles, he said, putting his IBM hat firmly on his head. These are dramatically lower cost, trusted privacy, and business model endurance, and in turn, those will be enabled by an open, decentralized, scalable and secure platform. Specifics of that might include new technologies for autonomous device management and secure distributed file sharing – the kind of innovations currently emerging from the labs of the big IT vendors like IBM itself.
“Today, we have IoT based on open access networks and a centralized cloud. In 2025 and beyond, we will need to have a distributed cloud in which the device itself is the edge of the cloud,” he predicted. With low value, long-lived devices, decentralization becomes essential, he added, saying: “By empowering devices to function autonomously, the edge will become a frontier of new economic value. You can create a new marketplace hosted by peer exchanges, and transform devices into points of transaction and value, creating an ‘economy of things’ – enabling things to participate in this new digital economy.”
Ericsson, despite its new pessimism about device numbers, is also looking towards common, highly scalable platforms to help the IoT develop. Speaking at the annual TM Forum event in France, CTO Ulf Ewaldsson said IoT operators would need ways to monitor and control cross-industry processes through a common platform, in order to tie the data from hundreds of different sensors together into something coherent without overloading the networks.
This would require a “very, very clever” architecture, he said, which would allow operators to use the same common infrastructure for many tasks, from process control to robotics. “One network is very important. The frequency resources are limited and we can’t just keep building network after network,” he added.
“We have clouds that are screaming for connectivity,” Ewaldsson told FierceWireless. “Industry after industry wants to bring software to their products, and they want to know who will provide their connectivity.”