EWIDS: an extended wireless IDS for metropolitan wireless networks based on kinematical analysis

• Autor(a) principal: Pirmez, Luci
• Data de Publicação: 2011
• Outros Autores: Vianna, Nilson Rocha, Correia, Reinaldo de Barros, Carmo, Luiz Fernando Rust da Costa, Farias, Cláudio Miceli de, Salmon, Hélio Mendes
• Tipo de documento: Artigo
• Idioma: por
• Título da fonte: Repositório Institucional do INMETRO
• Texto Completo: http://hdl.handle.net/10926/1592

Wireless metropolitan area networks (WMANs) are well known to subject users or applications and to a vast gamma of security risks, hindering security critical distributed applications from employing this type of network as a communication infrastructure. Most existing approaches for addressing WMAN security issues use cryptography-based mechanisms or ad-hoc adapted versions of traditional Intrusion Detection Systems (IDS) for wired networks. While the first approach may lead to unfeasible computation costs for mobile hand-held devices, the second exhibits a high dependency on the freshness of their attack-signature databases, besides not considering any inherent characteristic of wireless networks, such as mobility. Thus, we present EWIDS (Extended Wireless IDS); a lightweight IDS specially designed for WMANs, which detects anomalous wireless device transmissions by employing kinematical analysis on the motion of users’ mobile devices. EWIDS also takes into account the decision information generated by transmitter fingerprint mechanisms used to identify wireless device. Both information is integrated through a fuzzy logic engine in order to increase the system performance. Realistic simulations based on WMAN scenarios revealed that our approach is very promising, since worst-case results have shown high correct alarm rates associated with low false positive rates.