RICERCANDO: Data mining toolkit for mobile broadband measurements

摘要

Increasing reliance on mobile broadband (MBB) networks for communication, vehicle navigation, healthcare, and other critical purposes calls for improved monitoring and troubleshooting. While recent advances in monitoring with crowdsourced and network infrastructure-based methods allow us to tap into a number of performance metrics from all layers of networking, huge swaths of data remain poorly explored due to a lack of tools suitable for fast, interactive, and rigorous MBB data analysis. In this paper we present RICERCANDO, a solution that enables rapid exploration of large heterogeneous MBB measurement data as well as the identification and explanation of unusual patterns detected in such data. RICERCANDO consists of a preprocessing module ensuring that time-series data is stored in the most appropriate form for mining, a rapid exploration module enabling iterative analysis of time-series and geomobile data to detect and single-out anomalies, and the advanced mining module that lets the analyst deduce root causes of observed anomalies. We implement and release RICERCANDO in open-source, and validate its usability on case studies from a pan-European MBB measurement testbed.