Unicast Network Loss Tomography Using #R-Cast Probes Scheme

Feng Qian,Guang-min Hu,Xing-miao Yao

DOI: https://doi.org/10.1109/CMC.2009.365

2009-01-01

Abstract:Network tomography can not rely on the cooperation internal node, and use only a group of edge nodes to reveal the mathematical and statistical characterization of network behavior that is wished to be known. Therefore, network tomography has become one of the focused new technologies. Existing unicast network loss tomography need send large numbers of probe packet, which is lead to add network load. In the paper, we propose unicast network loss tomography using #R-cast probes scheme with two stage mechanism. Here R denotes leaf set (receiver's nodes) of an arbitrary tree, and # is the number of its set. In my study, to eliminate the redundancy probe packet, we send first #R-cast probes to all receivers so that all paths are covered, simultaneously. Then, we exact different flexicast from the #R-cast received by leaf nodes, for all the internal link loss rates identifiability. Our method hence can send much lesser probe than existing method. Finally, we employ fast least-squares algorithm to obtain the solution of our overdetermined equations of my loss tomography. Ns2 simulation experiments demonstrate the performance of the propose approach.