| |
PA2AGA > TCPDIG 02.01.98 23:01l 127 Lines 4305 Bytes #-10115 (0) @ EU
BID : TCP_97_99A
Read: GUEST
Subj: TCP-Group Digest 97/99A
Path: DB0AAB<DB0SL<DB0RGB<DB0ABH<DB0SRS<DB0AIS<DB0NDK<DB0ACH<PI8JOP<PI8ZAA<
PI8GCB<PI8MBQ<PI8VNW
Sent: 980102/1353Z @:PI8VNW.#ZH2.NLD.EU #:2833 [Hoek v Holland] FBB5.15c
From: PA2AGA@PI8VNW.#ZH2.NLD.EU
To : TCPDIG@EU
Received: from pa2aga by pi1hvh with SMTP
id AA129 ; Fri, 02 Jan 98 13:37:47 UTC
Received: from pa2aga by pa2aga (NET/Mac 2.3.64/7.1) with SMTP
id AA00005710 ; Fri, 02 Jan 98 14:17:56 MET
Date: Fri, 02 Jan 98 14:13:42 MET
Message-Id: <tcp_97_99A>
From: pa2aga
To: tcp_broadcast@pa2aga
Subject: TCP-Group Digest 97/99A
X-BBS-Msg-Type: B
TCP-Group Digest Fri, 2 Jan 98 Volume 97 : Issue 99
Today's Topics:
Multiple Routes (2 msgs)
NETCPA Meeting, Jan. 10
Send Replies or notes for publication to: <TCP-Group@UCSD.Edu>.
Subscription requests to <TCP-Group-REQUEST@UCSD.Edu>.
Problems you can't solve otherwise to brian@ucsd.edu.
Archives of past issues of the TCP-Group Digest are available
(by FTP only) from ftp.UCSD.Edu in directory "mailarchives".
We trust that readers are intelligent enough to realize that all text
herein consists of personal comments and does not represent the official
policies or positions of any party. Your mileage may vary. So there.
Loop-Detect: TCP-Group:97/99
----------------------------------------------------------------------
Date: Thu, 1 Jan 1998 13:39:50 -0800 (PST)
From: Bill Vodall <wa7nwp@bigsky.com>
Subject: Multiple Routes
I've always heard that the TCP network was designed to be
robust. That is, it will keep working if one path fails.
It appears, as I understand it, that this is not at all
the case with the 44 AMPR.ORG net.
Ignoring for the moment, the critical Mirrorshades router, is
it possible to have backup gateway paths?
Say I'm on a RF only subnet a couple hops from any gateway. My
local IP number would be 44.82.73.5. (All numbers here are made up.)
There could be 3 gateways serving the region. 44.82.201.1, 44.82.202.1
and 44.82.203.1.
Assume that 44.82.201.1 is my main gateway to which I'd have a nice
medium speed path (T1/2).
Assume also that I have a local route to the second gateway [44.82.202.1]
via a couple slow speed (56k) interties..
And to take it this example all the way, say I also have an even slower
and somewhat unreliable route to the third gateway, [44.82.203.1].
Now, under normal circumstances, all my 44.82.73.5 packets would
blast through the 201.1 gateway. Then, as happens all too often,
the gateway could die due to a software crash, the hard drive dieing,
or the supporting ISP would go out of business. Even though
there are still two useable routes, I'd be cut off from AMPR.ORG
connectivity.
Is there software magic to keep the routes alive? To update the
routing dynamically?
I'm aware that RIP handles some of this on a local basis but
I don't think it's used as far as the gateways are concerned.
73,
Bill - WA7NWP
------------------------------
Date: Thu, 1 Jan 1998 14:43:52 -0800 (PST)
From: Brian Kantor <brian@karoshi.ucsd.edu>
Subject: Multiple Routes
>I've always heard that the TCP network was designed to be
>robust. That is, it will keep working if one path fails.
No, the internet protocols PERMIT robustness. Designing a network
which *IS* robust is a difficult proposition. Few are.
>It appears, as I understand it, that this is not at all
>the case with the 44 AMPR.ORG net.
Not surprising, since it's supposed to be an EXPERIMENTAL network.
Anyone depending on any part of network 44 to be a robust or even
production grade network is due for disappointment.
>Ignoring for the moment, the critical Mirrorshades router, is
>it possible to have backup gateway paths?
> ...
>Is there software magic to keep the routes alive? To update the
>routing dynamically?
That's one of the things we're supposedly experimenting with.
For example, Fred's RSPF shows promise but is not widely deployed
so we really don't know whether it'll do the job.
>I'm aware that RIP handles some of this on a local basis but
>I don't think it's used as far as the gateways are concerned.
Correct - the gateways are currently statically routed. Without some
sort of gateway routing protocol (which we also don't have yet), that's
really about all we can do right now.
Feel like designing and testing solutions to this problem?
Well, GREAT! That's what the network is for.
- Brian
------------------------------
To be continued in digest: tcp_97_99B
Read previous mail | Read next mail
| |
