On Thu, 12 Aug 2004 17:07:37 -0700 (PDT), Guy Harris <gharris@xxxxxxxxx> wrote:
> Joshua Schoolcraft said:
> > Scenario #1: The size of the IP packet exceeds the size of the CCSDS
> > transfer frame.
> > In this case, there will exist telemetry transfer frames which contain
> > no IP header and only a partial IP payload, or contain an IP header
> > with a partial payload, padded by a preceding fragmented payload.
>
> > The transfer frame header contains a pointer to the first IP header in
> > its payload (if it exists), so I can at least reliably find a single
> > IP header in the mix. Is there any way, however, to decode the IP
> > payload when it spans across multiple CCSDS frames?
>
> Yes, *if* you reassemble the pieces of the IP datagram.
>
> > Scenario #2: The size of the IP packet is small enough that multiple
> > packets will fit in a single frame. In this case, I can use the
> > previously mentioned pointer to pass Ethereal the first IP packet, but
> > I'm not sure how to make it dissect each additional one inside the
> > frame's payload.
>
> Have a loop in the CCSDS dissector, which calls the IP dissector for each
> IP datagram, passing it a tvbuff that contains only the data for that IP
> datagram.
>
> Without a spec for that protocol and for the way IP datagrams are
> encapsulated within it, I can't give full details on how this would be
> done, as I don't know how you'd know where the IP datagram boundaries are.
> Are those specs available?
The official spec being used for the telemetry frame is here:
http://www.ccsds.org/docu/dscgi/ds.py/Get/File-105/100x0g1.pdf
To see the frame format, check page 30 (listed as 3-11 in the index).
We are not currently implementing the secondary header.
A rough spec for the IP encapsulating software can temporarily be found here:
http://fc.umit.maine.edu/~pokey/ip_over_ccsds/IP_OVER_CCSDS.doc
Figure 4 on page 4 has a flow chart for IP encapsulation.
Regards,
Josh Schoolcraft
