IoT Blog

IoT Blog

NB-IoT and LTE-M Interoperability Forecast: Blue Skies

by Gus Vos, Chief Engineer, Technology Standards
With NB-IOT interoperability in the news lately, I thought it would be a good time to sort through the facts of the matter, including a brief history lesson of how the NB-IOT and LTE-M standards were developed. The two similar technologies traveled very different standardization and interoperability paths that, as we’ll see, resulted in some interesting side trips.

Interoperability 101

Interoperability Device testing (IoDT) plays a critical role in new cellular technology rollouts, where chipset vendors provide very early products to infrastructure vendors to test. The first challenge is that the definition of when interop is “done” is NOT standardized, but most companies feel that when the first “couple” of chipsets work with all the infrastructure vendors, interop is “done”. This the definition that I’m using here in this blog.

The main hurtle comes from ambiguous specifications. Specifications are written in English prose, for the most part, so, like poetry, readers can interpret them in different ways. When there is a disagreement, who decides what’s right and, more importantly, who will burn the resources to change their design can be a costly game.

In addition, IoDT can be tricky because it relies on cooperation between an ecosystem of multiple vendors that are often rivals vying to be first to market. For example, a company that makes both infrastructure equipment and chipsets could refuse to provide its early chipsets to rival infrastructure suppliers for testing, which slows the process down for those rival infrastructures but more critically could lead to interoperability issues for all.

As with most difficult tasks in life, when successful, they bring large rewards, and IoDT doesn’t disappoint. IoDT brings massive value for our customers, including the ability to scale globally, vendor choice, lower prices, and supply chain security. Proprietary LPWA technologies simply can’t provide these benefits.

LTE-M and NB-IoT Had Very Different Standardization Paths

I know cellular standards history is probably not one of your favourite subjects, but understanding the standardization history behind LTE-M and NB-IOT will help you understand some of the unique IoDT challenges. 

Let’s start with LTE-M. The very basic idea of LTE-M can be traced back seven years ago to May 2010 (See Sierra contribution S1-101154), which proposed a simpler LTE device for the Internet of Things. After this, 3GPP started a study item (see work item in RP-121441) in September 2011 that lasted a year and a half and looked into optimizing two of the three C’s needed to make LTE a fit for IoT: Cost (device and services cost), and Coverage (extended reach). The first normative work started in June 2013, with Vodafone as rapporteur (see work item in RP-130848), where CAT 0 was defined and the coverage extension work was started; the second normative work started in September 2014 (see work item RP-141660) and finished in March 2016, where CAT-M1 and the coverage extension work was completed. All told, LTE-M took about three years of normative work.

NB-IOT started its standardization life in 3GPP’s GSM group (GeRAN) with a study focusing on clean slate solutions that could be deployed in re-farmed 2G spectrum (i.e. 200KHz channels). This study was stopped in August 2015, where five high-level solutions were captured in TR 45.820. The top three most popular solutions were: NB-M2M from Huawei, NB-OFDMA from Qualcomm and NB-LTE from Ericsson/Nokia. 3GPP then conducted a five-month study in the 3GPP RAN group (the UMTS and LTE specification group), where a single general direction was agreed to in December 2015. In typical standardization style, the final high-level direction selected was based on parts of all three of the above-mentioned popular proposals: NB-M2M, NB-ODFMA, and NB-LTE (see work item RP-152284). In January 2016, the detailed normative specification work started and a mere six months later, in June 2016, the NB-IOT specification was declared frozen. To accomplish this compressed timeline, long NB-IOT sessions were conducted in parallel with regular 3GPP sessions and two, one-week-long NB-IoT-only adhoc meetings were added.

So, in summary, LTE-M was specified in about three years, whereas NB-IoT was specified in only six months.  

Interoperability Side Trips 
Before the NB-IOT specification was completed, Huawei and Neul, acquired by Huawei, were demonstrating a proprietary version of NB-IOT based on the NB-M2M proposal defined in TR 45.820. Huawei was in a unique position to be developing both infrastructure and chipsets, which allowed it to create this proprietary NB-M2M solution. When the NB-IOT standards was finalized in June 2016, Huawei was again in a unique and advantageous position to be able to do IoDT using its own infrastructure and chipset, which helped accelerate their NB-IOT development. 

With the accelerated standardization and IoDT timeline for NB-IOT, some field and interoperability issues were inevitable. For example, because NB-IOT has a brand new broadcast channel, it was discovered through field trials that the broadcast channels interfered with each other, so in March 2017 two essential, non-backward-compatible changes to add scramble to these channels were introduced (see changes in R1-1703913 and R1-1703964). Given the number of field trials now complete, Sierra Wireless is reasonably confident the major field issues are now behind us.  

In summary, for NB-IOT all the above factors, combined with the short six-month specification time, created additional IoDT challenges.

IoDT and the Mythical Man Month
Looking at 3GPP history over the last 10 years, it normally takes one to two years between 3GPP standardization freeze and commercialization (including IoDT). 

Despite this history, some eager vendors were predicting NB-IOT IoDT could be done in a mere six months suggesting the early proprietary work that was already done and the large dedicated engineering effort as a reason for the accelerated schedule. It was a very aggressive assumption, especially for an essentially new protocol that was not field proven.  And, unlike some other engineering activities, IoDT and field testing can’t be rushed by throwing more engineers on the job—the mythical man-month applies. As of about June 2017, NB-IOT successfully completed IoDT with a limited number of chipset vendors, thus IoDT ended up taking the historically normal one-year period.

Sierra Wireless took a pragmatic outlook on timing, especially when talking to customers about technology roadmaps. So from our perspective, NB-IOT interoperability is not late, it is right when we expected it to be completed.  

Surprisingly, the IoDT and commercialization of LTE-M occurred in December 2016, which means LTE-M only took around nine months, which is unprecedented for such a significant feature. However, LTE-M had a much simpler and traditional IoDT path than NB-IOT. Given the longer runway for LTE-M, the IoDT planning and ecosystem organized itself in a more typical, collaborative and harmonious way. In addition, LTE-M has several optional features (e.g. +20 EC mode B and the control plane small data optimization), which could be in IoDT later. Also, LTE-M is a much closer variant to LTE, where most of the broadcast channels are re-used. These factors, made the LTE-M IoDT process closer to the typical process. 

Forecast: Blue Skies
Commercialization of both LTE-M and NB-IOT devices are well underway. In fact, Sierra Wireless has announced a number of carrier trials, including AT&T, Verizon and Telstra, as well as the launch of the AirPrime HL77 and WP77 devices—the WP77 devices are the industry’s first to support both LTE-M and NB-IoT on a single device.

IoDT is difficult but worth the effort and, unlike some other engineering activities, it can’t be rushed by throwing lots of manpower at it. The LTE-M commercialization was ahead of Sierra Wireless’ expectations but there are still some features that need to be deployed (e.g. EC mode B and e-DRX). Although the recent news has portrayed NB-IOT IoDT as being behind schedule, that schedule was very aggressive and in our view, NB-IOT IoDT and commercialization completed on schedule.