The MPLS-OPS Archive

[Christopher Lewis <chrlewis@cisco.com>]

FRR=Fast Re-Route
www.ietf.org draft-pan-rsvp-fastreroute-00.txt
Inter AS = having one VPN span two providers networks and CsC = Carriers' 
Carrier http://www.ietf.org/internet-drafts/draft-ietf-ppvpn-rfc2547bis-01.txt

At 01:14 PM 1/31/2002, Saqib Jang wrote:


>What is "FRR", "Inter-AS" and "CSC".
>
>thanks,
>Saqib
>
>-----Original Message-----
>From: Robert Raszuk [mailto:raszuk@cisco.com]
>Sent: Thursday, January 17, 2002 1:58 AM
>To: Nanda
>Cc: Krishna Kishore Y; mpls-ops@mplsrc.com
>Subject: Re: MPLS label switching vs IP routing
>
>
>Nanda,
>
> > So in your opinion what is the driving point for MPLS ?
>
>Application one can build on top of it. The key applications being
>deployed today are:
>
>L3 MPLS-VPNs (inf. RFC2547),
>CSC,
>Inter-AS,
>TE,
>FRR,
>DS-TE,
>L2 Transport,
>L2VPNs ...
>+
>and soon GMPLS.
>
>Of course you may find ways to provide the the same results of some of
>the above applications without MPLS at all, but still production
>deployments and scalability characteristics still make the MPLS versions
>of those much more attractive.
>
>R.
>
>
> > Nanda wrote:
> >
> > Hi Robert,
> >
> > So in your opinion what is the driving point for MPLS ? And in which case
>we
> > decide that MPLS should be deployed. Anyway i agree with Krishna's point
> > that it makes routing efficient (for cases next route is same) eventhough
> > finally the end result is same. What about IP checksum modification
>because
> > of the TTL decrements in each hop ?
> >
> > Regards,
> > Nandan
> > ----- Original Message -----
> > From: "Robert Raszuk" <raszuk@cisco.com>
> > To: "Krishna Kishore Y" <kishore@wipro.net>
> > Cc: "Nanda" <nanda@procsys.com>; <mpls-ops@mplsrc.com>
> > Sent: Wednesday, January 16, 2002 9:08 PM
> > Subject: Re: MPLS label switching vs IP routing
> >
> > > Krishna,
> > >
> > > > If I understand correctly, it might not make routing faster but
> > definitely
> > > > efficient, in terms of reduced route lookup time.
> > >
> > > You are putting the question a bit upside down ... if one can already do
> > > line rate lookup in hardware accelarated path for ip addresses it just
> > > doesn't give you much advantage if you make the lookup faster with mpls
> > > :-).
> > >
> > > R.
> > >
> > > PS. A few years back when the mpls started to roll out we measured and
> > > the gain was something about 5% (label swap) for our platforms in the
> > > LSRs and loss about 15% at LERs (label imposition) in compared to plain
> > > CEF switching.
> > >
> > >
> > > > Krishna Kishore Y wrote:
> > > >
> > > > But Robert,
> > > >
> > > > If I understand correctly, it might not make routing faster but
> > definitely
> > > > efficient, in terms of reduced route lookup time.
> > > > If one has large number of routes in the forwarding table and next hop
> > for
> > > > all of them is same, I understand that one single label replaces this
> > entire
> > > > table. So reduction in routing/forwarding table entries.
> > > > I think it's a step ahead of CEF (which is better than fast
>switching).
> > > >
> > > > Please correct me if I am wrong.
> > > >
> > > > Kishor
> > > >
> > > > -----Original Message-----
> > > > From: Robert Raszuk [mailto:raszuk@cisco.com]
> > > > Sent: Wednesday, January 16, 2002 3:58 PM
> > > > To: Nanda
> > > > Cc: mpls-ops@mplsrc.com
> > > > Subject: Re: MPLS label switching vs IP routing
> > > >
> > > > Nanad,
> > > >
> > > > > Now since the IP routing can be done at wire speeds for using the
> > > > > latest Network Processors - does the label based switching in MPLS
>has
> > > > > any advantage in terms of speed compared to the IP routing
> > > >
> > > > No it never did.
> > > >
> > > > > generally in the white papers i was reading they always project that
> > > > > Label based switching makes the IP routing faster - but this is no
> > > > > longer true.
> > > >
> > > > That is just marketing :).
> > > >
> > > > R.
> > > >
> > > > PS. The same misconception is often heard as: "MPLS gives you QoS" -
> > > > nothing more mistaken - if you don't have IP QoS MPLS will not help
>you
> > > > in making your netowrk QoS enabled It can enhance the existing QoS
> > > > feature set but not replace it.
> > > >
> > > > > Nanda wrote:
> > > > >
> > > > > Hi,
> > > > >
> > > > > Now since the IP routing can be done at wire speeds for using the
> > > > > latest Network Processors - does the label based switching in MPLS
>has
> > > > > any advantage in terms of speed compared to the IP routing (Since
> > > > > anyway the same result is got using the Network Processors) ? Anyway
>i
> > > > > know that MPLS gives lots of other features like QoS,TE,VPNs etc.
>But
> > > > > generally in the white papers i was reading they always project that
> > > > > Label based switching makes the IP routing faster - but this is no
> > > > > longer true.
> > > > >
> > > > > Can somebody give some more thaughts on this ?
> > > > >
> > > > > Regards,
> > > > > Nandan
> > > >
> > > > -------
> > > > The MPLS-OPS Mailing List
> > > > Subscribe/Unsubscribe:  http://www.mplsrc.com/mplsops.shtml
> > > > Archive: http://www.mplsrc.com/mpls-ops_archive.shtml
> > >
> > > -------
> > > The MPLS-OPS Mailing List
> > > Subscribe/Unsubscribe:  http://www.mplsrc.com/mplsops.shtml
> > > Archive: http://www.mplsrc.com/mpls-ops_archive.shtml
>
>-------
>The MPLS-OPS Mailing List
>Subscribe/Unsubscribe:  http://www.mplsrc.com/mplsops.shtml
>Archive: http://www.mplsrc.com/mpls-ops_archive.shtml
>
>-------
>The MPLS-OPS Mailing List
>Subscribe/Unsubscribe:  http://www.mplsrc.com/mplsops.shtml
>Archive: http://www.mplsrc.com/mpls-ops_archive.shtml

-------
The MPLS-OPS Mailing List
Subscribe/Unsubscribe:  http://www.mplsrc.com/mplsops.shtml
Archive: http://www.mplsrc.com/mpls-ops_archive.shtml