Traffic Classification Analysis Using OMNeT

There has been a lot of research on effective monitoring and management of the network traffic, where a large amount of internet traffic requires more accurate and efficient ways of traffic classification methods and approaches with an aim to improve network performance. In our research, we introduce the subject of packet classification in IP traffic analysis with a simple technique that relies on prototype classifier using OMNET++ (Optical Modelling Network using C++ programming language) which unfolds one new possibility for an online classification focusing on application detection in the absence of payload information. In this research, we evaluated our novel IATP (Inter-arrival time and precision) clustering algorithm with the help of OMNET++ scheduler for classification of network traffic. The analysis is based on the measure combined with inter-arrival time and precision which was able to distinguish fairly as a small different subset of clusters. With our implementation of a range of flow attributes, the simulation result demonstrates the effectiveness of 100% accuracy of classifying packets but does not constitute the same level of accuracy with real-time traffic classifier which operates under certain constraints. Accuracy for real-time traffic might normally varies from 80 to 95% and depends on the type of each application. Further study and heuristics are required for detecting much better methodologies for detecting applications with real-time traffic measurements.