Thread is a wireless standard meant to sit next to Bluetooth and WiFi.
Matter is a home automation protocol can that be used over Thread or WiFi. Ideal Matter devices use Thread instead of WiFi because running a bunch of home devices like light bulbs or switches on your WiFi is a recipe for disaster.
Matter is important because it provides native compatibility among different platforms.
There are also performance implications (a Zigbee coordinator can easily handle 100 devices, while many routers would struggle with that amount of clients), power saving (especially for battery powered sensors) - some Zigbee sensors can last years on a single coin cell battery.
Right, thanks. But this can be easily resolved by:
Removing devices’ access to WAN, which also vastly reduces the external actor’s ability to compromise them in the first place.
Isolating devices from each other with internal firewall rules, allowing them to only interact with the hub host.
Is this correct, or am I missing something?
With a good router, it’s not that hard to do. But even then it took me a long time to get around to setting up the separate network, and I don’t think I’ve migrated all of my devices over to it still (some got moved, new ones go there, but there are some older devices still sitting on the original network). So, yeah, there’s definitely extra effort, and it’s not really fun like getting that new smart device integrated. I will say the stuff on that network works perfectly - I haven’t noticed any side effects.
Oh I did allow them access to the pool ntp server so they can pick up the correct time, and some require temporary access to the internet while setting up (the linknlink RF device needed it to download the Home assistant version of their firmware, for example).
ZigBee/Thread are just better for this, you’re protected without doing anything.
Requirements like the ones you listed above make widespread adoption impossible, short of forcing routers to have a separate IoT network and forcing devices to use only that, with all the issues that may prop up along the way.
Good luck explaining how to do any of this to my parents, for example. For someone with a technical background that’s feasible, for someone with an it background it’s even easy. For the other 90 or 95% of people who are actually supposed to buy and use these things, it isn’t. They don’t even know something like this can be done, let alone that it should be done.
But that is not a fault of WiFi as a medium, but rather of the ecosystem of devices as we know them. Some company might launch a “Home Automation WiFi” product, which would be simply a home hub with a builtin WiFi router pre-configured with the recommended security settings. Zero config nor admin work required, just buy the right (hypothetical) hub.
Though the real problem is that every other device relies on cloud connectivity, which highly limits such hubs effectiveness. Again, that isn’t an inherent fault of how WiFi works, rather I see it as a problem with the ecosystem and how consumers want their devices to work without any hub. Hopefully with more local-only devices that trend can still be reversed.
but it is a fault of WiFi as a choice for that application. Just because it does wireless communication doesn’t mean it’s suited for any application that needs a wireless protocol. Using it for very-low traffic applications is simply not what it was designed to do, and it has significant negative effects if you do. Any device you add basically slows down any other device by a bit. And wifi network you add in a physical area decreases the effectiveness of all other wifi networks in it’s vicinity. In even medium densly populated areas, wifi is already borderline unusable due to congestion. Your proposed (dedicated) hub is a good idea for network isolation, assuming it’s designed and configured correctly, but that also comes with more and frankly just as bad security implications, just different ones. To be clear, having like a light bulb or two wifi is a fine choice. For 50 or a whole smart home network, it no longer is.
Both Zigbee and Matter do not rely on cloud connectivity as a protocol, though many of the manufacturers implementations do effectly add that on top: you get the exact solution you propose here as well. At least with these standards you can control everyhing locally, if you want to, and you don’t congest the spectrum nearly as much as wifi does.
Using it for very-low traffic applications is simply not what it was designed to do
Was it not? Sure, it wasn’t specifically designed for low traffic, but it was designed with a wide range of traffic levels in mind; after all, virtually every WiFi capable device, home automation or not, uses very little traffic most of the time as it idles. Now, I absolutely appreciate Zigbee or ZWave being optimized for minimal energy consumption, which is useful for some device types; but I feel it isn’t right to call WiFi a poor choice for low traffic just because it also handles very high traffic well.
You raise an interesting point about congestion, though, and that is very much was I hoped to learn about. I am sometimes under impression that device designers assume everyone lives in a detached house so interference can be ignored. Do you know how specifically is Zigbee better in this regard? Living in a condominium I always had very poor experience with Zigbee reliability, which might or might not have been due to local radio noise at various ares of the spectrum, so I’m curious to learn about details how exactly these purpose-specific transfer layers deal with noisy neighborhoods.
Both Zigbee and Matter do not rely on cloud connectivity as a protocol
In my ideal world, no devices would not rely on cloud connectivity ever, regardless of their choice wireless transport layer. The fact that the nature of Zigbee or Thread stop device manufacturers from stupid practices (such as relying on direct WAN access) is nice, sure, but does being less capable really make them better suited for the job? I would prefer to appreciate Zigbee or Thread for what they are and have, rather than by the fortunate side effects of what they can’t do.
but that also comes with more and frankly just as bad security implications, just different ones
What do you have in mind? Surely, no solution is perfect, but at least on the surface it sounds much better than the alternative of no isolation at all, so I’d love to learn what worries you.
