> From: Curtis Villamizar (curtis@workhorse.fictitious.org)
> Date: Wed May 09 2001 - 08:18:13 EDT
>
>
> MPLS can put the A-E traffic on one leg and the A-F traffic on the
> other leg but cannot accurately balance the load. The only technique
> so far that can accurately balance the load is OMP. With lots of MPLS
> LSP, the CR based MPLS/TE load balance is generally not as good but
> "good enough". With OSPF and ECMP metrics can be set so that load
> balance is not as good as MPLS/TE. Some consider the ability of the
> IGP to load balance to be good enough. Others don't consider it good
> enough and consider IGP load balance too hard to manage and that was a
> principle motivation for MPLS/TE.
Curtis,
I dont think creating many MPLS LSPs itself will help you on
traffic engineering. MPLS LSPs are *virtual circuits" - it is
the mapping of these virtual circuits to the physical links
that achieves optimization. Each edge-to-edge traffic demand
can be routed over an MPLS LSP, but how this MPLS LSP is routed
through the physical topology determines the capacity allocation
and traffic distribution.
In one of our early work, we developed a mathematical model
for MPLS-based TE (the equation is mentioned in that paper also).
Basically, the problem can be modelled as a linear programming
problem, and it is possible to get the precise optimal routes for
MPLS LSPs given a topology and edge-to-edge demands. However,
the optimal solution may also require the splitting of demands
over multiple LSPs. Nevertheless, some simple approaches
we simulated can achieve very good results without splitting
demands.
Cheers
Zheng
This archive was generated by hypermail 2b29 : Mon Aug 04 2003 - 04:10:03 EDT