Negin Fathollah Nejad Asl, Computer Science and Engineering

On the Design and Analysis of Consensus Protocols for Vehicular Ad Hoc Networks
Vehicle-to-vehicle communication technologies support diverse cooperative applications for intelligent transportation systems to increase safety and fuel efficiency of road vehicles. Vehicles participating in a cooperative pplication are expected to make coordinated and mutually consistent decisions. To ensure consistency, it is often essential that the participating vehicles reach agreement on the data they use as a basis for these decisions. This thesis deals with the fundamental problem of reaching agreement on a value, or a set of values, in a distributed system in the presence of unrestricted communication failures. It is known from the literature that this problem is impossible to solve perfectly, i.e., no matter what algorithm we use there is always a non-zero probability of disagreement. Hence, our aim is to design algorithms that minimize the probability of disagreement. We propose and analyse several agreement algorithms to solve three fundamental consensus problems. These algorithms are distinguished by their decision criterion, which determine whether a computer should decide on a value or decide to abort. Our analyses show that the probability of disagreement depends strongly on the number of computers in the system, the number of rounds of message exchange, the choice of decision criterion, as well as the probability of message loss. We identify two types of disagreement, safe and unsafe disagreement, and show that unsafe disagreement can be avoided if all computers know the number of computers in the system.
​Negin Fathollah Nejad Asl belongs to the Computer Engineering division of Computer Science and Engineering.

Opponent: Mohamed Kaaniche, Head of the Dependable Computing and Fault Tolerance research group at LAAS-CNRS, France.

http://publications.lib.chalmers.se/publication/251334
Category Thesis defence
Location: EB, lecture hall, EDIT trappa C, D och H, EDIT
Starts: 20 September, 2017, 13:15
Ends: 20 September, 2017, 14:15

Published: Tue 05 Sep 2017.