I see someone replied about security. But I was just taking about stability.
Most people don’t have super beefy wifi routers. Many have whatever shit their ISP sent them. These are fine for your average number of laptops and phones, etc. but if you then throw on 10 more 2.4GHz WiFi IOT devices, you are probally going to run into devices randomly dropping off the wifi, etc.
Additionally, wifi is usually chosen over other protocols by manufacturers due to the cost of hardware and development. So they are often lower quality. (This is only one reason)
But sure, if you have a super awesome 2.4GHz wifi setup and high quality wifi devices, maybe things will work out just fine. But my personal experience with WiFi tells me I shouldn’t clutter my WiFi.
Also, if you were curious, yes: almost all WiFi IOT devices are 2.4GHz only.
This sort of makes it sound like the advantage of Thread is merely that it’s new, and therefore nobody will be affected by having a poor pre-existing Thread setup. If ISPs were sending people Zigbee hubs, it sure would be the cheapest shit available, which could very well translate to similarly terrible Zigbee performance.
I see your point, but there should be much more merit to the specialized IoT protocols than just that nobody has yet flooded the market with terrible Thread/Zigbee devices.
I’m not sure manufacturers choose WiFi because of hardware costs. There are often other reasons (some good, some terrible) for this choice, but I’m certain Zigbee support has to be cheaper; having disassembled plenty of such devices it’s almost always a dedicated IC and a tiny PCB trace for an antenna. WiFi support requires a fairly complete TCP/IP stack implementation, basic certificate management etc. which will inevitably require a small SoC; and while prebuilt solutions are plentiful, Zigbee alternatives are an order of magnitude cheaper.
I can imagine software development costs being lower, though, given how every other programmer knows a fair deal about TCP/IP networking, while good comprehension of dedicated IoT protocols is a much rarer skill, there are also much less community resources and open-source solutions available etc.
Does Thread support pairing two or more devices so they can control each other directly without going through the coordinator? I do that a bit with my Zigbee network.
I never meant to imply that Thread was the only solution to WiFi based IOT being bad. Just that it is a forward looking choice that has interoperable support from the dominant computing platforms. This is why IKEA is moving to it.
Thread is a bit more power efficient, which matters for battery powered devices that aren’t connected to permanent power and don’t need to transmit significant data, like door locks, temperature/humidity sensors, things like that. A full wifi networking chip would consume a lot more power for an always-on device.
Dishonest question, your following comments show you clearly know the common points of these discussions.
There’s no arguing that using WiFi for this purpose is an obstacle to widespread adoption and a security risk, no matter how well you yourself are handling it.
There’s also the technical advantage of those IoT networks being meshed by design, whereas expanding a WiFi network with mesh functionality often carries challenges.
I may know some answers, but I don’t have a full picture, and I wish to complete my understanding by learning others’ perspectives and what they find most important about this choice of protocols. Thank you for sharing yours.
An important difference between thread and zigbee/wi-fi I’m not seeing mentioned is that all thread devices automesh in a hub/spoke model as long as they’re not battery powered. So your light bulbs, plugs, etc all become extenders and part of a self healing mesh network without a single point of failure. For me it works better than Zigbee for this reason.
They’re different in their implementation. Zigbee automesh is more of a centralized router-hub model with self healing relying on routing tables. This caused significant issues for me. Thread is true automesh with all devices acting as a hub in a hub/spoke model, so there’s no centralized routing table to act as a single point of failure.
Thread also works on the 2.4 GHz range but can utilize sub ranges of 868 in Europe and 915 in north America. The 868 and 915 GHz ranges are what LoRa operates on and provides a much greater range for low data rate transmissions.
In fact Meshtastic operates on LoRa on 915 here in the U.S. and I have a node in my second floor window with a 3db antenna and I have been able to message both ways up to 3 blocks away.
Long story short, utilizing 868 and 915 in these devices will make dead spots a thing of the past within a home, even with their lower gain internal antennas.
Man there are way too many IoT standards. What’s the difference between these two? How do they each compare to Matter?
Thread is a wireless standard meant to sit next to Bluetooth and WiFi.
Matter is a home automation protocol can that be used over Thread or WiFi. Ideal Matter devices use Thread instead of WiFi because running a bunch of home devices like light bulbs or switches on your WiFi is a recipe for disaster.
Matter is important because it provides native compatibility among different platforms.
What kind of disaster specifically? I hear everyone discouraging from using WiFi for home devices, but never understood what the actual risks are.
