> In answer to your question about X.25 in the comment in the handoff
> registration routine:
[snip]
>
> 1) the protocol has an OSI NLPID
>
> and
>
> 2) the CALL REQUEST was captured.
As far as I know, UCP nor SMPP have a protocol identifier within X.25 (as
indeed I expect many application protocols atop X.25 won't have one).
> We could add a heuristic dissector table, as well, which would be used
> for virtual circuits where either
>
> 1) the CALL REQUEST wasn't captured, and the user didn't specify
> SNA-over-X.25, and the packet doesn't look like an IP packet
>
> or
>
> 2) if the CALL REQUEST was captured but it didn't specify the
> protocol to be used on the connection.
One of the things I like so much about Ethereal is that it'll present you
immediately with a best guess (usually right) without required interaction.
Heuristics will do that for you.
>
> We could also either
>
> 1) add a mechanism - and the infrastructure necessary for it -
> to allow the user to specify what protocol to assume X.25
> traffic is using if we didn't get that from the CALL REQUEST
>
> or
>
> 2) add a mechanism in the UI to let the user specify what
> protocol is to be used for the virtual circuit to which the
> currently selected frame belongs
I'd say option 2 would be preferred (presuming that it would be like the
"Decode as" under TCP/UDP), option 1 might still not cover, for instance, a
trace with both UCP and SMPP traffic in it.
Lot of work?
Regards,
Tom Uijldert Email: Tom.Uijldert@xxxxxx
