On Mon, 11 Nov 1996, dennis wrote:
> According to a bunch of people.... > > >According to Daniel O'Callaghan:
> >> and is totally unnecessary, as Denis says. Simply put, the
difference is > >> that you are running a ppp link within a single IP
network (happens to be > >> class C), while James is running a ppp link
between two distinct IP networks. > >> You: 193.56.58.20 --> 193.56.58.234
> >> James: 203.16.20.1 --> 203.8.105.20 > > It still doesnt make a
difference....setting up routing over the PPP link > and defining the
netmask of the serial line are two different things... > the bottom line
is that you are using direct routes to hosts (not via a > net) when
getting from here to there on the link itself. From a > routing
perspective (where the issue is next hop), the next hop > is the host at
the end of a Point to point modeled network rather > than a gateway on a
network or subnetted network. Defining it > as a network is stupid,
because there is no net...there are only > 2 peers.
TrueAccording to a bunch of people....
>
> >According to Daniel O'Callaghan:
> >> and is totally unnecessary, as Denis says. Simply put, the difference is
> >> that you are running a ppp link within a single IP network (happens to be
> >> class C), while James is running a ppp link between two distinct IP networks.
> >> You: 193.56.58.20 --> 193.56.58.234
> >> James: 203.16.20.1 --> 203.8.105.20
>
> It still doesnt make a difference....setting up routing over the PPP link
> and defining the netmask of the serial line are two different things...
> the bottom line is that you are using direct routes to hosts (not via a
> net) when getting from here to there on the link itself. From a
> routing perspective (where the issue is next hop), the next hop
> is the host at the end of a Point to point modeled network rather
> than a gateway on a network or subnetted network. Defining it
> as a network is stupid, because there is no net...there are only
> 2 peers.
True, but sliplogin or slattach or gated (something, I can't remember
which) *does* apply the interface netmask to decide which hosts are
gatewayed by the remote end. Probably just some fudging, as you say.
Danny