Fri, 14 Oct 1994 22:43:28 +1000
> From: firstname.lastname@example.org (Christopher Fraser)
> But I thought Stephen Boyd Gowing said:
> > > > I've been avoiding answering the PPP question 'cause I'm not sure just
> > > > how feasible it is without any modems attached.. I expect it'd work fine
> > >
[then Raz said]
> > > Oh - not feasible at all! PPP would be useful (only) over modem lines.
> > > Concievably one could come in to somewhere through a modem (8 data, no
> > > parity, 1 stop) and rlogin to ftoomsh that attempt a PPP attach... most
> > > kernels would barf at such a thing though...
> Yup, this should work. Even via the SoCS modems; PPP is supposed to be able
> to probe the line, detect which characters don't work and only use that do.
> The SoCS modems are still pretty sick though -- no rotary, sick flow control
> when they work and 7e1 is just a sick joke.
Except that SOCS modems are typically connected to SOCS machines. Attempting
to run PPP over an 'rlogin ftoomsh' session seems revolting at best, but may
be impossible - I'd be suprised if SunOS/Solaris would accept a pty as a
> So, what would be have to do get an allocation of IP numbers to play around
> with (or better still, our own subnet?)
Apply to ITD I guess - but I doubt that they'll give us half a percent
of UTS's total IP address allocation. Maybe if we ask SOCS nicely, we can
steal some more of theirs :-)
> On a technical side, is it possible to do packet sniffing from a PPP link?
> (I can't see how -- you'd need to put the ethernet interface on the remote
> machine into promiscuous mode).
Ahh, not really. It certainly should be from Ftoomsh though.