Date: Sun, 7 Jan 1996 22:39:02 +0000 () From: James Raynard <fqueries@parody.tecc.co.uk> To: Gary Kline <kline@tera.com> Cc: kline@tera.com, wollman@lcs.mit.edu, jbarrm@panix.com, questions@freebsd.org Subject: Re: ppp from shell, & starting X Message-ID: <Pine.BSF.3.91.960107212651.288A-100000@parody.tecc.co.uk> In-Reply-To: <9601040142.AA01597@tera.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, 3 Jan 1996, Gary Kline wrote:
> > #static_routes="multicast loopback"
> > #route_multicast="224.0.0.0 -netmask 0xf0000000 -interface ${hostname}"
> > #route_loopback="${hostname} localhost"
>
>
> Why do you have the above 3 lines commented out? Or, more
> to the point, what's the purpose of multicast and loopback?
Loopback allows your machine to send packets to itself using a "fake"
IP address of the form 127.x.x.x. On BSD TCP/IP implementations, these
packets are intercepted by the kernel and don't pass through any network
devices, so you can run networking programs without even being connected
to another machine! As they do however go through all the TCP/IP code,
this makes loopback very useful for testing.
Traditionally, you could either send packets to a single machine
("unicasting") or send them to every machine on your network
("broadcasting"), which is inefficient if you only want to send
them to a limited number of machines.
Multicasting allows you to restrict the number of machines you send
packets to - as an added bonus, they don't even have be on your network.
To achieve this, a special set of IP addresses is used, in the range
224.0.0.0 to 239.255.255.255 (known as "Class D addresses"). It's often
used for audio and video transmissions - you may have heard of the MBONE
(multicast backbone) in this connection.
The reason why I have these entries commented out is that there is no
point in having static multicast and loopback routes on a stand-alone
machine, so I don't bother.
My policy is to ignore routing completely - I have PPP configured so that
it sets up a default route to the other machine when it connects. This
means that all packets which are not for my machine get sent to the other
machine, which takes care of getting them to the right destination (and
of getting the replies back to me, of course!).
Note that these sysconfig entries only set up static routes - they don't
make any difference as to whether or not use you can actually use the
loopback and multicast interfaces. I have also seen suggestions that
they don't work anyway, which is another reason for commenting them out 8-)
(Disclaimer - I've over-simplified in the above, for example implicitly
assuming that a "machine" only has a single IP address and that networks
aren't sub-netted, but I did say I'm not a guru 8-)
> I've been working on getting Taylor uucp working so that when
> I am _not_ working via PPP that mail will get across via uucp.
> Can I set it up so that mail will be received/sent across when
> I am connected by PPP *and* by uucp? Or is this too much to
> expect?
I know practically nothing about UUCP, but I don't see why not. In fact,
I've seen references to people running UUCP on top of a TCP/IP session
(which is presumably running on top of a PPP/SLIP connection!), so it
must be possible. *How* is another question, though.
James
Segmentation fault (core dumped): cannot find file '.signature'
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.3.91.960107212651.288A-100000>