New Sony LPWAN Deja Vu

Sony is entering the LPWAN radio space with “Eltres”, giving new competition to Semtech, SigFox, NB-IoT, and the rest. Some brief reactions:

Pros

• Haystack’s analysis of the physical layer (radio) is: this is the way a LPWAN radio should be designed! Well done.
• LDPC error correction, which Haystack previously announced for use over Semtech’s LoRa, is good forward error correction and may provide some differentiation vis-a-vis LoRaWAN, which is unlikely to ever support such high-grade error correction.
• Eltres appears focused on mobile LPWAN use cases. Here at Haystack we agree the mobile LPWAN segment is an underserved and very large market that neither LoRaWAN, nor Sigfox, nor NB-IoT is able to address with any seriousness. So nice to see Sony endorse our view of the market.
• Eltres includes GPS. Sony has its own GNSS/GPS receiver product so marrying it with the radio is a natural for them.
• Even though it is a uni-directional protocol today, the specification itself provides at least theoretically for a future bi-directional “upgrade”.

Cons

• Battery. Battery life according to the article is “possibly months.” Possibly? A previous bit of Eltres marketing info we found said six day battery life, which reminds me of this Samsung GPS tracker using NB-IoT, which hasn’t exactly generated a ton of press. Even if we assume it is “weeks,” if you are going to enter the LPWAN space, you should have the marketing team talk about — or the engineering team build for — years and worst case many months or it’s really not low power.
• The 1-Way Problem. Eltres is a one-way/uni-directional protocol. Uni-directional protocols are vulnerable to very high packet error rates in mobile (and fixed!) environments that even with LDPC error correction presents huge challenges. Eltres re-transmits each message four times in the hopes that if the first three messages didn’t arrive, the fourth will! Oh well — at least they don’t hide this approach. Regardless, the focus on mobile for Eltres is a huge question mark given this vulnerability alone.
• More 1-Way Problems. A unidirectional protocol results in weak security, no ARQ in error correction, difficulty in updating firmware, no multicast, no P2P, no ability to query endpoints … pretty much just like LoRaWAN or Sigfox.
• Weak Geofencing. A unidirectional protocol can only deliver a mediocre geofencing feature, at best. As mobile LPWAN feature requirements go, this is a pretty important one that Eltres cannot really address.
• No A-GPS. Sony’s unidirectional solution precludes the use of Assisted GPS to minimize power draw and location acquisition times.
• A so-so GPS component. We tested the standlone Sony GNSS/GPS receiver and it did not meet our requirements.

Overall:

• The market can benefit from a better LPWAN radio option but not another uni-directional one!
• When Sony gets around to rolling out its two-way radio, Eltres may be more compelling to developers and Sony may have a real shot at addressing the mobile LPWAN market. Until then, this product will have difficulty distinguishing itself from other one-way LPWAN’s.
• The “60 mile” long range claim is driven in part by LDPC error correction but hugely impacted by placing devices at high altitudes — I believe one may have been near the summit of Mt Fuji? — which is a time-tested way of maximizing range for marketing purposes that have little to do with real-world implementations.
• The addition of GPS is clever but at the cost of very short battery life. At Haystack we are radio agnostic but I think it is unlikely we will spend much effort on Eltres in its current incarnation. #

Get the latest about Haystack by signing up for our newsletter here.