An online-based approach for detecting connectivity anomaly in 802.11 networks

International Conference and Business Expo on Wireless Communication & Network

September 21-23, 2015 Baltimore, USA

Dossa M Massa

The Institute of Finance Management (IFM), Tanzania

Posters-Accepted Abstracts: Sensor Netw Data Commun

Abstract :

Wireless 802.11 networks are a popular technology that offers inexpensive ubiquitous access to the Internet in campuses, enterprises, homes, coffee shops, airports, and other public places. Their wide-scale adoption has brought great convenience to many people, giving them anytime and anywhere access to the Internet. As a result, people are becoming more and more dependent on these networks and are increasingly demanding reliability and high performance when connecting to the Internet through them. The task of diagnosing and fixing connectivity problems using online analysis of 802.11 network usage data is one of the key challenges that campus and corporate 802.11 network administrator�s confront. This paper proposes an online method for detection of 802.11 Access Point (AP) abrupt ending of connections that happen when a large number of sessions in the same Access Point (AP) end within the same second window. Complex event processing techniques are employed to detect abrupt ending events and their associated anomaly-related patterns in real-time based on the collected 802.11 network usage data. Patterns such as AP interferences, across AP vicinity interferences, AP persistent interferences, AP overload, and AP crash are adequately detected and characterized.

Biography :

Email: mmohamed@inesctec.pt