Hi
I have been following the release of Matter 1.0.
I'm intrigued, is it possible to run a Matter 'Hub' for connecting smart things in the home?
I know there is already MQTT packages, I am thinking these are not supporting the new Matter standard? ( CSA-IOT - Connectivity Standards Alliance)
Keen to hear the thoughts on this, excited what this could offer
What is exactly required here? What is exactly a Matter Hub? Is there any OSS implementation?
From what I can see, there is no OSS implementation.
Maybe here?
It's literally just a document - offer what?
Perhap then, you can tell us what it is?
Also, here's a link that doesnt take 5+ clicks to navigate to some relevant information: https://csa-iot.org/all-solutions/matter/
TBH, I'm not sure this post isn't marketing/clickbait SPAM or to just generate traffic.
Refrence:
NPoD (Necropost of the Day) brings you:
I just recalled that there was a post on the openwrt-dev mailing list back in July, https://lists.openwrt.org/pipermail/openwrt-devel/2023-July/041343.html with a link to the current Matter-into-OpenWrt work:
Home Assistant uses their Matter server written in Python.
Would be cool if this would be built for OpenWRT.
It uses WebSockets. There is also a controller/client needed that is able to talk/send commands to the server, preferably an app with a nice user interface. Otherwise it is practically not useful for most users.
In their server is also a matter client included. I don't know how this client is accessed (Python code?). Their own app connects to the Home Assistant platform which in turn uses the matter client, through the Home Assistant Matter integration (extension).
But why then not simply use Home Assistant? Don't get me wrong, I get the "would be cool" factor. But realistically is it worth the effort to create a "poor man's" replica of Home Assistant on OpenWrt just to have a subset of its functionality on what probably needs to be similarly specced hardware?
this document might help.
Tasmota supported the matter without being a member of the matter alliance.
there are routers which do have matter-over-thread compatible hardware. Otherwise, Matter over Wifi can be also be supported if someone does look into it.
...just to have a subset of its functionality...
Because that is like buying a Playstation 5 just to watch Blu-ray movies (while you already have a Blu-ray capable mediabox which is only missing the video codec). HA is used for home automation. Many people (like me) just want to dim their matter light bulbs for example, without any automation or IoT involved.
...on what probably needs to be similarly specced hardware?
I don't think the Matter server needs the same specs and could run on most router hardware (unlike Home Assistant).
In short: to avoid needing an extra (overpowered) device.
That still requires creating the whole "home automation" underbelly, you'd just be stopping at manual triggers short of the automated ones. If you build all of the necessary device integration and a user interface with manual triggers, you're most of the way to Home Assistant.
But, again, I'm not trying to poo on the idea. If you want to create all of this for OpenWrt, more power to you.
In my view you don't need Homeassistant to run on OpenWrt hardware.
OpenWrt should only be used as matter bridges/Matter-hub.
Hardware is there, it just needs to be utilized using same firmware i.e OpenWrt with matter packages.
For the most part it lacks the hardware side (as well as sensible means to communicate with some kind of supervisor, like home-assistant, web sockets aren't exactly great from a security point of view…). Even if just looking at plain old zigbee, you can pretty much count the number of routers with zigbee hardware on one hand (only google onhub srt-1900ac/ tgr-1900 comes to my mind, with rather old silabs silicon). For bluetooth there are a few more devices around, but not a whole lot either - and DECT ULE is pretty much AVM-only (and with zero FOSS driver support); z-wave is also trying to find customers. As we see here, there are already 5 competing rf standards in widespread use, most in the WLAN frequency spectrum; before even looking at 433 MHz and LoRa (868/ 900 MHz) and halow.
Yes, USB might bridge the gap (only <50% of all modern routers have USB in the first place), but as all three (except DECT ULE) technologies share the 2.4 GHz band with WLAN, there are considerations to deal with interference from this non-WLAN traffic (similar to bluetooth coexistance), which is not exactly forthcoming for USB solutions either (USB3…).
(at least for zigbee, there are esp32 based hubs with tasmota'ish firmwares that can provide tcp sockets for zigbee communication to z2m/ zha, the advantage here would be to get some distance between WLAN and zigbee radios, while profiting from the ethernet infrastructure of your router/ APs).
Like takimata I'm not at all dismissive of the idea, just a bit realistic of how likely this might be - and how useful in practice. I would like to see routing hardware with bt, zigbee, thread radios to become commonplace - and some idea of how to manage that, if only for my own curiosity. But in practice this is a chicken-and-egg problem, in the sense that we''d have to start seeing this kind of hardware in the wild - and some form of orchestration to make it work (something more than websockets, but less than full home-assistant on the router).
Zigbee USB with OpenWrt.
On OpenWrt hardware with Thread radios you probably don't want to have to run hassio right on the device, and instead run OTBR or it to extend your existing smart home network.
Yes exactly that is my point.
OpenWrt should be used as matter or zigbee router not as zigbee coordinator or main automation hub.
Where's the point in that?
Zigbee router functionality just needs mains power, everything else is done inside the proprietary zigbee firmware. Every (most) zigbee lightbulb or zigbee power plug acts as router, there are also USB-powered dedicated zigbee extenders you could plug into your router's USB for around a tenner (they only use USB for power, so no drivers necessary).
I could very much understand a desire for zigbee/ matter coordinator use (and I'd point towards zb-gw03/ tasmota for one cheap alternative to that), maybe even running home-assistant in a VM or container on your router, but not 'only' as zigbee-router, there are cheaper and more flexible options for that (e.g. a strategically placed zigbee lightbulb for 2.50 EUR).
Again, I'm not trying to be dismissive here, I'm genuinely interested in the 'why' (I've only scratched the surface of smarthome stuff myself).
I am personally running all those devices as zigbee routers.
Sonoff and Tuya relays. Ikea Bulb. Dedicated USB zigbee router, but I still think if we have got Zigbee/matter compitable chips in wifi routers supported by OpenWrt why not use those?
A wifi router with Zigbee/matter chip and large 2.4GHz antenna will obviously better zigbee router.
The more zigbee router devices you have in network the stronger and reliable the network is.
Edit: it is similar to what we use to discuss what was the purpose of USB in a router back in those days and what is the use of hyper terminal in windows(if somebody remember those days).
Zigbee/Matter incoming atleast this device supports it.
If neither the guy constantly yelling at you nor its product name won't, then certainly its price will give you a nosebleed.
Overpriced under-specs yes.
Aqara Presence Sensor FP1 was USD 80 on launch it started the presence sensor trend. The presence sensor I am using in this room is LD2420 on ESP32S2 board. Total cost Pkr 1200 (USD 4.32) sensor for Pkr 200, ESP32S2 for Pkr 600 and case for Pkr 400. Software opensource ESPhome based.
If software support for ZigBee and matter is added in OpenWrt somehow we will be able to find cheap hardware for it too.
W1700K a wifi 7 router running old OpenWrt 18 based firmware was like USD 28 for used and USD 36 for new.
