MPLS TE Offline Computation

Introduction

The main method of computing MPLS TE is using information from a Link State IGP. Another method is using an offline computation and importing the topology based on that computations.

MPLS TE Offline Computation

  • The way offline computation works is by exporting the network topology to some software that is able to create optimal LSP placement.

Advantages

  • Exact control of paths - no dynamic/unknown path placement.
  • Ability to cross area and AS boundaries - as long as the topology information imported has a global view, areas and AS boundaries don't matter when creating these LSPs.
  • Account for normal and failure instance - offline computation can account for normal functionality where all links are up and in failure scenarios where multiple links are down. The software could iterate through all possible failure scenarios and pick the best LSP topology.
  • Optimality of the solution - the optimal solution can use other than CSPF algorithm and take much longer to analyze all possible scenarios. Other algorithms can minimize bandwidth usage, minimize the number of LSP changes, achieve best protection from failures…etc.

Disadvantages

  • Order of upgrades counts - following a re-computation not only the final topology is important but also the order of changing it to that final state.
  • Not very dynamic - any changes to the network's topology would require re-running the offline computation which might take a lot of time.
  • Hard to do a global change - incremental might be the only method of deploying this new computed LSP topology. Global implementation might be too intrusive.
  • Quality of output dependent on the input - if the input of the topology is not very accurate, the output will not be very good either.

Tools

Comments

rating: 0+x
Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-ShareAlike 3.0 License