The MPLS WG Archive

[Curtis Villamizar <curtis@workhorse.fictitious.org>]

In message <009401c06704$892669a0$f23987cb@faisal-s>, "Faisal S. Naik" writes:
> 
> Hello friends,=20
> At a recent discussion with my MS class students, i was discussing MPLS =
> and related issues. The discussion went into a complete confusion when =
> we got stuck at the very question=20
> WHY use MPLS?
> Is it only because of the fact that the efficient swapping mechanism, =
> that MPLS is considered good. What about already accepted switching =
> techs like ATMs, incase they are integrated with IPv6 networks for =
> better services like QoS through flow labels?
> would anyone be kind enough to clarify?
> Regards,
> Faisal


Hi,

I just read the answers on this an you guys all got it wrong.  :-)

The scaling argument was close, but hierarchy was needed in some way
that could make the overall solution scale, not just allow the PNNI
routing to scale but prevent scalability in the upper layers.

There were a number of motivations for getting rid of ATM in IP
networks.  One was the cell tax which at a typical 20% would not alone
kill ATM.  Another was SAR.  You can't get fast ATM router interfaces
because SAR speed and SAR buffering becomes a problem.  This alone was
enough to kill ATM.

Perhaps even larger was the problems of independent control planes and
the effect on IGP scaling.  Flooding is needed for reliability in the
IGP and flooding in a full mesh of N routers has some N^3 properties
with regard to messages sent when a router in the mesh fails and N^2
when a VC fails.  Attempts to constrain the flooding tended to slow
convergence and sometimes cause long lived IGP inconsistency when
change occurred.  This is not a problem is the flooding follows the
physical interconnections as it does with MPLS since the number of
adjacencies per node drops dramatically.

The IGP scaling problem was address by ISPs by partitioning their
network into core and regions but this reduced the effectivenes of TE.

There was also no way to change a BW reservation amount and no
make-before-break rerouting.  If a better path was available, ATM
implementations would not use it (or would not do so for very long
periods of time).  There was more attention to adaptivity in MPLS.

Lack of local-protect was at one point seen as a limitation.

There were also implementation issues.  ATM vendors were slow to add
buffering to the switches and never added enough.  They were slow to
do EPD/PPD and never did RED, pushing ABR way too long (which never
worked well in practice).  There was reluctance to give up on the CAC
religion and allow overbooking.  Layout algorithms were not very good
for what the ISPs wanted to do and results were poor.

Curtis