Replies: 2 comments 1 reply
-
|
Thank you very much for your efforts, it is appreciated! |
Beta Was this translation helpful? Give feedback.
-
|
This is very cool! I've read through your current status, and it certainly looks like very impressive work! I really like the status page from a technical perspective. It's very satisfying to get a look at the status of the machine under the hood 👍 Can you say a few words about what the expectations are right now, for real-world situations? For instance, there are a few things missing, like this:
For "normal" use cases with "reasonable" devices, is this likely to have a big impact? Or are any of the other missing features likely to cause issues? I completely understand that this is the first working draft of this new system - I'm just trying to wrap my head around roughly what state it's in right now. Again, thank you for this work, it looks really cool :-) |
Beta Was this translation helpful? Give feedback.
-
It's not yet ready for "main live network", for obvious reasons. I'd say, most devices should be fine (pairing/using). The ones that are a bit more "off-spec" may exhibit more issues, although I did do a few tests with the all-too-common "poor-quality Tuya" and it did work fine... I don't have any Aqara (really bad pairing system)... I haven't done much stress-testing, so, optimizations may be in order in some areas. I haven't spent much time on all the LQI stuff yet (most values are currently raw dBm from Spinel metadata), nor on the inner workings of ZigBee groups (I don't have sniffs ready, I need to make some, to confirm and add some CI tests). I'll be working on improving the state saving next (implementing "reset" support as used by Z2M, to make testing easier). Some users have very specific use cases for ZigBee, this is where we'll likely see things needing more implementation (queueing, groups, heavy broadcasting, etc.). We also need some feedback on the RCP firmware stability -regarding its specific use by ZoH-. Anyone can help move this along, even without testing ZoH: |
Beta Was this translation helpful? Give feedback.
-
Important
This is experimental, not intended for use in live networks and subject to breaking changes outside the Zigbee2MQTT release cycle.
As its name indicates, this driver provides a host-based ZigBee implementation.
Using OpenThread RCP firmware as a base allows to communicate with various radios (802.15.4) and use the same ZigBee stack across hardware/manufacturers. Currently Silicon Labs and Texas Instruments firmware are available, with more to come in the future!
The OpenThread RCP firmware is in fact only used for its ability to communicate directly with the coordinator's radio (Radio Co-Processor).
This provides numerous benefits:
Information on testing
https://github.com/Nerivec/zigbee-on-host?tab=readme-ov-file#testing
Tip
Make sure to run with logging set as
debugso you can share them in case you encounter any issue!Current status
https://github.com/Nerivec/zigbee-on-host?tab=readme-ov-file#current-status
Important
Want to help even without testing ZoH? You can provide sniffs from live networks! Nerivec/zigbee-on-host#14
Current release discussion
Nerivec/zigbee-on-host#10
Reporting issues
Since most of the issues will be about the ZoH implementation rather than Zigbee2MQTT, please report issues in that repository:
https://github.com/Nerivec/zigbee-on-host/issues
Important
Make sure to check the current release discussion for issues that are already known.
Discussing features / Asking questions
https://github.com/Nerivec/zigbee-on-host/discussions/new/choose
Beta Was this translation helpful? Give feedback.
All reactions