What’s the best IoT technology to replace 2G and 3G based solutions?

What’s the best IoT technology to replace 2G and 3G based solutions?

Many operators have completed or are busy sunsetting 2G (GSM/GPRS) and 3G (UMTS). End user organizations facing 2G and 3G sunsetting needs to go through their IoT use cases and select suitable replacement technologies or end up scrambling to handle communication outages and features unexpectedly missing.

Summary

No single IoT connectivity technology is optimal for all use-cases. They all come with pros and cons where some of the cons can de-rail an IoT project. At the core it’s a compromise between cost, coverage, power consumption and capacity. To further complicate things, different MNOs (Mobile Network Operators) may choose to implement different versions of the technology standards as well as selectively implement the features in the standards. Hence, it’s imperative to fully understand the current and future IoT use-case(s) as well as their current and future footprint(s). Mapping the use-cases to the optimum technology usually benefits from leveraging a knowledgeable third party.   

What are the cellular IoT technology options to replace 2G/3G on a high level

There are really only three options that are based on the 3GPP standards evolution. They all come with pros and cons. It’s imperative to understand the IoT use-cases in order to select the best possible technology. Selecting technology first may lead to challenging surprises later on when needed features may be totally missing or only available in part of the footprint. The three technologies are:

  • LTE Cat-1 – part of the original LTE specification (3GPP Release 8) frozen in December 2008.
  • NB-IoT – a LPWAN (Low Power Wide Area Network) standard introduced in 3GPP Release 13 (LTE Advanced Pro) frozen in June 2016
  • LTE-M – a LPWAN standard also introduced in 3GPP Release 13

The GSMA uses the term MIoT (Mobile IoT) that refers to the LPWAN technologies using licensed bands.

Figure 1 below shows the 3 technologies rated across 10 capabilities

Looking at the challenges with LPWAN technologies compared to traditional LTE

It’s important to understand the most important difference between LTE Cat-1, which was part of the original LTE specification and supported by all MNOs having an LTE network as well as all of their roaming partners, and the 2 LPWAN technologies. LTE Cat-1 is universally supported by approximately 800 operators on a global basis and uses the same network resources and protocols that a normal LTE capable smartphone. The two LPWAN technologies on the other hand are in layman’s terms “mutilated” compared to the original LTE standard in order to deliver chipsets with up to 80% lower Bill of Material (BoM), a significant reduction in power requirements as well as significantly better geographic and indoor/underground coverage compared to full-fledged LTE devices. The reduction of complexity and added features in the two LPWAN technologies means that MNOs need to allocate specific resources in their networks to support NB-IoT respectively LTE-M. LPWAN uses different modulation, single antenna and repetition to achieve better indoor and underground coverage compared to traditional LTE. Some of the challenges an end user organization can meet when deploying LPWAN technologies include:

  • There is frequently a discrepancy between what the standard supports and what a specific MNO has chosen to support for each technology.
  • MNOswith mobile network equipment from multiple vendors (due to sourcing strategy, acquisition or operating a shared network with another MNO may not be able to roll out LPWAN across their entire network at the same time as some equipment may require a time consuming and costly HW refresh to support LPWAN in addition to the SW upgrade.
  • Roaming is still nascent when it comes to LPWAN except within a couple of operator groups or alliances.
  • eSIM and iSIM is not fully supported until the eSIM standard SGP.32, that was published in May of 2023, is implemented which is expected to happen by end of 2024 across the ecosystem.  

There are currently approximately 140 NB-IoT and 120 LTE-M networks live. In North America, Western Europe and developed countries in APAC, it’s increasingly common that MNOs support both technologies in their networks. The total number of MNOs providing any type of LPWAN technology is approximately 160, which is around 20% of all MNOs providing an LTE services today. 

