26.01.17


Where next for IoT connectivity?


2G Sunset, LoRa vs. SigFox vs. NB IoT vs. LTE CAT 1M… it’s easy to see how the IoT connectivity market is becoming tricky to navigate. 

A global patchwork of legacy network switch offs, a blizzard of new technologies, open vs proprietary standards – together with new start-ups and incumbent operators all arguing that they alone have the key to the future. It’s little wonder companies are hesitating about which direction to take for long-term IoT connectivity. So, what are the main options? 

So long 2G?  

For the last decade, the GSM 2G / 2.5G GPRS standard has been pretty much the global workhorse for IoT. For large numbers of sensors generating small amounts of data over a wide area, 2G has advantages over later rivals, including national and global coverage, lower-cost devices and affordable transmission costs. 

Many of these features made it ideal for the telematics industry for example, but 2G has always been slightly over-engineered for simple uses such as smart-city infrastructure. A few bytes of data are all you need to know if a fridge is at the right temperature, a gate has been opened, a streetlamp is switched on or a dustbin needs emptying. 

The full-strength 2G GSM mobile standard has some significant downsides: including the effect on battery life of the high-power radio, the chatty long-winded set up for a data session and the licence fee on top of the manufacturing cost of the modules. In the context of current and emerging IoT requirements, and with its demise planned or underway in many countries, it’s easy to see why 2G’s time might be up for future long-term use. 

Getting over the 3G hump  

Gradual retirement of 2G means that IoT businesses are faced with choices about which technology should replace it. 

3G investment is now frozen as 4G and 5G technologies are being developed, however 4G and 5G, designed for the rich media smart phone user, with high device costs and power hungry modules, mean they are not really ideal standards for low data IoT use. 

So, what are the considerations and appropriate choices for IoT now?  

Best for heavy IoT data users -  for Fleet and UBI Telematics, or Telehealth for instance, LPWA doesn’t offer blanket geographical coverage, fast enough data throughput or (for Telematics) seamless international roaming. Here 4G can step in when 2G exits and it offers great advantages, such as coping with the demand for data driven by the burgeoning growth of streaming video cameras and other data-hungry components. 

Best for lighter data users: With every widely distributed, light-data, long-lifecycle deployment from smart cities to agribusiness, some type of LPWA is ideal. The current dilemma is whether to go for a proprietary unlicensed-spectrum LPWA solution (SigFox, LoRa etc.) or wait 6 to 12 months for the promised 3GPP licensed MNO alternatives, ‘Cat1M LTE’ and ‘NB IoT’, to be realised. Looking at these options in more detail:

LPWA 

Low Power Wireless Access (“LPWA”) solutions offer extremely low hardware and operating costs and long battery life. However, they also come with tiny data throughput, a lot of closed proprietary technology in silos with patchy geographical coverage. 

The low price points make LPWA an attractive option for smart technology applications, such as dustbins, but the downsides above need to be taken into account. 

LPWA solutions offer numerous options including DIY vs. third party-provided; and “open” (ish) vs. proprietary standards. The main players are: SigFox, LoRa and Ingenu: 

 

  • SigFox – perhaps the most promoted brand in this market is effectively a single network provided by many national third parties, “Sigfox Network Operators”, such as Arqiva in the UK, who are slowly increasing international coverage. 

Sigfox make the technology standard openly available on licence to module manufacturers, and charge end users for access to the network. With pricing as low as €1.00 per device per year this can be much lower than current GSM alternatives, but minimum volume / revenue requirements mean SigFox only costs in for very large deployments, and the extremely low bandwidth and throughput restrict it’s target uses to the very low end of data use such as utility metering. 

 

  • LoRa – is a proprietary technology belonging to Semtech Inc. with the standards set by an independent body the LoRa Alliance. LoRa offers similarly low cost network access plus it offers greater business model flexibility. National networks have been built by large MNOs such as South Korea and the Netherlands (by KPN), but the wider range of LoRa deployments to date are private networks built by and for Smart Cities and large enterprise “campus” environments such as goods yards, ports, factories or universities. 

Semtech owns the technology IP and manufactures the modules (although starting to offer licences), however these are far cheaper than even 2G GSM modules, so the economics work from medium to very large scale deployments such as utility metering. 

 

  • Ingenu – has a business model that is even more proprietary, insular and vertically integrated, as they produce the modules, access points, build and manage the entire network based around a patent technology RPMA which seeks to more efficiently manage the spectrum-use, bandwith and throughput vs. other LPWA standards. 

In the USA, Ingenu have a publicly accessible IoT network, but their main overseas uptake is in private networks for the utilities & energy industries.  

GSMA 3GPP

With the launch of the latest, much anticipated GSMA Standards under 3GPP the “big mobile” industry’s open IoT standards aim to offer IoT businesses friendly pricing by: 

 

  • avoiding using expensive licensed hardware
  • exploiting existing infrastructure and spectrum
  • addressing battery life by simplifying data session set up etc. 

The speed at which the industry has coalesced to agree two global, open standards under 3GPP is perhaps indicative of the threat posed by LPWA. The new standards are known as: 

 

  • ‘LTE Cat M1’, which has been ready to ship from Q4 2016
  • ‘NB-IoT’ which is being finalised for launch later in 2017.

LTE Cat M1 ought to be a winner, as compared to the proprietary LPWA solutions, LTE Cat M1 offers the prospect of a rapid build-out to blanket coverage, similar battery life (c. 10 years), greater bandwidth and many times more data throughput capacity (up to 1 Mbps). 

Since Mobile Operators only need to drop a software upgrade into their 4G networks, LTE Cat M1 should rapidly become the most widely accessible dedicated IoT solution with national and global coverage – moving almost instantly up to the level of the best 4G voice & data networks. 

Devices will be slightly more complex (requiring a SIM, although possibly e-SIM) and therefore not quite as cheap as the LPWA alternatives, however LTE Cat M1’s other benefits are clear. The operators are well known, large scale, safe bets for the 20-year+ longevity of fixed remote infrastructure like substations or lamp-posts. Global enterprises and public utilities will find it easier to trust them with their critical infrastructure monitoring. 

We are yet to see 3GPP NB-IoT modules arrive, but given the specification as it stands, as it only offers marginal battery life improvements and a far smaller data throughput - compared to LTE Cat M1, (250 kbps vs. 1 Mbps) – so NB-IoT may struggle to take hold in the market when it arrives in late 2017. 

So where next?

Every business looking at a long term IoT deployment needs to arrive at a carefully costed and considered conclusion, but I am clear where my bet is going. 

I love the idea of upstart independent IoT technologies taking on the vested interests of Big Mobile. I can also understand the economic arguments, especially for a wholly owned campus network, but the history of global open standards winning out over closed proprietary technologies, combined with the compelling permanence of the Tier-1 MNOs backing LTE Cat M1, makes it the safe bet -  if not quite a sure-fire winner!