날짜: 3월 17일 화요일
장소: 301동 412호
시간: PM 1:00 ~ 2:00 (1시간)
Seminar Topic: Quality of security paradigm for end-to-end security in integrated future wireless networks
Presenter: 정종민 박사님 (성균관대학교 연구교수)
Education
• Kangwon National University, Korea, Ph.D., 2000 – 2004
• Kangwon National University, Korea, M.S., 1998 – 2000 –
• Kangwon National University, Korea, B.A., 1994 – 1998
Work experience
• Research Professor in Sungkyunkwan University, Korea, Sep 2008 – present
• Visiting scientist & post-doc in Wireless Network Lab of Cornell University, USA, Jan 2006 – Aug 2008
• Research professor Kyungpook National University, Korea Mar 2005 – Dec 2005
• Invited faculty in Information and Communication University (ICU), Korea, Mar 2004 – Feb 2005
• Internship in Korea Electronics Technology Institute (KETI) Korea, Aug 1999 – Jan 2000
Abstract:
In this talk, I will mainly present about one of my previous research issue, time information based key distribution in wireless sensor networks. Since sensor networks are typically closed communication environments (e.g., the set of network nodes is constant), one would expect that securing sensor networks is an easier task as opposed to open communication environments, such as ad hoc networks, for example. However, because of the massive number of nodes in a sensor network and the fact that compromising a single node may bring down the whole network, securing a sensor network is not a trivial task. In this paper, we propose and analyze a new time-based pre-deployed secure key distribution (TPSKD) scheme. To secure the operation of the key distribution, TPSKD implants individual security material in each node before the network deployment. The individual security materials serve as initial time values at the nodes and are simultaneously incremented in all the network nodes. For the scheme to operate properly, the frequency of the clocks at the network nodes need only to be approximately equal. When a communication between two nodes in the network is required, a simple protocol is executed which allows the two nodes to generate a common secret session-key. To prove the security of the proposed scheme, we used the concept of reduction to a secure pseudorandom function. The strength of the security scheme was demonstrated for a set of possible attack scenarios, showing that the computational complexity of the adversarial attack equals that of the brute-force attack. We also analyzed the effectiveness and the efficiency of the scheme, where the former is measured as the fraction of node pairs that can established a common session key, while the latter is evaluated by resource requirements, such as memory, processing, and communication overhead.