Active Measurement of the AS Path Prepending Method

It is widely believed that the popular AS path prepending (ASPP) method is very effective in controlling the inbound traffic for multi-homed AS. The ASPP method artificially increases the length of the AS path advertised in BGP routes with the hope of discouraging the upstream AS from using the prepended routes. But the result of this method is usually unpredictable, i.e. the amount of traffic shifted across the inbound links after prepending cannot be predicted accurately. As a result, this method was often performed in a trial-and-error basis which could increase the convergence time and even introduce congestion in the unprepended links. Moreover, very few measurement studies were reported for the ASPP method. The main objective of this work is to evaluate the effectiveness of the ASPP method from the perspective of a stub, multi-homed AS.