[ippm] [e2e] Mathematical analysis of e2e lable switching path
Ayyasamy, Senthilkumar (UMKC-Student)
saq66 at umkc.edu
Thu Oct 3 16:33:17 PDT 2002
> > My thought is, even for traffic engineering MPLS is not
> > necessary. The benefits you get from constrained routing
> > can easily be achieved by means of some OMP in OSPF. Researchers
> > have achieved the same Traffic engineering objective by optimizing
> > OSPF weights in backbone network.
> Some operational IP networks have successfully used OSPF weights
> for traffic engineering in a large backbone; so this claim is
> more than
> just research. This approach requires a backbone eng staff with
> more than the usual amount of clue about OSPF and usually seems
> to require at least one person with clue about statistics.
I agree. The basic cost/objective function has to come from a
operational staff who has experience in tweaking OSPF weights.
Then, minimization/maximization of the objective function has to be
carried by some heuristic method... combinatorial/linear
programming/simulated annealing methods/search heuristic(which
ever one is familiar.) Some of AT&T research's works will throw more
light on such techniques.
> In each case that I'm personally familiar with, S-Plus (or R) was an
> important part of the analysis tool set. It can work quite well.
> The networks I'm familiar with considered this ability to tune their
> OSPF deployment to be a major competitive advantage.
More information about the end2end-interest