There are also performance implications (a Zigbee coordinator can easily handle 100 devices, while many routers would struggle with that amount of clients), power saving (especially for battery powered sensors) - some Zigbee sensors can last years on a single coin cell battery.
Let’s say your lightbulb(s) become comprised. A bad actor now has a toehold into your network.
Maybe they run a spam relay through the bulb. Maybe they attack your network by attacking the printer that hasn’t had a firmware update in years.
Point is, there’s someone with access to your network that isn’t you.
https://old.reddit.com/r/Ubiquiti/comments/1fjspoc/what_is_the_point_of_having_a_separate_devices/
Right, thanks. But this can be easily resolved by:
Sounds like an admin job I don’t want to do.
With a good router, it’s not that hard to do. But even then it took me a long time to get around to setting up the separate network, and I don’t think I’ve migrated all of my devices over to it still (some got moved, new ones go there, but there are some older devices still sitting on the original network). So, yeah, there’s definitely extra effort, and it’s not really fun like getting that new smart device integrated. I will say the stuff on that network works perfectly - I haven’t noticed any side effects.
Oh I did allow them access to the pool ntp server so they can pick up the correct time, and some require temporary access to the internet while setting up (the linknlink RF device needed it to download the Home assistant version of their firmware, for example).
ZigBee/Thread are just better for this, you’re protected without doing anything.
Requirements like the ones you listed above make widespread adoption impossible, short of forcing routers to have a separate IoT network and forcing devices to use only that, with all the issues that may prop up along the way.
Good luck explaining how to do any of this to my parents, for example. For someone with a technical background that’s feasible, for someone with an it background it’s even easy. For the other 90 or 95% of people who are actually supposed to buy and use these things, it isn’t. They don’t even know something like this can be done, let alone that it should be done.
But that is not a fault of WiFi as a medium, but rather of the ecosystem of devices as we know them. Some company might launch a “Home Automation WiFi” product, which would be simply a home hub with a builtin WiFi router pre-configured with the recommended security settings. Zero config nor admin work required, just buy the right (hypothetical) hub.
Though the real problem is that every other device relies on cloud connectivity, which highly limits such hubs effectiveness. Again, that isn’t an inherent fault of how WiFi works, rather I see it as a problem with the ecosystem and how consumers want their devices to work without any hub. Hopefully with more local-only devices that trend can still be reversed.
but it is a fault of WiFi as a choice for that application. Just because it does wireless communication doesn’t mean it’s suited for any application that needs a wireless protocol. Using it for very-low traffic applications is simply not what it was designed to do, and it has significant negative effects if you do. Any device you add basically slows down any other device by a bit. And wifi network you add in a physical area decreases the effectiveness of all other wifi networks in it’s vicinity. In even medium densly populated areas, wifi is already borderline unusable due to congestion. Your proposed (dedicated) hub is a good idea for network isolation, assuming it’s designed and configured correctly, but that also comes with more and frankly just as bad security implications, just different ones. To be clear, having like a light bulb or two wifi is a fine choice. For 50 or a whole smart home network, it no longer is.
Both Zigbee and Matter do not rely on cloud connectivity as a protocol, though many of the manufacturers implementations do effectly add that on top: you get the exact solution you propose here as well. At least with these standards you can control everyhing locally, if you want to, and you don’t congest the spectrum nearly as much as wifi does.
Was it not? Sure, it wasn’t specifically designed for low traffic, but it was designed with a wide range of traffic levels in mind; after all, virtually every WiFi capable device, home automation or not, uses very little traffic most of the time as it idles. Now, I absolutely appreciate Zigbee or ZWave being optimized for minimal energy consumption, which is useful for some device types; but I feel it isn’t right to call WiFi a poor choice for low traffic just because it also handles very high traffic well.
You raise an interesting point about congestion, though, and that is very much was I hoped to learn about. I am sometimes under impression that device designers assume everyone lives in a detached house so interference can be ignored. Do you know how specifically is Zigbee better in this regard? Living in a condominium I always had very poor experience with Zigbee reliability, which might or might not have been due to local radio noise at various ares of the spectrum, so I’m curious to learn about details how exactly these purpose-specific transfer layers deal with noisy neighborhoods.
In my ideal world, no devices would not rely on cloud connectivity ever, regardless of their choice wireless transport layer. The fact that the nature of Zigbee or Thread stop device manufacturers from stupid practices (such as relying on direct WAN access) is nice, sure, but does being less capable really make them better suited for the job? I would prefer to appreciate Zigbee or Thread for what they are and have, rather than by the fortunate side effects of what they can’t do.
What do you have in mind? Surely, no solution is perfect, but at least on the surface it sounds much better than the alternative of no isolation at all, so I’d love to learn what worries you.
firewall rules? who cares about firewall rules if the switch still forwards packets to the printer?
firewall rules only work on end devices, and routers when the destination is in a different subnet and broadcast domain.
