v0.2.x

[Nerivec]

What's Changed

• fix: improve routing by @Nerivec in fix: improve routing #70
• feat!: refactor saves to TLV-based by @Nerivec in feat!: refactor saves to TLV-based #71
• feat!: revamp structure by @Nerivec in feat!: revamp structure #72
• feat: various compliance features by @Nerivec in feat: various compliance features #76
  • feat: NWK replay protection
  • feat: install codes support
  • feat: APP link keys support
  • feat: APS duplicate detection
  • feat: APS fragmentation support
  • feat: NWK key switching support
  • feat: ED timeout support
  • fix: improve NWK GP duplicate detection
  • fix: improve APS ACKing (pending, wait window, etc.)
  • fix: improve handling for various spec commands
  • fix: extract raw MAC frame handling out of driver
  • fix: improve routing
  • fix: add compliance tests for every layer
• fix: encoding cleanup by @Nerivec in fix: encoding cleanup #77
• fix: MAC association join vs rejoin via authorized field

Full Changelog: v0.1.13...v0.2.0
v0.2.1: https://github.com/Nerivec/zigbee-on-host/releases/tag/v0.2.1

---

This discussion was created from the release v0.2.0.

[chris-1243]

This is a huge step forward!!!

I hope I will be able to use a CC2652P2 with OT firmware. My recent last try was not successful...
I need to keep a Silabs chip to sniff trafic😉

[chris-1243]

I hope I will be able to use a CC2652P2 with OT firmware. My recent last try was not successful...

This has been resolved in latest ZoH release👍

[burmistrzak]

@chris-1243 How does ZoH compare to a more traditional NCP setup in terms of performance/latency?
This entire thing sounds sick as hell! @Nerivec Excellent job! 🤩

[chris-1243]

Well, the idea behind this project is awesome and the work made by a single man is just incredible!

How does ZoH compare to a more traditional NCP setup in terms of performance/latency?

The network I made is not that big (~15 devices) but it performs well. I have not seen any latency or bad performance. It does the job like an NCP setup would do. Performance/latency should be tested on a larger network which I may not provide and I also do not own any kind of really chatty device to test with.
You would be surprised how fast it loads when you start Z2M...

As stated in the documentation, there is still some work to do in order to have this project released in a larger way. Some firmware need to be adjusted or fine tuned (Silabs and TI as well depending on the chip used). Innr devices are kind of nightmare to pair but this issue is also known.

Generally speaking, I would appreciate to move my main network on this driver but it is too early. What would be nice is to find more tester and skilled people who could help Nerivec to clean and improve the code. I do not have the knowledge to modify the code but at least I can test it and provide information to improve this project.

If you have some free time and motivation to help Nerivec, well, this would be nice !

N.B. ZGP devices work as they should and the Silabs bug does not exist in ZoH. It is how it should be in both TI and Silabs NCP stack.

[Nerivec]

We're still exploring how different hardware behave (and what configs might need tweaking in firmware), number of devices, features and whatnot, but from the current tests with chris, ZoH is about as fast (sometimes faster) than zstack, with a TI chip (measured in packets timing with sniffer).

[burmistrzak]

@chris-1243 Would love to spend a lot more time on free software/hardware projects but health, edu & life stuff are setting me limits atm unfortunately. 🫤
I'm doing my best juggling these things right now and hope to eventually have enough bandwidth for bigger, quality contributions.

@Nerivec I really like that we're keeping these adapters as dumb as possible. Not having to rely on opaque, vendor-provided stacks is a HUGE win.