Research‎ > ‎PlanetLab experiments‎ > ‎

Reputation Framework for Behavioural History

Background information about this research can be obtained from the IFIPTM 2010 conference publication and my DPhil thesis.

The aim of the PlanetLab experiment is to deploy a set of servers and clients that interact with each other to emulate clients and servers on the real Internet. We would then want to test two hypotheses:
  1. That the increased overhead of checking of a client's reputation, and the monitoring and reporting of behaviour would get lost in the noise of typical Internet applications.
  2. That there is no significant difference in the performance of a centralised versus a distributed reputation management system.
These are the null hypotheses, and whilst we are happy for 1 to be found true, we would hope 2 to be false, but we would like to know by how much. The client and server requests would be synthetic derived from various email logs, very similar to those described in the thesis.

[description by Ian Wakeman]
Comments