REVIEW: End-to-End Effects of Internet Path Selection

REVIEW: End-to-End Effects of Internet Path Selection

The paper quantifies the effects of factors influencing internet path
selection on the end users in an attempt to provide a qualitative
measurement of the efficiency of internet routing. It is able to provide
such qalitative analysis by first measuring the the quality of alternate
paths, looking at parameters such as round trip time, loss rate and
bandwidth. Its finding is that in 30 to 80% of the cases there is a better
path that can be selected, therefore implying thet from the users
perspective internet routing is inefficient. Moreover, for 10% of those
cases, the alternate path in better by at least 50%.

The use different datasets with geographically diverse internet hosts
constitues a good attempt to have a representative set of hosts from the
internet. An argument for the representativeness of two of the datasets
used is referenced. It is difficult, however, to produce an agreeable
representative set.

The experiment contains, however, a number of biases, as pointed out by
the authors. The study is limited by the inability to determine alternate
paths for packets as well as the hops in those paths, and therefore the
only used alternate paths the authors were aware of. For most of the
datasets used, the paths have not been measured simultaneously, which
means that some paths may have biased estimates as they were measured at a
period when congestions were occurring, while others were not.
Received on Thu Nov 03 2005 - 10:34:49 EST