| Summary: | Recovery from network failures is a very important and challenging requirement
for current Internet service providers. Real-time applications and other mission-critical
tasks require networks to respond to network faults expeditiously in order
to not degrade the required quality.
In IP networks, failures are detected and handled through the convergence
of network topology (topology change reflected in all routers' routing information
bases). This recovery mechanism, however, has an obvious disadvantage in that
it reacts to network faults very slowly, because it may take a long time for the
entire network topology to converge. In this thesis, we present the design of a fast
recovery mechanism in MPLS networks, implemented over the Linux platform. We
investigate its performance over our testing network composed of Linux routers. Our
experiments show that our recovery mechanism responds to network faults quickly,
with less packet loss. === Science, Faculty of === Computer Science, Department of === Graduate
|
|---|
