I have my imp running, and returning temp data to Cosm.
However, the application needs the imp to send a local http request to other devices on the local (Imp) side of the firewall/router. Im doing this to ensure resilience in my high uptime application… reliance on the internet isn’t an option.
My ‘receiving’ apps are embedded devices that simply need a http request such as ‘http://192.168.1.3/changepower=234’ where 234 is the variable data the IMP generates.
How do I get the imp to send a local http request ? is all the http handling via the server side ?
Probably not the answer you’re looking for, but you could get a cheap UART->Ethernet module - I haven’t used this particular one, but it’s 3.3V-compatible and shouldn’t be too terrible to get up and running based on experiences I’ve had with the ENC28J60 before.
The issue is that the current Imp firmware (as posted in a different thread) requires a live internet connection via its WiFi every cold boot. This is considered a limitation that will eventually be removed (again from that other thread) but is probably a strong blocker for your application - in the event of a power blip and network outage the Imp would basically stop responding.
I am considering an Imp as a replacement for the hopelessly retarded controller for my home water conditioning system. A good solution could appeal to the manufacturer. I need to be able to get near-realtime scheduling information to the Imp from my own desktop appl. My preference would be through a TCP socket on my desktop appl in my home enclave (either side of the socket could be listening). An alternative could be to make the scheduling info available to the Imp server-hosted agent (if I understand the overall architecture - not guaranteed). It is not clear to me how to get user information, from anywhere, to the Imp. Neither is it clear to me how sensor information sent from the Imp could be made accessible to me in my prospective desktop appl.
http://electricimp.com/manufacturers/ shows connectivity between the Imp and “my” server, but this seems to contradict Hugo’s statement that the Imp can connect outbound to only Imp servers. Surely the Imp does not honor http requests, right? How is the connection shown in the diagram implemented? I really need local connectivity using the Imp’s seemingly well-implemented wi-fi, or some good alternative, to make this work.
I have shelved an arduino solution as do-able but too costly.
The imp itself only connects to the imp service. However, the agent (another squirrel VM running on our servers that you write code for) can both make and receive HTTP requests.
Basically, the way you talk to an imp is to talk to our servers, that forward the traffic. This means the imps are accessible no matter what firewalls and NAT might be in the way.
There is no way to talk to an imp locally. If you want a wifi module, there are plenty of those available, and they provide that functionality, but that’s not what the imp is for - the imp is for connecting things to the internet.