From: hedrick@geneva.rutgers.edu (Charles Hedrick) Subject: Re: Simple Q's on SLIP, DIP and a bad trip. Date: 1 Aug 1993 07:15:45 GMT
cs911461@ariel.yorku.ca (CHRISTIAN D. ARMOUR) writes:
>(1) Does DIP default to CSLIP? Our server will not do CSLIP.
>If it is doing CSLIP, can I make it operate in the other mode?
The kernel defaults to CSLIP. To disable compression, go
into ..net/inet/slip.c, and remove the call to slhc_compress
in sl_encaps. That is, remove the two lines marked below
with >. This is an FAQ. Could it be put in some FAQ document?
static void
sl_encaps(struct slip *sl, unsigned char *icp, int len)
{
...
p = icp;
> len = slhc_compress((struct slcompress *)sl->slcomp, p, len,
> sl->cbuff, &p, 1);
>(2) Our terminal server only gives out an IP address. This, I
>figure, is to prevent unauthorized persons from using yorku as
>a gateway (no authentication precedes connection to terminal
>server and hence, slip connections). The terminal server
>operates like so:
> nova> slip
> Your IP address is 130.63.122.87.
>As you can see, no server address is given. Having read the FAQs
>approx. 500 times, I cannot at this point see any clear solution.
>I know the address of nova.ts.yorku.ca to be 130.63.122.11 but
>what are the odds that this machine is itself the SLIP server?
Frankly, it doesn't matter what address you use, as long as it's on
the same network as your own address. You're going to set up a
default route so that all packets get sent to that address. So even
if you get the address wrong, it doesn't make any difference: All
packets except those addressed to your loopback address are going to
get sent down the SLIP line anyway. It's up to the terminal server at
the other end to figure out what to do with them.
But the correct address would be the IP address of the machine that's
doing the SLIP handling. That would simply be the address of nova, in
your case. By SLIP server, I simply mean the machine that's providing
SLIP service for you. At least with Cisco terminal servers, there's
not a separate address for SLIP. (This is not necessarily true for
other implementations of SLIP. Sometimes the SLIP server has a
different address for its end of each phone line.)
By the way, according to the HINFO record, that box is a cisco ASM/3.
The current release of Cisco's software (which has been out for quite
some time) will support compressed SLIP. It's just a matter of
turning it on in the configuration file. I strongly recommend it.
Without header compression, interactive response over a SLIP
connection is painfully slow. You might be better off using term or
some other connection method, if you can't get your staff to turn on
SLIP.
>(3) The NET-2 FAQ is not too clear on one point... if you are
>using SLIP and loopback, then I would imagine that you would have
>to fully configure your net (ie. get a unique IP number and all
>that terribly unfortunate business). Just how wrong am I?
For loopback you use 127.0.0.1. For SLIP you use the address printed
when you enable SLIP on the terminal server. You don't need any other
address.
>(4) Will SLIP, ideally, allow me to say, ftp remote machines in
>all the world with my local software, telnet to anyplace in the
>world with local software, play netrek on remote servers with
>local clients (my main motivation! :-) or is it just a glorified
>kermit?
SLIP is real TCP/IP. In principle you can do anything you could do on
an Ethernet. However performance at dialup speeds may not be good
enough for some applications to be usable. I don't know enough about
netrek to know whether it's practical at 9600 bps or not.
The primary advantages over kermit are (1) the ability to do a wider
variety of things -- not just logging in and doing file transfer, but
things like X. (2) the ability to have more than one session open at a
time. I often need to use several different machines. I have a
window open on each one. That's not possible in Kermit, though there
are ways other than TCP/IP to do it (e.g. term).