Looking at the differences/challenges in detail

In figure 1 above, the three technologies are compared on 10 different capabilities that are more or less important in different use cases:

  • Battery Life – The technology that typically consumes the least energy in PSM (Power Save Mode) is NB-IoT that can consume less than 1 micro A. LTE-M modules typically consumes 1.5 – 1.8 micro A. A typical LTE Cat-1bis module in sleep mode consumes 160 micro A, that is 200 times what a NB-IOT consumes.  
  • Coverage – Both LTE-M and NB-IoT have better coverage compared to Cat-1 due to different modulation and repetition. This becomes very visible when the device is deep indoor or underground. Most operators typically put NB-IoT in the best possible spectrum so in real life situations NB-IoT typically have an advantage over LTE-M.
  • Throughput – NB-IoT operates in GPRS (2G data) like speeds 26 Kbps downlink and 17 Kbps uplink theoretically (single-tone). LTE-M is significantly faster with 1 Mbps in both down and uplink theoretically. LTE Cat-1 is the fastest technology with 10 Mbps downlink and 5 Mbps uplink theoretically, enough for streaming video.  
  • Latency – NB-IoT have a typical latency of > 1.6s. LTE-M and LTE Cat-1 have “traditional” LTE latency around 50 ms.
  • S/W updates – The downlink speed of NB-IoT makes it less suitable for OTA (Over The Air) upgrades of SW and FW. LTE-M have enough throughput to handle medium size OTA updates. LTE Cat-1 can easily handle big OTA updates.
  • SMS – Although it’s technically possible to support SMS over NB-IoT, hardly any MNOs support it currently. Most operators support SMS over LTE-M and when it comes to LTE Cat-1 it’s a mandatory part. SMS is the low-tech way to get an eSIM profile onto a device with the current eSIM standards.
  • Voice – Voice in an LTE (as well as 5G) network is packetized and the standard is called VoLTE (Voice over LTE). VoLTE is not supported in NB-IoT. VoLTE is part of the LTE-M specification but not all operators support it. VoLTE is a mandatory part of LTE Cat-1.
  • Mobility – NB-IoT was designed for stationary devices and don’t support switching to a new cell if the devices is moving. This is changing in NB2 but at low speeds. LTE-M and LTE Cat-1 have full support to move to a new cell when the device moves, even at high speeds.
  • Roaming – Roaming is nascent when it comes to NB-IoT, somewhat better when it comes to LTE-M. LTE Cat-1 is covered by “every” LTE roaming agreement in place between MNOs.
  • # of MNOs supporting the technology - The number of LPWAN networks (both NB-IoT and LTE-M) have been growing slowly over the past 5 years. There are currently more NB-IoT networks than LTE-M networks but there seems to be a higher momentum with LTE-M. In North America, Western Europe and mature countries in APAC, it’s common to have MNOs with both NB-IoT and LTE-M networks. There are a few countries in LATAM that only have LTE-M. There are a number of large countries like China, India, KSA, Russia, South Africa that only have NB-IoT networks currently. There is a large number of countries, probably around 50% that don’t have any LPWAN networks currently. There are about 800 MNOs supporting LTE Cat-1 on their LTE networks which means that if a country has an LTE network, it’s almost guaranteed that LTE Cat-1 will work.  

Non-3GPP based technologies

LoraWAN is the most obvious alternative for regional or countrywide implementations. It operates in license free spectrum like 2.4 MHz and have similar characteristics as NB-IOT.

For local requirements BLE (BlueTooth Low Energy), Wi-Fi (predominantly 802.11ah dubbed HaLOW) and ZigBee may be an option for select use cases.

Future evolution of the 3 technologies

All three technologies have an evolution forward:

  • The next version of NB-IOT is formally named LTE Cat NB2 and offers higher throughput of 127 Kbps down and 159 Kbps up theoretically, better positioning and the ability to move between cells at low speed (introduced in 3GPP release 14).
  • The next version of LTE-M is formally named LTE Cat M2 and offers higher throughput of 4 Mbps down and 7 Mbps up theoretically (introduced in 3GPP release 14).
  • The next version of LTE Cat-1 is named LTE Cat-1bis and basically offers identical performance as LTE Cat-1 but using only one antenna which lowers the price and reduces the size of the module (introduced in 3GPP release 13). LTE Cat-1bis is designed to work on any existing LTE network without changes.

