Cell Relay Archive

Cell Relay Retreat>List Archive>month:1997-Nov> msg00176



[Date Prev][Date Next][Thread Prev][Thread Next]  
  [Date Index][Thread Index][Author Index][Subject Index]

Re: IP Switching-TTL

  • From: holdrege@eisner.decus.org (Matt Holdrege)
  • Date: Tue, 25 Nov 1997 10:29:57 -0500 (EST)

Organization: DECUServe
Message-ID: <25NOV199710131296@eisner.decus.org>
References: <64uchc$lv3$1@dismay.ucs.indiana.edu> <64vcuc$3q$1@shell6.ba.=
best.com>
X-Trace: news.decus.org 880470793 19733 HOLDREGE [192.67.173.2]
X-Nntp-Posting-Host: eisner.decus.org
Date: Tue, 25 Nov 1997 15:13:00 GMT
Lines: 38
Xref: news.indiana.edu comp.dcom.cell-relay:12008

In article <64vcuc$3q$1@shell6.ba.best.com>, bannai@best.com (Vinay Banna=
i) writes...
>In article <64uchc$lv3$1@dismay.ucs.indiana.edu>,
>Nafez Aboulibda  <nafez@igdv.fh-darmstadt.de> wrote:
>>My name is Nafez Aboulibda, I'm doing my MSc project.
>>There is some thing relating to the TTL field in the Flow identifier in=

>>Ipsilon=92s publications
>>"IP switching" that I couldn't understand.
>>Since this field is not transmitted within the packet and their issue i=
s
>>to retain the function of
>>TTL.
>>My question is : where is  the recovery procedure explicitly done to
>>decrement the TTL field
>>on each node .
>>Ipsilon's words" the TTL field is not transmitted in the packet but it
>>is recovered from the
>>info. stored at the destination"
>>
>>with regards.
>>Nafez Aboulibda
>>
>>
> =

>This is something of an issue for all IP switching and tag switching
>folks. Since the packets (or the cells of the packets) are switched till=

>they hit the egress node, there cannot be any TTL decrements. Tag folks
>want to check the TTL value at the ingress point and based on the tag fi=
nd
>out how many swithing nodes it has to traverse to the egress point and
>decrement that value from the TTL at the ingress. I have not been keepin=
g
>up with the MPLS for a while so I am not sure what their position is rig=
ht
>now. As far, the IFMP, I think what they do is somehow code the TTL valu=
e
>into the flow identifier and at the egress point of the switching core
>network, they re-introduce the TTL value back into the IP header. I don'=
t
>think they decrement the TTL value either at the ingress or the egress. =


Ascend's IP Navigator which is based on MPLS does decrement the TTL field=
=2E It
also explicitly prevents routing loops which is obviously important with =
the
high bandwidth connections used.