作者: Brian Trammell , Dominik Schatzmann
DOI: 10.1109/IWCMC.2011.5982724
关键词:
摘要: This work applies snack, a flow-based algorithm for detecting connections between clients and supernodes to the Skype network, study of two separate outages on network in August 2007 December 2010. We examine both detail, comparing them each other, previous first outage. find that simple metric derived from connection events per unique client as measured by snack acts proxy would have provided forewarning outage, may applicability an indicator general health. also faster onset recovery 2010 discovering distribution supernode is measurement centralization, detects apparent present strategy such outages. These metrics be applied any operator monitor health network.