Tuesday, 16 October 2018

Nanananana

B.A.T.M.A.N.

I think I'm going to have a go at implementing a mesh network scheme/algorithm inspired by the ones used in batman-adv IV for my project. I can't actually implement batman-adv it's too much work and inappropriate for the ESP platform. The bare bones of the mesh algorithm on the other hand looks like a good candidate and unlike a lot of things floating around online this is an active open source project that's been through multiple iterations of working on the difficulties of mesh networking.

I need to think through how ESP-Now not having a broadcast mechanism will affect it. ESP-Now can send a packet to all a node's peers, but that's not the same thing, those peers need to be found and actively added in the first place.

I've done a trivial test of ESP-Now that involves running SoftAP and periodically scanning for the beacon frames, then switching the SoftAP off once you've joined the mesh. However this would very swiftly end up with multiple isolated meshes that can't find each other.

Current thinking is to have some nodes, at least one per mesh, run as SoftAP, to maintain options for peer discovery, spinning up more periodically if none are directly visible from a section of the mesh.

Likewise ESP-Now adds another layer of complication/admin as the peer table is limited to 20. This peer list will have to be pruned/updated, not something that happens in batman-adv, which only considers if a node is one-hop reachable. So there will be three potential states for a node, one-hop reachable peers, one-hop reachable non-peers and multi-hop reachable.

Thing is I know I could fix this easily if I just started sending my own crafted vendor action frames, especially broadcasts, but this way lies an even deeper rabbithole of reinventing the wheel than the one I'm already in. Using the standard SoftAP and Wi-Fi scanning libraries are an expedient way to avoid this I hope.

Monday, 15 October 2018

Cellar refurb

I've not been doing as much project work as usual recently as I've had a specialist company in to 'tank' my cellar.

The cellar has been used for years as my occasional workshop, but has always suffered from being a cold dank nasty place and about once a year it would flood, courtesy of backwash from a drain in the floor. The water would almost immediately run back out of the drain but it left an awful mess behind, occasionally ruining things I'd been stupid enough to leave on the floor.

I've run a dehumidifier 24x7 for maybe five years down there and it only ever managed to keep things vaguely acceptable. Until it flooded, when I'd have to move everything out, clean it all and mop up all the little pools of water left behind on the uneven floor.

I had written off the idea of getting it professionally tanked as too expensive, but when I recently got a quote it was much more affordable than I expected.

The whole project has taken about two months and I've spent lots of my free time in first clearing the cellar out, decorating it and then moving everything back down.

I have purposefully gone with the 'industrial' look of exposed joists to give me a little more headroom and painted everything stark white for light. There is enough height to walk around, unless you're well over 2m tall, but a proper ceiling would have made it quite claustrophobic.

Now the paint and plaster skim has dried out, it sits nicely at the ambient temperature and humidity of the rest of the house, you don't really feel you're in a cellar.

I'm very pleased with my new den.

Monday, 1 October 2018

Am I barking up the wrong tree?

After a successful test of painlessMesh at the weekend I posted about it on the ESP8266 Facebook group and somebody wondered if I'd considered ESP-NOW.

Which I may have seen mentioned a couple of years ago but have never fiddled with and completely forgotten.

While painlessMesh is doing what I want I have in the back of my mind that it's layering heavyweight stuff together (AP associations, WPA2 encryption, TCP servers, NTP-like time sync) to give you a simple messaging protocol.

Despite my assertion I didn't want to get sucked into writing my own networking code I had actually been Googling to see if you could send/receive packets directly between ESP8266 stations without actually associating with an AP at all. Most of the chatter about that is building antisocial Wi-Fi de-auther devices, and it didn't look a very fruitful route.

Of course this is exactly what ESP-NOW is. It's a proprietary messaging protocol that's part of the Espressif suite of dev tools for the ESP8266/32. It leverages standards compliant 'vendor action frames' which allow vendors to send frames directly between stations for their own proprietary reasons.

As this is very low level ESP-NOW supports star networks with 20 nodes, rather than the 5 nodes of the full WPA2 encrypted IP connection that painlessMesh is built on. So a very similar 'star of stars' mesh built out of ESP-NOW connections should scale more gracefully. It also leaves some resource for a conventional IP device to bridge in to the mesh and talk to it. It is unencypted by default but you can switch this on and it lowers the capacity to 10 nodes.

As painlessMesh seems to handle scaling quite well, I'm more interested in the greater power efficiency and range promised by ESP-NOW.

If you've ever noticed you can see a Wi-Fi SSID way outside the range you can make useful connections to it, that's the kind of range ESP-NOW promises. It's not doing all the encryption, background keepalive and management work needed for a full AP association, so it's much more tolerant and fewer packets means less power.

Digging deeper there is also ESP-MESH in the dev tools which the documentation promises is a self organising mesh built out of ESP-NOW connections.

Why have I not noticed these protocols before?

The answer seems to be nobody in the 'maker' community bothers with either ESP-NOW or ESP-MESH. I've found a handful of people talking about it but very much proof of concept builds of the example code, no mention of big mesh projects. I guess the IoT spin that's put on the ESP product lines has meant the non-IP protocol suite is mostly ignored.

As this is all changeable in the code I'm going to continue working with painlessMesh, build the rest of my static nodes and start on the wearable GPS-equipped nodes. This will allow me to conduct a bigger test including them at the end of October. If I subsequently find a way to drop in ESP-MESH or engineer my own thing with ESP-NOW it'll be a 'range upgrade'.