- Mon Apr 13, 2020 9:13 am
#86506
That's what my sensors do - I use SSL and certificates at either end, but the data isn't that exciting really.
What is more important is to make sure that when one of my devices is stolen and the code is reverse engineered, and the whole memory is made available (and the assumption is that this is what will happen), that I have the smallest possible attack vector.
My devices all register themselves using a time locked key that is generated per installer, and when they are successful they get a device key that is unique to them and that only they know. They also transmit their location when registering and are geolocked.
When a device is stolen, the thief can only potentially spoof that one device and if it doesn't show the correct location it will be sent a command to update its firmware but will actually be sent a self destruct. Device copies will not be able to register themselves. Attempts to spoof the server side will not work because each device has commands signed by their unique key, so remote takeover will be limited to just that one device. Attempts to listen in on the feed will only be able to get one device worth of data. Not perfect but good enough for what I need on the public internet.
All this is my own and the platform's code - you have to build it yourself. Nothing in the SDKs
What is more important is to make sure that when one of my devices is stolen and the code is reverse engineered, and the whole memory is made available (and the assumption is that this is what will happen), that I have the smallest possible attack vector.
My devices all register themselves using a time locked key that is generated per installer, and when they are successful they get a device key that is unique to them and that only they know. They also transmit their location when registering and are geolocked.
When a device is stolen, the thief can only potentially spoof that one device and if it doesn't show the correct location it will be sent a command to update its firmware but will actually be sent a self destruct. Device copies will not be able to register themselves. Attempts to spoof the server side will not work because each device has commands signed by their unique key, so remote takeover will be limited to just that one device. Attempts to listen in on the feed will only be able to get one device worth of data. Not perfect but good enough for what I need on the public internet.
All this is my own and the platform's code - you have to build it yourself. Nothing in the SDKs
