The development and deployment of cellular machine-to-machine networks will lag that of mobile broadband by at least five years. The two families of LTE-based IoT-centric specifications, NB-IoT and LTE-M, are only being commercially deployed at scale now, so although they will have 5G upgrades standardized in 2021’s 3GPP Release 17, most MNOs will wait for some years to roll these out. If they are deploying NB-IoT and/or LTE-M now, they will not want to implement another new technology, even one that is mainly a software upgrade to the main RANs, until they see return on investment, or until compelling use cases emerge that really require 5G.
The lag between broadband and low power M2M is unsurprising. When the 4G standards were conceived, the need for a low power, wide area network (LPWAN) to support M2M applications beyond GSM was hardly recognized. There was virtually no commercial impetus behind these standards until a few years ago, when systems in unlicensed spectrum, such as LoRa, emerged to address the new opportunity. At that point, low power 4G specs were dusted off, and new ones defined, leading to the creation of the two families of specs (LTE-M is slightly higher-powered, for devices such as wearables, while NB-IoT is a true LPWAN).
But there is a need for 5G standards to be defined. The 3GPP community needs to avoid a repeat of the situations it has faced before, when it has not pre-empted demand for new services and left MNOs outmanoevered by alternative technologies while they waited for their new specs (LoRa versus NB-IoT is one example; an even more serious one is Voice over WiFi versus VoLTE, since voice was inexplicably sidelined in early LTE standards processes).
The 5G upgrade to NB-IoT will not appear in the next release of 5G standards, 3GPP Release 16, which is due late this year. Instead, it is on the program for Release 17, which should be frozen in June 2021, driving commercial shipments about a year later. At this point, an evolution of NB-IoT will be part of the 5G massive machine-type communication (mMTC) specs. Their main aim is to support even larger numbers of sensors and devices than 4G does, and to address the ITU’s IMT-2020 ‘massive IoT’ requirement, which demands one million devices square kilometer (compared to 60,680 devices for current NB-IoT).
In the nearer term, there are moves to extend coverage of NB-IoT to achieve the true ubiquity that some M2M applications require. This need for ubiquity is driving the increasing convergence of satellite and radio technologies, particularly in satellite-over-5G developments that are largely geared to ultra-wide area and remote IoT.
One start-up, Skylo, is targeting the pre-5G satellite IoT opportunity, emerging from stealth mode with $116m of investment to pursue its plan to build a global NB-IoT network using existing satellite infrastructure. Its $103m Series B round was led by Softbank.
Skylo says its network is hosted on existing geostationary satellites, “enabling immediate connectivity for customers, without the need to add new infrastructure in space”. It claims that this provides industries like agriculture, logistics, and utilities with an easy way to connect IoT sensors “for the first time” (clearly untrue).
It has not yet talked about existing customers or pilots, although it does claim that it is 95% cheaper than existing satellite solutions, costing $1 per user per year, with hardware costs less than $100. However, these figures show it comparing itself against the conventional crop of satellite operators, and ignoring the emerging gang of Low Earth Orbit (LEO) nanosatellites that have generated a lot of buzz in driving down IoT costs.
Some of those rivals are using non-standard protocols, but others are exploring how to use LoRaWAN (usually via ground terminals). Sigfox is in on this action too, via a partnership with Eutelsat, and while we thought it wouldn’t be long before somebody tried adding licensed LPWAN to the mix, we had assumed it would be via LEO nanosatellites, rather than big fat GEO sats.
Another player is the USA’s Ligado, which is trying to build its own Cat-NB satellite network. That would differentiate it from Skylo, but it is still trying to secure FCC approval to use its L-Band (1.5 GHz to 1.7 GHz) spectrum for cellular services, and has recently been rumored to be on the verge of running out of money.
The Skylo Hub is the end device in the new network – an 8-inch square that might be all of an inch tall, the white plastic unit aggregates data and beams that data up to the satellites via Cat-NB. This data is then beamed back down to the ground units that connect these satellites to the wider internet, and then into Skylo’s Data Platform – the cloud environment where you can view the sensor data and hook up APIs (application programming interfaces).
The initial target markets on Skylo’s website are disaster management, maritime, fleet management and agriculture. It points to its hubs being IP67-compliant, as proof of their suitability for deployment in rugged field conditions. For vehicular use, Skylo is also claiming location accuracy of less than three meters.
However, that’s about as much information as Skylo wants to share publicly. There is mention of Bluetooth and OBD-II, for connecting things to the hub, but there must be more I/O ports to cater for the massive variety of sensors out there on the market.
It says it has tested the network with early commercial partners, but won’t say who, only that they were in both enterprise and government. LightReading says that Indian state telco BSNL is working with the firm to connect polling stations, and that Skylo has tested connections to Indian fishing boats and will boast Indian Railways as its first commercial customer.
With offices in California (HQ), India and Israel, and founded in 2017, it plans to begin global roll-out this year.
Of course, the NB-IoT approach is going to encounter many of the same problems that conventional satellite deals with. Line-of-sight with the sky is going to be necessary, as we severely doubt it is going to be any good at indoor or underground propagation, and the battery life of the Hub itself is a big wild card, as it is going to need a power supply to act as a NB-IoT gateway for nearby devices. In theory though, it shouldn’t need more power than regular NB-IoT communications, as there’s a lot less interference firing beams straight up into the sky than trying to blast them through obstacles at ground level.