Author of the "renard" Sigfox specifications here.
Just to give you and everyone else reading this some background: These
specifications are not "official" documentation by Sigfox, but the result of my
protocol analysis that I recently presented at 35C3 ("Hunting the Sigfox:
Wireless IoT Network Security").
> 1) Would the link-layer header include any radio metadata? For 802.11,
> there are various forms of radio metadata headers, such as the radiotap
> header:
>
> http://www.radiotap.org
>
> If so, what would the format of the radio metadata be?
I would suggest something similar to what the LoRa people did with LoRaTap:
https://github.com/eriknl/LoRaTap
For the uplink, except for version / padding / length of course, I don't think
we need anything other than the uplink frequency (Sigfox uses a continuous
uplink band) and an RSSI value.
For the downlink we should again have frequency and RSSI, but maybe also some
metadata (sequence number / device ID) about the corresponding uplink (Sigfox
is an uplink-initiated protocol, i.e. downlinks are only transmitted after they
have been requested by a corresponding uplink and their scrambling depends on
the uplink metadata, something that is described in patent EP3259864A1).
But I'm not sure whether the link-layer header is the correct place for this
kind of information?
> 2) The 2.2.2 "Implementation by Sigfox" section of the spec shows, on
> page 14, a frame structure.
>
> In the captures, is the first byte of frame data (following the
> radio metadata, if there is radio metadata, or at the beginning of the
> packet, if there is no radio metadata), the first byte of the Preamble/Type?
IMO the first byte of frame data should be the first byte of the "Type", since
the "Preamble" is constant and just something that is used to detect the
presence of a Sigfox uplink inside the uplink band and to synchronize the
receiver.
At least, that is how I currently handle it with renard-phy.
> 3) Is the CRC-16 included with the frames?
Yes, it's included in the uplink transmission.
Thank you two for your work, I'd be very happy to see this being integrated
into libpcap :) !
If you happen to have any further questions, feel free to ask!
- Florian Euchner (Jeija)
_______________________________________________
tcpdump-workers mailing list
tcpdump-workers@lists.tcpdump.org
https://lists.sandelman.ca/mailman/listinfo/tcpdump-workers
