The MPLS WG Archive

[neil.2.harrison@bt.com]

Zhi Hao Hong wrote 25 September 2001 09:12
> I ever ask this question to Peter Lothenberg, archetect of Sprint.
> He just ask me " why we need mpls".
> He means he can do very good without mpls. Maybe this is the 
> simple answer.
> 
> At 02:05 2001-9-24 +0000, k m a wrote:
> >hi all,
> >i have a question,y the industry is not accepting 
> MPLS......what r the hindrances they r having....what exactly 
> is the reason.see if u could help and we can work out things 
> to convince them.
> >thanx
> >kapil

This is actually an excellent question that should not be dismissed lightly
by the readers of this list.  Like Sprint, people here in BT ask the same
question.  It's no good pointing to the *technical* advantages of MPLS (see
Note).....you have to address the *cost* issues for operators.  That is, if
you can't show that implementing MPLS will either reduce operational costs
(which BTW is perhaps the most critical issue) and/or increase revenues then
I am afraid that those who control the purse strings in operators won't buy
it.  So you are right, it is as simple as that.

Note:  Who is the user? is a key question here.

1	A new ISP might have a different view to an existing traditional
carrier/operator and simply want a 'CO tool' to better manage IP via greater
separation of forwarding/routing functions.  However, if they then want to
be able to offer a wider range of services they will need to consider
emulation of trad L2 managed BW technologies.....and this is exactly why we
see PWE3 activities, PPVPN and MPLS/ATM interworking activities happening.
BTW - This means we simply *have* to  start treating MPLS as a proper layer
network and stop pretending its something 'special'...see next item.

2	A traditional operator comes with important *revenue earning* fine
granularity managed BW services that you ignore at your peril.....in TDM, FR
and ATM flavours.  So from this user's perspective, it is absolutely
critical that MPLS can do more than just be an IP-helper, it needs to offer
a migration solution for all these self-similar managed BW products.  If it
can't, then that will a real hindrance to its uptake/acceptability IMO
because we simply can't afford to have lots of self-similar
technologies.....this a key facet of the critical operational cost issue I
mentioned at the outset.  However, a pre-cursor requirement for this to even
have a chance, is that MPLS must have carrier-strength fault-manangement
(aka MPLS OAM).  If we don't get this then CTO people like me will have a
very hard time 'selling' MPLS internally just on technical merits.  So those
of you who are against getting proper MPLS OAM in place please reflect
carefully on this observation.

regards Neil