The Myth of NB-IoT

Screen Shot 2017-02-08 at 9.30.54 AM

I ran across an analyst piece on the future of LPWAN’s and it does a better-than-average job of providing some realism about the future of the cellular industry’s best hope at penetrating the LPWAN space: NB-IoT.

LTE is a notoriously expensive and battery-unfriendly technology that may be OK for certain “M2M” IoT applications where mains power is available, but for the low powered endpoints which will make up 90% (or more) of all future IoT endpoints, LTE was never a serious option since form factor, price, and even battery life on a brick-sized device was still poor. But the LTE folks have been working hard and appear to be on the cusp of … something. And like so many cellular industry promises of the past few decades (anyone remember CDPD?), the NB-IoT promise conforms to the pattern of over-promising, delay, and most likely, under-delivering.

No sane person would underestimate the potential of the cellular industry to figure out some good-enough solution (at least) to the IoT threat, but as the article states, we are years, maybe even more than a decade, away from this. And that assumes the promises are real. Assuming for a moment that NB-IoT silicon can actually price at $1 per part (note the Huawei reference) and that endpoint battery life will be measured in the 3-5 year range, which customers will prefer a carrier-anchored solution like this over comparably priced and comparably performing solutions from Semtech, TI, and others operating in unlicensed bands? Will we really see (pricey) cellular base stations — untethered from carriers — placed on campuses and in industrial facilities like the LANs and WANs of today?

The author argues – erroneously – that NB-IoT will eventually supplant all the various LPWAN standards now competing in the unlicensed bands. NB-IoT, when it arrives and works out its kinks, may find a home with consumer and SMB customers who seek the ease of use that a carrier might be able to solve for. This alone is a big market and lots of exciting opportunities here for everyone including folks from the unlicensed LPWAN world like LoRa. That said, this is not where the bulk of the LPWAN market lies and it’s worth attempting some market segmentation now rather than to engage in (some people’s) continuing hope that somehow the cellular folks are going to sort all this IoT standards stuff out for us proles. Au contraire, there’s a better likelihood due to the sheer number of endpoint installs that will dwarf NB-IoT in the short- and medium-term of the opposite: an unlicensed band LPWAN emerging as the de facto LPWAN standard and the cellular industry conforming where it can.

Fortunately, Haystack provides a robust networking stack that rides on top of NB-IoT (really well suited for NB-IoT, actually) as well as LoRa and other LPWAN radio technologies. You can read more here and here. So developers don’t have to worry about “picking wrong” in their LPWAN decision if they use a universal/unified networking stack like Haystack.

UPDATE: here is the usually-carrier-friendly Cisco opining on the poor futures for NB-IoT in the enterprise. Deductively, if you remove enterprise (and industrial), and you assume government/defense follows enterprise/industrial in their adoption behavior, this leaves … hmm … consumer markets for NB-IoT to fight over.


Leave a comment

Your email address will not be published. Required fields are marked *