Sounds like homework
I see someone replied about security. But I was just taking about stability.
Most people don’t have super beefy wifi routers. Many have whatever shit their ISP sent them. These are fine for your average number of laptops and phones, etc. but if you then throw on 10 more 2.4GHz WiFi IOT devices, you are probally going to run into devices randomly dropping off the wifi, etc.
Additionally, wifi is usually chosen over other protocols by manufacturers due to the cost of hardware and development. So they are often lower quality. (This is only one reason)
But sure, if you have a super awesome 2.4GHz wifi setup and high quality wifi devices, maybe things will work out just fine. But my personal experience with WiFi tells me I shouldn’t clutter my WiFi.
Also, if you were curious, yes: almost all WiFi IOT devices are 2.4GHz only.
This sort of makes it sound like the advantage of Thread is merely that it’s new, and therefore nobody will be affected by having a poor pre-existing Thread setup. If ISPs were sending people Zigbee hubs, it sure would be the cheapest shit available, which could very well translate to similarly terrible Zigbee performance.
I see your point, but there should be much more merit to the specialized IoT protocols than just that nobody has yet flooded the market with terrible Thread/Zigbee devices.
I’m not sure manufacturers choose WiFi because of hardware costs. There are often other reasons (some good, some terrible) for this choice, but I’m certain Zigbee support has to be cheaper; having disassembled plenty of such devices it’s almost always a dedicated IC and a tiny PCB trace for an antenna. WiFi support requires a fairly complete TCP/IP stack implementation, basic certificate management etc. which will inevitably require a small SoC; and while prebuilt solutions are plentiful, Zigbee alternatives are an order of magnitude cheaper.
I can imagine software development costs being lower, though, given how every other programmer knows a fair deal about TCP/IP networking, while good comprehension of dedicated IoT protocols is a much rarer skill, there are also much less community resources and open-source solutions available etc.
Thread is the replacement for zigbee.
It’s zigbee plus more features.
There is nothing zigbee does better than thread.
Lower power usage, ipv6, standard tls security, not proprietary, etc.
Does Thread support pairing two or more devices so they can control each other directly without going through the coordinator? I do that a bit with my Zigbee network.
I never meant to imply that Thread was the only solution to WiFi based IOT being bad. Just that it is a forward looking choice that has interoperable support from the dominant computing platforms. This is why IKEA is moving to it.
2.4GHz*
Whoops! Fixed.
Thread is a bit more power efficient, which matters for battery powered devices that aren’t connected to permanent power and don’t need to transmit significant data, like door locks, temperature/humidity sensors, things like that. A full wifi networking chip would consume a lot more power for an always-on device.
Dishonest question, your following comments show you clearly know the common points of these discussions.
There’s no arguing that using WiFi for this purpose is an obstacle to widespread adoption and a security risk, no matter how well you yourself are handling it.
There’s also the technical advantage of those IoT networks being meshed by design, whereas expanding a WiFi network with mesh functionality often carries challenges.
I may know some answers, but I don’t have a full picture, and I wish to complete my understanding by learning others’ perspectives and what they find most important about this choice of protocols. Thank you for sharing yours.
An important difference between thread and zigbee/wi-fi I’m not seeing mentioned is that all thread devices automesh in a hub/spoke model as long as they’re not battery powered. So your light bulbs, plugs, etc all become extenders and part of a self healing mesh network without a single point of failure. For me it works better than Zigbee for this reason.
Zigbee does that too tho, right?
The wiki on zigbee says so at least
Pretty sure than an underlying feature of both zigbee and zwave.
They’re different in their implementation. Zigbee automesh is more of a centralized router-hub model with self healing relying on routing tables. This caused significant issues for me. Thread is true automesh with all devices acting as a hub in a hub/spoke model, so there’s no centralized routing table to act as a single point of failure.
Thread also works on the 2.4 GHz range but can utilize sub ranges of 868 in Europe and 915 in north America. The 868 and 915 GHz ranges are what LoRa operates on and provides a much greater range for low data rate transmissions.
In fact Meshtastic operates on LoRa on 915 here in the U.S. and I have a node in my second floor window with a 3db antenna and I have been able to message both ways up to 3 blocks away.
Long story short, utilizing 868 and 915 in these devices will make dead spots a thing of the past within a home, even with their lower gain internal antennas.
I exclusively use ZigBee. It automeshes.
Thread is a network layer thing, comparable to WiFi or Bluetooth. Matter is an application later thing, comparable to HomeKit or Google home.
Zigbee is both network and application layer.
This article has a decent overview https://www.smarthomeperfected.com/zigbee-vs-thread/
Thread works with matter, these devices will use Matter over Thread.