|
Descripción
|
|
|---|---|
| The specific nature of routing in sensor networks has made possible new sorts of attacks that can have closer insight and effect on the networks packets, the most important being the packet tampering. Routing attacks on the network level are the first step in tampering with the packets. In this work we propose a solution for detecting and eliminating these attacks that couples reputation systems with clustering techniques, namely unsupervised genetic algorithm and self-organizing maps, trained for detecting outliers in data. The algorithms use the feature space based on sequences of routing hops that provides ability of detecting wide range of attacks. We further present a flexible way of integrating the solution into targeted sensor network that can easily adapt its computational requirements to the existing network resources. The solution offers many benefits: scalable solution, fast response to adversarial activities, ability to detect unknown attacks, high adaptability, flexible integration, and high ability in detecting and confining attacks. | |
|
Internacional
|
Si |
|
Nombre congreso
|
ACM Conference on Computer and Communications Security, CCS 2010 |
|
Tipo de participación
|
960 |
|
Lugar del congreso
|
Chicago, Illinois, USA |
|
Revisores
|
Si |
|
ISBN o ISSN
|
978-1-4503-0088-9 |
|
DOI
|
10.1145/1866423.1866426 |
|
Fecha inicio congreso
|
04/10/2010 |
|
Fecha fin congreso
|
08/10/2010 |
|
Desde la página
|
8 |
|
Hasta la página
|
13 |
|
Título de las actas
|
CCS'10 Proceedings of the 3rd ACM workshop on Artificial intelligence and security |