Posts made by Stevanov

@skywatch while i obviously agree with the fact that i we can best change channels, i figured that while we're at it we might as well add encryption too. I don't insist on using any specific channel. I couldn't care less what channel's being used

That said, i tried recompiling my pi gateway wih the password and one sensor to start with using #define MY_ENCRYPTION_SIMPLE_PASSWD "mypassword" (shut off all other sensors). But i'm seeing all sorts of erroneous messages coming by, the one node is showing up on the gateway as a different ID every time.

I'll whip up a simple serial gateway next time i get the chance and try to use #define MY_ENCRYPTION_SIMPLE_PASSWD "mypassword" on both to see if the problem lies with how i configure the pi gateway

THanks for all the input everyone so far

@alowhum said in Problems with the neighbours:

MY_ENCRYPTION_SIMPLE_PASSWD

From what i can gather on the git repo i see that MY_ENCRYPTION_SIMPLE_PASSWD is only used in MyTransportRFM69.cpp, so it seems that for the NRF24 that we use we are going to have to do some other trickery to get encryption to work.
I prefer the encryption over switching a channel but i'm struggling to find out how it works

Thank you all, i'll definitely try to change channels and use the

#define MY_ENCRYPTION_SIMPLE_PASSWD "mypassword"

This last options means i have to compile the raspberry gateway again, correct? I see it has an AES flag in the ./configure and the config. I assume that's what i have to use?

Were utilizing the cheaper NRF24 radios for our network

I'm a volunteer at the local Makerspace and we use mysensors as a nice and simple way to hook up peoples projects to our local network without having dozens of ESP8266s or whatever all over the place. About half a year ago a new tech company moved in next door and it seems that during the evening hours the employees there get to work on hobby stuff because since then our Controller is being flooded with temperature, gas, light, motion etc, sensors that isn't ours. There are at least 40 sensors in the list that aren't created by us. Weve also had problems of their sensors that have an ID that one of our sensors also uses, overriding stuff in our database, creating some very odd power usage and temperature graphs.

I've thought of solving this with the signing or encryption thing. But it seems that that is computationally complex and wed need to add the little SHA chip to most our devices. Sadly i think we need to reprogram all our stuff. It doesn't seem like the neighbours mind our sensors messing with their network.

My question is does anyone have any recommendation? We do have code available for almost all devices in our net so it should be doable to reprogram most of them. I'm not so much worried about security, it's just annoying that the neighbours' stuff gets into our database.

Maybe theres an option to switch channels or to do some sort of handshake, a passphrase, or whatever? The signing and security page on the mysensors website seems sort of theoretical and i cant really find any examples of what to do. (I can't imagine it's as simple as only setting #define MY_SIGNING_SOFT and #define MY_SIGNING_SOFT_RANDOMSEED_PIN 7 in every sketch)

Any input would be very much appreciated!

Thank you for the thorough reply @mfalkvidd

I already had a look the pages you linked but some of the questions I had were only partially answered by those pages so I appreciate you taking the time to answer all of them.

Judging by your answers i'm not going to try two gateways. I will instead look into building an ethernet gateway or perhaps forwarding data from one controller to the next.

Signing is definitely something i want to implement too as my gateway seems to be picking up pacakges from a mysensors user near me!

Especially your statements about undefined behaviour in regards to ACKs were helpful.

Thanks again and apologies for replying as late as i did.

I'm looking into switching controllers (or perhaps using two controllers for a while to see which i prefer). What's worrying me is that i'm not 100% sure on what happens when i expose sensors to two gateways.

My question therefore is where can i find information on how sensors and gateways and controllers actually interact. So i can find out things like;

• is there a routing table on each sensors and is it possible for a sensor to transmit to two gateways attached to two separate controllers
• i am aware that radio is literally just a broadcast and anyone listening will hear, but will a gateway/controller 'drop' a package from a source it doesn't know? (like they have not gone trough pairing
• what is the behavior of a sensor broadcasting and receiving an ACK if there are two?
• does the controller actually do much?, i know some behavior is built into the (i'm generally using serial) gateways like sending out a discovery request on boot, does the controller keep track of ID's and assign them, or does a controller merely do responses for data that sensors ask for
• etc

I'm aware that i could just dig through the code and find this out myself but i'm hoping somebody has a write up on this or some design documentation i haven't been able to find myself yet.

Any input will be greatly appreciated!