We've been using some tracking hardware with firmware from a company called Wirepas that has implements mesh networking. They've been doing this at scale for quite some years. I their technology uses both wifi and bluetooth. They don't do hardware themselves and instead work with a large amount of OEMs. Cool company to check out and I absolutely love their website and approach to marketing. If you have a few minutes, check out their videos. They are hilarious.
The attraction of their solution is that the devices can be wireless and don't require a lot of infrastructure. You basically just mount them to the wall or the ceiling. Battery life is pretty good as well depending how you set these devices up. Years typically. Installation cost and effort is one of the big hurdles for companies to take this stuff into use.
So, interesting to see some standards for mesh networking emerging here.
Wirepas requested a meeting with us at our office several years ago. They were very polite, but in all honesty, we knew even less about implementing their product afterward, and felt as though they weren't really interested in working with /us/ at all.
Most of our questions were about implementing their firmware in our hardware (which largely went unanswered), and most of their questions were focussed companies we have ties with, that they were interested in reaching but have so far been unsuccessful.
I hope they're doing well as it was an interesting product.
They raised 22 million a few months ago. So, it looks like they are doing well. Might be worth reaching out again if that's still relevant to you. I've talked to a few of their people on a couple of occasions and they were pretty competent and pleasant to deal with.
No, one of our partners is a company called Bornemann and they are one of the OEMs listed there. So we've mostly worked with their devices and platform.
Connecting the dots from another article on HN front page today, about ESP32.
ESP-BLE-MESH Architecture
> Bluetooth® mesh networking enables many-to-many (m:m) device communications and is optimized for creating large-scale device networks.
> Devices may relay data to other devices not in direct radio range of the originating device. In this way, mesh networks can span very large physical areas and contain large numbers of devices. It is ideally suited for building automation, sensor networks, and other IoT solutions where tens, hundreds, or thousands of devices need to reliably and securely communicate with one another.
> Built on top of Zephyr Bluetooth Mesh stack, the ESP-BLE-MESH implementation supports device provisioning and node control.
Counter censorship via RF is dicey with something rolled out that pervasively. If you're gonna spin up meshtastic at a protest that's one thing, it'll be off everyone's radar. If you put a radio on every iPhone though, the powers that be will absolutely enact countermeasures. This leaves you two options, neither of which work:
1) make the signal hard to detect and jam/locate via modern EW techniques: spread spectrum (but not the kind lora uses) which is typically illegal already so the devices will never be certified to operate in any country
2) make detection/tracking a pain in the ass by using the existing 5G uplink band in a simplex fashion. This is a better option all round, jamming it results in killing everything and hunting down everyone with a phone is impractical. The issue is this is generally still illegal in a regulatory sense, though might be possible in at least some jurisdictions.
I've done some work in this space and there are pros and cons. BLE gets very complicated quickly. The mesh idea itself is nice, since there's no pairing / setup time - your device can be mostly asleep, and it simply powers up, sends a few bytes quickly, and powers down - that's a common industrial sensing use case.
However there are broader issues that are not adequately addressed - onboarding, key management, credential revocation and key rotation and so on.
I've also done some work earlier on 802.15.4 based mesh networks, there are standards for access control, payload encryption, frame integrity and sequential freshness, but delivering complete working solutions still becomes a bespoke offering that needs a fair amount of work to get right.
To see a broad adoption of these mesh standards we need to address the broader engineering, integration, deployment and operational issues.
Looking at the landscape of products at the local Home Depot or Lowes, it seems like most offerings are Wifi Only, need internet access, and require use of some proprietary app. The cynic in me thinks this is by design to lock people in to needing additional services (aka revenue streams)
I grabbed a bunch of very cheap (via aliexpress) zigbee based products, and have been amazed at how well they can work. I find myself wishing I could find more ZigBee based devices locally (US-Based, at common retailers) in order to more quickly implement and build up more automation.
Search tuya zigbee on aliexpress, avoid tuya wifi as that requires an app that is managed in China. All the tuya zigbee devices I bought (lights, plugs, and temperature/ humidity sensors) are reliably controlled locally by homeassistant and they appear to be durable.
I've had a few glitches, but I'm blaming my rookie experience with this before I blame the product directly.
Expect a long delivery time from anything at aliexpress though.
I was originally seeking a whole-house energy monitor, but the 200A CT Clamps are too big to fit on the main lines into my distribution center, because there are too many other wires around it.
One reason is that ZigBee got there first. It was the established player, people had worked around the wrinkles and it was generally pretty good. And you have a network effect of installed devices.
The main advantage (AFAIK) of BLE mesh is that your radio can also easily do normal BLE which can be helpful e.g. to talk to phones. This hasn't (so far) been enough to make it worthwhile to switch from ZigBee.
It was a bit of a pain to work with, but I think it was more the SDK/implementation than the spec itself. Iirc this company had created its own "mesh" before the official Bluetooth spec so working backwards to conform to the official spec was rough.
There were a few annoying limitations around max hops and leader election that were tricky. This mesh needed to have a single "gateway" node to go between the mesh and a normal BLE connection to the phone.
I still think BLE mesh is a better solution than ZigBee/Z-Wave.
I don't know if there's been good progress on client SDKs to have an easier time interacting with BLE. Getting bluez to do what I wanted was not simple, and a lot of the time it wasn't compatible with hardware.
This is true, but this was never really a problem for the adoption. Spec delays, no firmware upgrade support, and most importantly no proper backing by phone or gateway vendors. Thread and matter seem to have fumbled their opportunity though, so it might not be too late.
Licensing is an issue for me. I don't want to pay to use name 'Bluetooth'. That being said who would buy an "IEEE 802.15.1 compliant" or "IEEE 802.15.1 enabled" device? No real way about getting clever about it like calling it "cyan enamel"
Mesh networking. Now there's a technology that went nowhere. I still struggle to share files wireless P2P, forget about running any actual apps like video chat. Is that even possible?
ELI5: What is the benefit of using Bluetooth for Mesh instead of WiFi? Generally speaking? I assume one is about energy consumption. WiFi has more coverage, right?
Assuming we're talking about IoT applications: Lower power (better battery life), less congestion on Wi-Fi channels, no fussing over IPv4/subnets/etc., smaller attack surface (it isn't a trojan horse to your LAN), simpler setup/enrollment procedure than dealing with WiFi authentication. Most of these things are true with Zigbee/Z-Wave/Thread too, with the exception of being able to use a smartphone to directly connect and manage devices.
The attraction of their solution is that the devices can be wireless and don't require a lot of infrastructure. You basically just mount them to the wall or the ceiling. Battery life is pretty good as well depending how you set these devices up. Years typically. Installation cost and effort is one of the big hurdles for companies to take this stuff into use.
So, interesting to see some standards for mesh networking emerging here.