Network Traffic Generation Focused on Flash Crowd Anomaly

Please enable javascript in your browser.

Saleh, Zahra | 2016

573 Viewed

Type of Document: M.Sc. Thesis
Language: English
Document No: 49708 (52)
University: Sharif University of Technology, International Campus, Kish Island
Department: Science and Engineering
Advisor(s): Jahangir, Amir Hossein
Abstract:
Flash Crowd traffic generation can be used as a metrics for measuring the resiliency and performance of a server. Also, it can provide a framework for verification and test of Intrusion detection systems (IDS) and Intrusion protection systems (IPS). Common traffic generation methods mimic timing and content of input traffic or regenerate input traffic by extracting its statistic distribution. So all of them need input traffic, while properties of Flash Crowd are different in the various servers and situations and there is no guaranty in existence of such samples of traffic for all servers. In this thesis, we introduce and use a new method for traffic generation without the need for input traffic. Therefore, we concentrate on properties of legal user and combine user behavior modeling with Flash Crowd effects. For this purpose, we implement add-ons for Firefox browser in order to capture user behavior. Afterwards, we fit the statistical behavior of users to a well-known distribution and model it by an approximate state machine. Finally, our traffic is generated online by the combination of this behavior and Flash Crowd effects
Keywords:
Flash Crowd ; Realistic Traffic Generation ; Behavioral Analysis ; Intrusion Detection System

No TOC