This is the forum archive of Homey. For more information about Homey, visit the Official Homey website.
The Homey Community has been moved to https://community.athom.com.
This forum is now read-only for archive purposes.
The Homey Community has been moved to https://community.athom.com.
This forum is now read-only for archive purposes.
Closed
Stress testing Homey
I searched the forum for information on results of stress testing Homey but could not find what I was looking for. So here I go ...
A short intro: I use 15+ 4-in1 Z-Wave sensors, 10+ Z-Wave switches with ON/OFF and power readings, 3 Z-wave remotes, 4 Z-Wave door/window sensors, 4 temp/humid 433MHz sensors, IR blaster for audio/video, 11 KAKU switches/dimmers, 3 IP camera's and a multi element Sonos system that will be connected to Homey. I might add some HUE stuff. I am and will be using multiple (10+) and extended scripts to control stuff with some degree of interaction and user notification.
- How many devices (and different devices on different protocol) will Homey be able to control simultaneously?
- Will there be a point (and when) on which Homey starts to hamper in performance (or even crash)?
This discussion has been closed.
Comments
+1
+1
Well, to be honest, we will have to find out ourselves as well. But I don't see any reason why it wouldn't work. Of course there are some limits, but just turning sockets on and off isn't that resource-intensive. Firing a flow every second is, of course.
Thanks for you honesty ... but also a bit worrying.
Any chance to (stress) test Homey with multiple scripts/flows to check if some or more requirements are met AND running 2-4 multiple flows for activation Z-Wave devices and and IR-blast combined with a some Voice recognition and responses?
[I'm pretty sure you can think of something more complex and challenging for Homey]
... and then see what happens if all (100%) resources are being used ?
How "easy" are limits hit?
Result example: I rather see a very slow Homey then a crashing Homey.
If I'm not mistaken, Homey's hardware is better than for example the Fibaro Home Center Lite which is able to control up to 230 devices.
So I guess it won't be a problem very quickly.
Due to the homey supporting allot more protocol's then other home automation systems we should not use the hardware comparisation on any other home automation device.
Homey is the first and only device that is able to control more than a few protocols in one.
I agree. My experience is limited to HomeWizard. Since they added hue (via wifi), the whole thing has become a lot slower. It can take over 10 seconds before a scene is updated (10 seconds is a very long time).
Overall, I think the hardware is powerful enough, so software optimizations can make all the difference if there would be any problems. I'm not sure how all the antennas interact with one another and if that could cause additional bugs. We will just have to see...
I guess I have to wait with buying those 50 hue lamps :-p
In only a few situations all these protocols will be used.
Most of us will be only using a few dozen KaKu and/or Z-wave devices which will not be switched every 5 min.
Anyone remember the wizards in winter youtube video?
That would be a nice stress test!
You cant stress test the system by sending on of signals (like the ones on the above video). Emile is right this is not resource intensive. Proper stress testing is vital and I just a hope a million euro company will take it seriously before sending it to production. A proper stress test can reveal not just software issues but hardware ones. For proper stress test you need to run resource intensive processes simultaneously and for long (not just for 15 minutes): streaming HD video(s), syncing online accounts (gmail, IFTT, etc), voice recognition, sending instructions via mobile apps, using internal speaker, running scripts and receiving and transmitting using all frequencies, protocols. All sensors must be set to the highest refresh rate.
You could test many things with: overheating, fire risk, will all RF/IR transmitters will have enough power to reach devices further away, will the receiver miss messages, will it freeze the system.
Some of you would say why do you need this, it is far from a real life scenario? Well, not. Things happens, it can go in infinite loop and more importantly it is an industry standard thing to do with a good reason. I dont want to name them but many large and experienced manufacturers runs in the issue of over heating. You dont want to have a potential security risk (melt down, fire or the signal of the smoke alarm, flood sensor, intrusion sensor cannot get through, etc..).
What is the sphere made of?
I might use a lot more devices in the future then TS :-)
Rather than testing the entire system, it is very common to choose standard components build on their specification. If I would build a PC, I'm not going to test it for spontaneous combustion. Since the Homey has an external power supply and does not include a battery, the change of it setting your house on fire is 0. You'd be surprised how little is actually tested in the real world by the way.
My main worry is that some of the antennas could interfere with each other on rare occasions or that the horizontal plane of the Homey is a bit of a deadzone. The fact that the hardware is relatively powerful allows room to fix some issue in software.
The more devices I add to my HomeWizard, the slower it becomes. Scenes with 3 hues and 4 KaKu dimmers are set one by one in a slow sequence. I can switch 5 hues though the hue app instantly and I have seen it scale to 60 hue lamps updating very frequently in sync during a demo.
There is always a risk with preordering new hardware before release. If you are really worried about it, you should wait for the reviews to come it. I will definitely post my findings ;-)
Our additional microcontroller, so next to the 1GHz compute module, is comparable to HomeWizard's core module. Just sayin'
The white spheres are ASA (UV resistant plastic, we don't want Homey to look grey or yellow after a few years), the LED ring is PMMA.
Oh - now I understand how you make the LED-ring so beautiful:
PMA (Paramethoxyamphetamine) and PMMA (Paramethoxymethamphetamine) are stimulants with halluncinogenic effects similar to MDMA.
Lol +1
Perhaps the name 'Spacy' is maby better
Lol.. That could cause some shipping issues at the borders
ASA is great. Thanks a lot, you are great guys!
Very good example JaapPelt. I have seen a burned/melted computer before and also laptops (plural), and one of them was mine and it was not the battery!! Also the laptops have external transformer, but it does not matter. So it is a good example for "Why to do a test it has no explosives inside". As I mentioned even with testing large experienced manufacturers can have troubles. However I thing we should be fine with acrylonitrile styrene acrylate.
What can I say... I'm an optimist...
Just don't try to tastetest the led ring OK?
and @Emile never get high from your own supply
Feels good to say this, doesn't it. It sure feels good to me to read this! :-D