Not sure if this question has been asked yet or not.
Let’s say I have one IMP. How often does it communicate (transfer)
between the IMP and the “cloud” server? Would it be noticeable on
my WiFi connection at home … slowing down my other computers
and devices using my WiFi?
What if I had 6 appliances in my house that all had IMPs?
What impact on WiFi would that pose?
I guess I’m trying to determine how much of a “hog” the IMP might be with my Wifi.
It communicates very rarely, and takes very little data.
Right now, the boot process maybes passes 4kB of data (TLS startup), then the only packets flowing are data to and from the imp, which depend on what you’re using it for (there’s TLS and BSON overhead, but this is miniscule) and server pings (to keep the link alive and ensure NAT doesn’t timeout). We’re talking about maybe 4kB per hour.
If you had a really bad network connection - say 512kbit, ie 64kB per second - an imp would be using <0.002% of your external bandwidth and generally a tiny fraction of that on your local wifi (as that’s higher performance).
You’ll never notice it, even if you have many of them…
I have a related question (not sure if I should start a new topic). I am not concerned about LAN data requirements, but am considering the use case of home monitoring. Specifically in rental properties where I don’t maintain a broadband connection and don’t want to rely on the tenant’s connection. In this case I would install a WiFi->3G router in a secure location, and buy some pay per byte SIM card (e.g. roamline.com).
Wifi traffic isn’t a concern, but I am wondering how much data usage I should expect after the initial firmware download. E.g. 4kb/hour * 10 devices would work out to about $13 / month in data fees, which seems high given that the only information I actually want to transfer would be alarm conditions (raw data would merely be a few bytes per year, plus all the protocol overhead).
Is there any way to control how often the imp sends data to a host outside the Wifi LAN?
Except for during cold boot and firmware upgrades, the imp only sends data when it’s told to. So for an alarm-conditions-only application, the imp would probably be in deep sleep (not on wifi at all) except when real data needed to be transferred. There is a trade-off, though, as an imp only gets told about any updated code when it checks in with the server – so if it never checks-in except if some emergency happens, you’d never be able to push updated software to it.
How about waking up once or twice per day to report any problems? That might be a good compromise. I assume you might be looking for things like water detection on the basement floor, or temp/furnace/air status?