Algorithms and Applications of Leader Election in Unstructured Networks

博士 === 國立中興大學 === 電機工程學系所 === 103 === For self-organized networks with highly decentralized and self-organized natures, e.g., Wireless Sensor Networks (WSNs), neither the identity nor the number of processes is known to all participants initially, even at the end of the computation, because no centr...

Full description

Bibliographic Details
Main Authors: Che-Cheng Chang, 張哲誠
Other Authors: 蔡智強
Format: Others
Language:en_US
Published: 2015
Online Access:http://ndltd.ncl.edu.tw/handle/02228759338902763075
Description
Summary:博士 === 國立中興大學 === 電機工程學系所 === 103 === For self-organized networks with highly decentralized and self-organized natures, e.g., Wireless Sensor Networks (WSNs), neither the identity nor the number of processes is known to all participants initially, even at the end of the computation, because no central authority exists to support such context information. Hence, solving some existing distributed issues in such networks cannot be accomplished in the ways for traditional fixed networks. In the literature, the consensus problem has been studied in a new computational model corresponding to self-organized networks. Such a variant of the traditional consensus problem is called Consensus with Unknown Participants (CUP), which relaxes the requirement for the original knowledge owned by every process about participants in the computation. Particularly, the aforementioned work considers asynchronous networks without process crashes, and has determined the necessary and sufficient knowledge connectivity condition for solving the problem, which is called One Sink Reducibility (OSR). In this dissertation, we investigate the possibility of finding a less restrictive knowledge connectivity condition than OSR, based on which consensus can still be achieved in self-organized networks. Later, the CUP problem was extended to consider process crashes. This new problem is called Fault-Tolerant Consensus with Unknown Participants (FT-CUP). One sufficient knowledge connectivity condition based on which the FT-CUP problem can be solved with a minimal synchrony assumption has been proposed in the literature. In this dissertation, we continue to investigate such a problem. In particular, we provide a less constrained condition, which is still sufficient for solving the FT-CUP problem. On the other hand, leader election is a fundamental building block for many applications. Particularly, there have been many methods proposed for solving the leader election problem in the literature. Solving the leader election problem in static networks is easier than in dynamic networks because dynamic behavior of processes must be considered in the latter. A simple way to solve this problem in dynamic networks is attaching a synchronous clock to each process. But doing so violates the assumption of asynchrony. Moreover, a leader had better be a process with the best performance-related characteristic among all nodes within a connected component. In this dissertation, we present an efficient leader election algorithm with regard to performance-related characteristics for dynamic networks, without any synchronous clock assumption. Furthermore, based on our leader election algorithms, we devise a novel collaborative mechanism for a visual surveillance system. Such a mechanism helps a visual surveillance system with more than one monitoring camera assign the proper number of cameras to track a particular object automatically and efficiently. Hence, upon a moderate number of objects entering the monitoring area, the monitoring cameras will be appointed to each object almost evenly. At last, the Distributed Trigger Counting (DTC) problem is to raise an alert while the number of triggers received by the system reaches a pre-specified amount. Particularly, there have been several algorithms proposed to solve the DTC problem in the literature. However, these existing algorithms are all under the assumption that each node knows what kind of topology the whole system is as well as what kind of role it plays in the system. Obviously, such an assumption is not practical. In this dissertation, we propose a novel distributed algorithm to solve the DTC problem without any global assumption. Moreover, we further reduce the message complexity of the foregoing algorithm, and then propose a more message-efficient version, which is still not based on any global assumption. More importantly, we also address the issue of solving the DTC problem on a dynamic network, of which the topology will continually change due to process moving, leaving and joining.