| KEYWORD |
A Predictive Peer-to-Peer Network
Thesis abroad
keywords AUTONOMIC NETWORKS, PEER-TO-PEER
Reference persons PAOLO GIACCONE
External reference persons Prof.Antonio Liotta (Technische Universiteit Eindhoven, The Netherlands)
Research Groups Telecommunication Networks Group
Thesis type EXPERIMENTAL, SIMULATIVE
Description Peer-to-peer (P2P) networks are currently being applied on a range of applications such as file sharing, IPTV and video conferencing. A crucial element of the P2P system is the algorithm that keeps the nodes connected. Existing algorithms tend to be application-specific ' they are optimized for a particular application ' and use the network resources inefficiently.
This project explores autonomic algorithms for the development of a novel approach to P2P networking. The aim is to come up with prediction models, which correlate the status of the physical network to the quality that is actually perceived by the end user (quality of experience). The hypothesis is that a 'predictive' P2P overlay can improve the network utilization.
In order to assess this hypothesis, the student will design the protocol internals, implement it within a network simulator, and carry out a comparative analysis.
Required skills The student is expected to have a strong background in networking, excellent programming skills in C/C++, and a good knowledge of Unix/Linux.
Notes This study will be executed at Technische Universiteit Eindhoven, The Netherlands. The student will be responsible for the travel and life expenses. Travel can be possibly funded through ERASMUS mobility grants for which the student has to apply in advance.
Deadline: December 2010
Deadline 25/10/2010
PROPONI LA TUA CANDIDATURA