Self-stabilization in dynamic distributed systems. Doctorat thesis
No Thumbnail Available
Date
2022-12-07
Authors
Saadi Leila
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Dynamic distributed systems are used widely in everyday life by dealing with exploitation and treatment on information, documents, services, medias, and all entertainment means. Many companies creating software and systems compete to offer users powerful and tolerant services. This dynamic DS are obliged to response the user demands in time and with trustworthy information, they offer all levels of security; authenticity of all services and personnel information. This kind of systems appears in our life in all areas using different means of sensors that collect all type of information and requests where the challenge is to ensure a permanent service availability since failures occurs. Many algorithms are used here to fault tolerant and exceed those situations, among the self-stabilizing algorithms that take the system to a legitimate state even the presence of failures and errors. This thesis deals with the problem of finding dominating set using self-stabilizing paradigm in dynamic distributed systems. Usually, members of a dominating set are selected to be as cluster heads in Wireless Sensor Networks (WSN) in order to ensure a permanent service availability. Since failures occurs frequently inside WSNdue to limited battery energy, selfstabilizing algorithm allows recomputing the dominating set, and hence the network returns to its ordinary running. Existing works have introduced many variants of self-stabilizing algorithms that compute minimal dominating set S where each node out of S has neighbours in S more than it has out S. In this thesis, we introduce a generalized self-stabilizing algorithm called minimal (α, β)-dominating set. An α-dominating set is a subset of nodes S such that for any node υ out of S, the rate of neighbours of υ inside S must be greater than α, where 0 < α ≤ 1. In the same way, an (α, β)-dominating set is a subset of nodes S such that: S is α-dominating set and for each node υ in S, the rate of neighbours of υ inside S is greater than β, where 0 ≤ β ≤ 1. Mathematical proofs and simulation tests showthe correctness and the efficiency of the proposed algorithm. Through our proposed variant (α, β)-domination, we prove rigorously the conjecture of Carrier et. al. (Self-stabilizing (f, g)-alliances with safe convergence) who have proposed a self-stabilizing algorithm for a domination variant called (f, g)-alliance set only when f ≥ g. We prove the correctness of the case f < g.