Overcoming Detection Rate Bottlenecks In New Qos Violation With Combining Hmm And Information Fusion Theory

Keywords

Dempster-Shafer evidence theory; HMM; Kaufman algorithm; Power spectrum density; QoS

Abstract

A novel detection scheme with combing HMM (Hidden Markov Models) and information fusion theory was proposed. First, the PSD (Power Spectral Density) features of both IPT (IP address-Port Triple) and WASPQ (Weighted Average Size of Packet in Queue) were calculated, which reflect the inherent periodicity of LDoS Attacks. Second, Two HMMs were constituted with these PSD features. Finally, the judgment result was determined by fusing the output results of HMMs based on Dempster-Shafer evidence theory. And it was applied to detect new QoS (Quality of Service) violation hidden in legitimate TCP/IP streams. In addition, Kaufman algorithm is used to dynamically adjust and upgrade threshold value. The experiments show that the detection scheme effectively reduces the false-positive rate and false-negative rate.

Publication Date

1-1-2015

Publication Title

Electronics, Communications and Networks IV - Proceedings of the 4th International Conference on Electronics, Communications and Networks, CECNet2014

Volume

2

Number of Pages

1759-1763

Document Type

Article; Proceedings Paper

Personal Identifier

scopus

DOI Link

https://doi.org/10.1201/b18592-320

Socpus ID

84960103573 (Scopus)

Source API URL

https://api.elsevier.com/content/abstract/scopus_id/84960103573

This document is currently not available here.

Share

COinS