Network Tomography: A novel algorithm for probing path selection

The impressive growth experienced by the Internet in the last years, together with the diversity in network technologies are posing more and more challenges in the field of network monitoring, where novel approaches, able to obtain an integrated view among multi-technological and multi-vendor domains, are required. In such a context, Network Tomography (NT) has emerged as a very promising monitoring technique, being able to infer unobservable network performance just relying on end-to-end (E2E) measurements. This work proposes a novel algorithm that permits to overcome one of the main limitations of NT, that is the proper choice of a minimal set of E2E probing connections, which provides a full monitoring coverage of the network. The proposed approach allows to find automatically the minimal set of probing paths thanks to a proper combined use of the Yen algorithm and the Gauss reduction method. We demonstrate the effectiveness of our algorithm via simulation analysis.