6th Annual Judith Ramaley Celebration of Research and Creative Scholarship
|
Poster #122 Vulnerability of Wireless Network Security due to Parallelized Brute Force Attacks Matthew Wolfe Faculty Mentors: Gerald Cichanowski and Mingrui Zhang Wireless networks are becoming one of the most prominent methods for businesses and consumers to share digital information. While wireless networks are convenient, they also create a potential vulnerability for companies who choose to deploy them. The current security encryption schemes for wireless networks are susceptible to dictionary attacks as well as traditional attacks that compromise the 802.11 standard. This paper will discuss the possibly of a malicious user’s ability to gain unauthorized access to a wireless network using highly parallelized brute force dictionary attacks designed to run on a Graphics Processing Unit (GPU). In order to determine whether a GPU optimized attack would increase probability of gaining unauthorized access, the attacks will run under three different configurations: serial code run on a single processor, multithreaded code on a quad-core processor, and GPU optimized code. In order to provide quantitative data for analysis, two different benchmarks were used. These benchmarks produced results that showed the GPU optimized code, while producing more overhead, would ultimately reveal a wireless network key faster than single and multiple core optimized code. |
|