Project Members  

Man-Kit Leung     : jleung     at berkeley.edu
Bor-Yiing Su         : briansu   at bekeley.edu
Karen Hsu            : karenhsu at berkeley.edu

Security Analysis on Defenses against Sybil Attacks in
Wireless Sensor Networks

Project Report

Project Proposal

Motivation

        There is a need for practical security measures in wireless sensor networks, as ad hoc networks are inherently vulnerable to a plethora of attacks.  Moreover, the limited resources of wireless sensor networks present additional challenges to the classes of defenses that are practical in real deployments.
        The Sybil attack [1, 2] is an attack where a node can illegitimately claim multiple identities, either fabricated or stolen.  There are several variations of Sybil attacks.  It is also the entry point for other attacks to build upon (for example, one can easily use false identities for stuffing ballots in voting, directing traffic in a blackhole attack, etc.).
        Several defenses have been proposed against the Sybil attack, i.e. radio resource testing, random key pre-distribution [3], fingerprinting, etc.  Each has their own tradeoffs, and we are interested in studying these defenses using in different metrics (i.e. dynamic topologies, scaling compromised nodes/insider attacker, points of failure) and analyzing their performance.  Once we have a solid understanding of the available defenses we will attempt to propose an augmented/improved defense to the attack.

Problem Statement

1.      Model the various Sybil defenses in the Ptolemy II simulation environment.

2.      Simulate various Sybil attacks under various conditions.

3.      Perform analysis, comparing the different defense models on our proposed metrics.

4.      Provide insights on an improved defense.

Key Challenges

1.      Setting up Ptolemy II to simulate wireless sensor networks.

2.      Come up with a set of comprehensive metrics used to measure these defenses.

3.      Dealing with the limited resources of wireless sensor networks.

Plan & Schedule

1.      Study the Sybil attack and the existing defenses.
      --------------------------------------------------------------------------------------------------------------------10/29/07

2.      Implement modules that can simulate motes, base stations, attacker nodes, and defense schemes in Ptolemy II.
      --------------------------------------------------------------------------------------------------------------------11/19/07

3.      Do simulations on defenses with different metrics, and analyze their performances.
      --------------------------------------------------------------------------------------------------------------------11/25/07

4.      Try to improve the existing defense against the Sybil attack.
      --------------------------------------------------------------------------------------------------------------------12/03/07

5.      Write a formal report of the project.  
      --------------------------------------------------------------------------------------------------------------------12/10/07

References:

[1] Tanya Roosta, S. P. Shieh, Shankar Sastry. "Taxonomy of Security Attacks in Sensor Networks and". The First IEEE International Conference on System
      Integration and Reliability Improvements, December, 2006.

[2] Chris Karlof and David Wagner. Ad Hoc Networks, vol 1, issues 2--3 (Special Issue on Sensor Network Applications and Protocols), pp. 293-315, Elsevier,
      September 2003.

[4] Haowen Chan and Adrian Perrig. "PIKE: Peer Intermediaries for Key Establishment in Sensor Networks."
      In Proceedings of the 24th Conference of the IEEE Communications Society (Infocom) 2005.

Last Update: 12/17/2007