The MPLS WG Archive

Cell Relay Retreat>MPLS WG Archive>month:2001-Aug> msg00156



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

Fwd: RE: [IP-Optical] Re: An algorithm for calculating allpaths between an s-d pair

  • From: Durga Gangisetti <durga@UU.NET>
  • Date: Thu, 16 Aug 2001 23:42:21 -0400
  • Cc: tziou@ics.forth.g, awduche@movaz.com, mpls@UU.NET, IP-Optical@lists.bell-labs.com, mpls-ops@mplsrc.com, ilazar@mplsrc.com

Ramesh,

>Summary: KSP (not the naive algorithm, but the algorithm as given in my
>book) is optimal and computationally fast.

I agree with you.  

But I think in case of multicommodity flows where non-linear variables like
QoS
and dynamic routing (not fixed routing - i.e; pre-determined) by splitted
traffic
with identical characteristics come in effect, it is at that point :
'  KSP doesnot may always generate the optimal disjoint path set. Whereas MF 
   is kind of guaranted to produce the max no:of possible disjoint paths
and may
    not produce the actual path set. '
it may be valid. Ofcourse above lines are sensitive and arises
architectural and
engieering issues in building an optimal and minimum cost optical network and
differs based on context.


Regards,

Durga,
UUnet Technologies Inc,
A Worldcom Company.