“…At the same time, the incorrect identification of user sessions has clear consequences on traffic analysis: counting multiple users as one, or vice versa, results in incorrect estimates of traffic distribution, user activity, and resource allocation, which affects modeling and simulations and potentially biases the conclusions. Traffic traces containing RNTI information have been proven useful to identify radio resource utilization of mobile traffic patterns [19,20], to fingerprint applications [21,22] and possibly reveal the user identity [23], to proactively identify user behavior for resource optimization [24] and to perform classification of Downlink Control Information (DCI) messages [25]. While previous studies on mobile data traffic were only able to differentiate traffic profiles according to time usage patterns [26,27], the use of BS traffic traces allows more sophisticated studies, for example to identify application classes [19] and standalone applications [21,22].…”