Your browser is not supported. Please update it.

20 October 2020

Low power WAN will complement, not compete, with 5G

The LP (low power) WAN protocols are sometimes misleadingly presented as clear alternatives to cellular, when increasingly they will be deployed alongside and be seen as part of an expanding portfolio of mobile communications options. There are cases of course where LPWAN protocols will be deployed in standalone mode, in some agricultural and campus settings, but even there they will often either be provided as part of a public cellular service, or in parallel with a private 5G network.

In fact two of the leading LPWAN protocol contenders will operate in licensed spectrum as part of the 5G effort to support mMTC (massive machine-type communications). These are 4G-based NB-IoT and LTE-M, and their future 5G iterations (though there is some doubt about the long term survival of NB-IoT’s roadmap, especially after NTT Docomo switched off its network based on this technology).

These two will, between them, account for at least 60% of global LPWAN IoT connections between them by around 2025, perhaps as much as two-thirds. That will leave one third or a little more for those LPWAN protocols that operate in unlicensed spectrum although with ability to interoperate with public networks, where LoRa and Sigfox are the two leading protagonists.

The 5G-based standards for LPWAN are only in the early stages of being ratified, with some functions included in the recently finalized Release 16 and others to come in future releases. In reality, operators which are deploying LPWAN now, or in the coming couple of years, are relying on 4G-based, and/or unlicensed technologies, and the timing means these are often part of a ‘5G’ deployment, with the LPWAN roll-out designed to complement a 5G mobile broadband network.

Although the LPWAN protocols can be perceived as belonging to a ‘5G’ group of technologies, they serve quite distinct use cases from those commonly associated with data under cellular. They cover relatively large areas for often intermittent low power and low bit-rate communications, yet  where there may be large numbers of devices and need for scalability. LPWAN is also distinct from those more local LAN protocols that support both high and low bit-rate communications, such as WiFi at the top end, down to the very short-range ZigBee, Z-Wave, Bluetooth Low Energy (BLE) and Thread at the bottom.

LPWAN, as well as to some extent the lower power LAN protocols, need to be analyzed in the context of an expanding IoT world and growing demand for applications under that banner of mMTC. While there are some common features, this covers a wide range of scenarios where LPWAN must be able to adapt, which is one reason there will continue to be several alternative protocols, including those two under the 5G umbrella.

Common features or expectations include low device cost below about $6, a range of data packet sizes per transaction from 10 up to around 1,000 bytes, quite modest uplink speeds up to around 250Kbps, long battery life of at least 10 years, high coverage range up to 1,000 kilometers (although most commonly just a few kilometers), and finally capability to support big device populations potentially up to 100,000 or even a million.

Against this common requirement set are some huge variations in coverage, capacity, bandwidth cost, delay tolerances and other factors. It is true that LPWAN is targeted chiefly at a subset of the overall mMTC field, some of whose use cases will be met by full blown 5G, or even WiFi when only local coverage is required.

LPWAN’s emphasis is on coverage, large numbers of low-cost devices, high energy efficiency, and comparatively narrow bandwidth range up to a low ceiling, but even within that there are significant variations in demand. The data itself may vary in terms of message number and size, as well as the reliability required.

Some applications, such as smart metering, require security and robustness but are very tolerant of delay, while others such as home security or detection of dangerous hazards like radiation leaks may require high priority and low delay. Some will trigger actions, others will not.

There may also be a need for a single LPWAN network to support multiple use cases and devices types, each of which may have different SLAs (service level agreements) overlaid. Support for different communications speeds and delays over both the uplink and downlink may be needed, with some devices being mobile and others static. It could be that more than one LPWAN protocol will need to coexist over the same infrastructure to support this. We have discussed the pros and cons for the main LPWAN protocols before and there is also a good discussion from the IETF (https://tools.ietf.org/id/draft-ietf-lpwan-overview-09.html).

Meanwhile, some of the technical consideration underpinning LPWAN deployment decisions are covered in a comprehensive paper from Future Internet (LPWAN Technologies: Emerging Application Characteristics, Requirements, and Design Considerations. Future Internet 2020, https://doi.org/10.3390/fi12030046.) This paper identifies and discusses the wide range of use cases where LPWAN will find application because of the need for low bit-rate connectivity spanning large areas with ability to handle a massive number of connections. These will include eHealth, smart cities, energy utilities, agriculture, automotive and industrial.

There is however one aspect of LPWAN that is often not fully addressed in such papers, which is the relationship with 5G. This has at least three aspects:

  • First there is the genuine coexistence aspect, which could be particularly relevant for smart city and also some campus nor industrial cases. LPWAN for example would score for applications such as the smart metering, already mentioned, as well as environmental monitoring, smart lighting and also intelligent parking, entering domains currently served in some cases by 4G. Then the applications requiring higher data throughput and sometimes low latency would be served by what we might call 5G proper, including CCTV surveillance, as well as autonomous drone operation in security and traffic monitoring.
  • The second aspect of the 5G relationship is LPWAN backhaul. Like all access technologies, LPWAN will require backhaul and that can be provided by 5G, often within a common service or infrastructure. This is the case regardless of the use case, but may be particularly relevant in the industrial, agricultural and smart city sectors. The relevant points are firstly that cellular is too expensive and power hungry to serve many of these use cases directly and secondly that it may not be available at all the desired locations. However, cellular might be more readily available than say wired connections for backhaul and would then provide the IP-based capacity to take aggregated data back to cloud or central platforms.
  • The third aspect of the cellular/LPWAN relationship is perhaps more of a direct replacement, addressing some situations where legacy 3G or even 2G services eventually become obsolete. Clearly they will be upgraded to 4G and 5G in many cases, but there will be situations where that is nether affordable nor necessary.

Spectrum will be reallocated away from 2G/3G and in any case devices will no longer support it. Yet LPWAN protocols will prove perfectly capable of supporting those use cases that do not require higher bit-rates or ultra-low latency, while being cheaper and more energy efficient, increasing battery life.

In some cases 2G and 3G will be replaced by the 5G LPWAN options of NB-IoT or LTE-M in any case and so stay entirely under the cellular stable. Indeed, Ericsson made the point in a recent blog that “engineers should start considering LTE-M and NB-IoT as the foundation of their future IoT developments”. The blog presents these as straightforward replacements for 2G/3G and a simple upgrade path to fully integrated 5G LPWANs in years to come.

Ericsson pointed out that while, less surprisingly, 2G has already been terminated across North America, Oceania, and much of Asia, leaving just 3G of the older cellular generations, in Europe it is often the other way round. There, “it’s likely that 2G will have greater longevity and 3G will be phased out”, the blog stated. This is because there are sufficient IoT nodes already installed in Europe dependent on 2G connectivity to justify retention of that for now, before being upgraded later to either LTE-M or NB-IoT.

So it can be seen that LPWAN is emerging from the chrysalis of cellular as well as an alternative or complement to it.