P2P optimization: evaluation of different trust and reputation metrics

From NET Wiki
Revision as of 14:45, 17 September 2009 by Ftegele1 (talk | contribs)
Jump to navigation Jump to search

Details

Supervisor: Florian Tegeler
Duration: 3 to 6 months, we can adapt the tasks to
Type: Bachelor/Master Thesis or Student Project
Status: open


Trust and Reputation systems are proposed to optimize P2P systems by identifying and punishing malicious user behavior (e.g. freeriding, content pollution etc.). After a while, cooperation between well behaving users increases and misbehaving nodes are slowly excluded from the system. A variety of systems has been proposed by the research community, primarily differing in the ways how good or bad transactions are judged and weighted and how information from others is included in the local decision process of a single node. The student's task will be to analyze different metrics/algorithms and work out the deciding differences. Additionally the performance and stability characteristics of such systems should be investigated using simulations.

Required skills:

  • Basic understanding of P2P systems
  • Ability to read and understand research papers (assistance is provided)
  • Programming skills to create relevant simulations (any programming language (pref. Java or C/C++)

The thesis topic is very open and can be adapted together with interested students. In this project, multiple thesises can be done, e.g. a theoretical analysis to find optimal rating schemes, simulation work, implementations, analysis of optimal storage location of trust and much more. Please feel free to drop in at any time to my office of contact me via email.