WND-Identifier: Automated and Efficient Identification of Wireless Network Devices

The widespread application of wireless communication technology brings great convenience to people, but security and privacy problems also arise. To assess and guarantee the security of wireless networks and user devices, discovering and identifying wireless devices become a foremost task. Currently...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: Fangzhou Zhu, Liang Liu, Simin Hu, Ting Lv, Renjun Ye
Formato: article
Lenguaje:EN
Publicado: Hindawi-Wiley 2021
Materias:
Acceso en línea:https://doaj.org/article/8e8fca6fe28d45a1a60cd8357f46090c
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
Descripción
Sumario:The widespread application of wireless communication technology brings great convenience to people, but security and privacy problems also arise. To assess and guarantee the security of wireless networks and user devices, discovering and identifying wireless devices become a foremost task. Currently, effective device identification is still a challenging issue, as device fingerprinting requires huge training datasets and is difficult to expand, and rule-based identification is not accurate and reliable enough. In this paper, we propose WND-Identifier, a universal and extensible framework for the identification of wireless devices, which can generate high-precision device labels (vendor, type, and product model) efficiently without user interaction. We first introduce the concept of device-info-related network protocols. WND-Identifier makes full use of the natural language features in such protocol messages and combines with the device description in the welcome page, thereby utilizing extraction rules to generate concrete device labels. Considering that the device information in the protocol messages may be incomplete or forged, we further take advantage of the application logic independence and stability of the device-info-related protocol, so as to build a multiprotocol text classification model, which maps the device to a known label. We conduct experiments in homes and public networks and present three application scenarios to verify the effectiveness of WND-Identifier.