Both NB-IoT and LTE-M are carried forward into 5G unchanged and supported in 5G NR (New Radio). In 3GPP release 17 the concept of NB-IoT over NTN (Non Terrestrial Networks – i e satellite) is introduced. In release 17, RedCap (Reduced Capability) was introduced, not to replace NB-IoT and LTE-M in 5G, but to provide higher speed than LPWAN at an affordable cost compared to 5G eMBB (enhanced Mobile BroadBand) modules.

There are “combo” models available from many of the IoT module manufacturers that support both NB-IoT (typically NB2) and LTE-M (typically M1) which offers greater flexibility and geographic coverage as it can switch depending on requirements and network availability.

Where to seek guidance on technology selection

Engaging with a knowledgeable 3rd party to ensure the optimum technology is selected for the specific use-case and its footprint is a best practice. These 3rd parties can be found in these four categories of organizations:

  • IoT connectivity service providers like leading MNOs and MVNOs (Mobile Virtual Network Operator) with a strong IoT practice can advise on technology selection and know exactly the capabilities of their local OpCos and respective roaming partners on a country-by-country basis. In addition, they usually have a sizeable ecosystem of vertically aligned partners. Some of them can also project manage complete pilots or production roll outs.
  • IoT Module Manufacturers like Quectel, Semtech/Sierra Wireless and Telit/Cinterion to name a few.
  • SIs (System Integrators) with sizeable IoT practices
  • IoT focused Technology Advisory organizations like Transforma Insights, CCS Insight, Berg Insight and Lionfish Tech Advisors to name a few.

A historical outlook on LPWAN

When it became obvious back in 2015 that what we now know today as NB-IoT was not likely to make it into 3GPP release 13, an alliance predominantly with Vodafone, Huawei, Ericsson and Nokia, basically took it out of the 3GPP standardization process, completed the standard and fed it back into the 3GPP process as a change request to release 13. This “sprint” effort worked and NB-IoT became an official part of 3GPP release 13, dubbed LTE Advanced Pro. A limited number of MNOs hedged their bets and choose to implement LoraWAN instead of waiting for NB-IoT (most notable MNOs like Orange in France, KPN and Swisscom).

Initially, LPWAN implementations in North America was predominantly based on LTE-M and in Western Europe and China almost exclusively based on NB-IoT. These days, most of the leading operators support both of the LPWAN technologies in their networks to better meet diverse customer requirements.

Recommendations

By now, organizations should have inventoried all their remaining 2G and 3G connections and have identified the use-cases they support. Fully understanding the use-cases and how they’ll evolve from a functional as well as geographic perspective is key. Once you have each use case documented:

  • Map it to the optimum technology and take into consideration the dual mode modules that exist on the market.
  • Validate the technology selection using 3rd party organizations that can ensure that operators in the desired footprint can support the required functionality.
  • Factor into the decision that a single truck roll to fix a problem in production may change the TCO calculation dramatically. In many situations, it may be prudent to go with a slightly more expensive module that have more capabilities.
  • If you have the power budget for a LTE Cat-1 module, it’s usually the most versatile and “safe” bet.

The sweet spot for the different technologies can be summarized as:

  • No external power source and stationary position deep indoors or underground. Typically a good candidate for NB-IoT or a dual mode module that operates on NB-IoT most of the time and can switch to LTE-M for SW and FW OTA upgrades. Sensors in water and sewage mounted underground would be a great example.
  • No or limited external power source, moving around at modest speeds, requirements for voice and SMS, support for eSIM in a limited footprint. Typically a good candidate for LTE-M. Track and trace of trucks, trailers and construction equipment would be a great example.
  • External power source, requirements for a large and unpredictable footprint, moderate to high data rates, eSIM, voice and SMS support. Typically a good candidate for LTE-Cat1 and Cat1-bis.

Sources

This article is based on publicly available information from mobile operators, IoT module manufacturers and the GSA (Global mobile Suppliers Association) as well as my own analysis.


To view or add a comment, sign in

More articles by Leif-Olof Wallin

Insights from the community

Others also viewed

Explore topics