Abstract:

In order to reduce the limits of time synchronization and cooperation between nodes and reduce the measurement traffic in topology inference techniques based on tomography, a selfadaptive measurement method that measured delay variation and a selfadaptive topology inference algorithm based on delay variation are put forward. The feasibility and correctness of selfadaptive topology inferencealgorithm based on delay variation are analyzed. The algorithm is validated through simulations by NS2. The simulation indicates that the selfadaptive topology inference algorithm based on delay variation can infer network topology better than the topology inference based on one way delay and had little limit.