Design of Realistic Network Traffic Generator and Traffic Anomalies

Please enable javascript in your browser.

Azadmanesh, Maryam | 2011

434 Viewed

Type of Document: M.Sc. Thesis
Language: Farsi
Document No: 44251 (19)
University: Sharif University of Technology
Department: Computer Engineering
Advisor(s): Jahangir, Amir Hossein
Abstract:
Generating realistic network traffic serves capacity planning, traffic engineering, understanding internet service performance and anomaly detection. A useful traffic generator must generate characteristics in flows and packets, and add known internet anomalies to legitimate traffic. In this dissertation, we develop a tool for generating realistic and anomalous traffic. This generator can generate traffic in a scalable, reproducible and stochastic manner. Two methods are proposed for this purpose. In the first method, the traffic is generated according to statistical distribution of packets and flow charactristics. In the second method, traffic is generated by modeling the behavior of network agents; they affect statistical characteristics of traffic. This tool can generate scanning port and denial of services attacks. The results of own experiments show that this tool can reproduce packet burstiness and flow characteristics of realistic traffic. Moreover, this tool can be used successfully in anomaly detection experiment
Keywords:
Traffic Generator ; Traffic Generator ; Anomaly

No TOC