Re: [ipwave] about V2I to traffic lights controller

[Alexandre Petrescu <[e-mail]>]

Le 11/04/2019 à 19:09, Alexandre Petrescu a écrit :
[...]
> can see the preliminaries of it if you try to browse a real traffic
> lights controller but situated - for now - in-lab at 193.48.19.10).  The 
> practical issues we have at this time is the support of IPv4 (RSU only 
> supports IPv4 and that will not change in the next few years) and 
> certificates (you will understand the certificate problem if you try to 
> browse it).

Speaking of which, I am currently looking for someone to test the DIASER 
UDP protocol on IPv4 to query this traffic lights controller.  In a 
secure manner, because the data is sensitive.  I have the python scripts 
ready for DIASER on UDP on IPv4, to query the traffic lights status and 
their expected (well like in weather prediction) state duration.

If successful, then we can talk about writing an individual Internet 
Draft containing an IPv4 RIO in an IPv6 RA over OCB, to allow making a 
direct IP path from the car to talk directly to the traffic lights.  A 
colleague programmer is already interested.

I also asked a few of the current project partners, but they are not 
interested for the moment - lack of time, at this time.  They tell me it 
should be me on the RSU who does that DIASER and then convert it to SPAT 
and upload it to cloud.  When I reply with the end-to-end principle (let 
the cloud query the controller, rather than in-middle convertor), they 
tell me that everybody does it that way (RSU reads controller data, and 
RSU uploads to cloud).  So I am trying to show a different way is 
possible: make the cloud query the traffic lights controller, and the 
car query the traffic lights controller, and anybody who so wishes, yet 
in a secure manner (probably some form of certificates).

That is a current problem that I have.

Alex