A fault-tolerant and distributed capacitated connected dominating set algorithm for wireless sensor networks

Highlights

• A capacitated connected dominating set (CapCDS) algorithm for wireless sensor networks is proposed.

• The proposed algorithm is the first distributed and self-stabilizing CapCDS algorithm.

• Theoretical analysis for proof of correctness and move count is proven.

• The proposed algorithm is compared with the previous work through extensive testbed experiments and simulations.

Abstract

Energy efficiency is one of the major issues in wireless sensor networks (WSNs) that lack a fixed infrastructure and centralized control. In order to prolong the network lifetime, a connected dominating set (CDS) has been widely used as a virtual backbone in WSNs. The sensor nodes in WSNs are prone to failure due to a lack of battery, hardware damage, link failure, or environmental interference. Therefore, designing an energy-efficient and fault-tolerant CDS algorithm is quite vital in WSNs. A non-masking fault tolerance method denoted self-stabilizing tolerates any finite number of transient faults. In this paper, we propose a fault-tolerant distributed algorithm for a minimal capacitated CDS (CapCDS) construction in WSNs. To the best of our knowledge, this is the first distributed self-stabilizing CapCDS algorithm. It makes an illegitimate system legitimate at most $(\frac{n^2}{3}+2n)$ moves by using an unfair distributed scheduler where n is the number of nodes. The performance of the algorithm is validated through extensive experimental testbeds and simulations.

Introduction

Wireless sensor networks (WSNs) are the infrastructure-less wireless communication networks that are decentralized and composed of thousands of spatially distributed autonomous sensor nodes deployed in an area [1], [2], [3]. WSNs are popularly used in health care monitoring, vehicle tracking, environmental monitoring, precision agriculture, smart buildings, animal tracking, security, and surveillance [4]. Due to the fact that WSNs have no fixed physical backbone, the sensors which have limited power, memory, and computational capacities act as routers, mainly use a broadcast communication paradigm, and the data of the entire sensor network can be collected by a sink node.

Distributed systems such as WSNs can subject to failure by their nature. Fault tolerance is the ability to maintain the overall task of the network without any interruption due to sensor node failures. Self-stabilization concept is a significant approach supporting non-masking fault-tolerance [5]. Even though a distributed system is initially started from an illegitimate configuration, a legitimate configuration is supported by a self-stabilizing algorithm in a finite time (convergence). Without any external interventions, it stays so (closure). Self-stabilizing algorithms are generally given as a collection of rules of the form “if $<guard>$ then $<move>$” where guard is a predicate and move is a change of the local state. In general, the rules are run in an atomic way. That a true predicate enables the corresponding rule. That one rule is enabled gives rise to make it privileged. In order to make a move, the privileged nodes have a chance to be selected by a scheduler. In an atomic step, only one privileged node is able to be selected under a central scheduler. While all privileged nodes are selected under synchronous scheduler, an asynchronous scheduler selects any non-empty subset of them. Additionally, as indicated in [6], using an asynchronous scheduler is quite convenient and realistic for WSNs.

A WSN can be modeled as a unit disk graph (UDG) G(V,E) in a Euclidean plane in which V denotes the set of nodes and E denotes the set of edges. In order to cope with the scalability and energy efficiency, constructing a virtual backbone is vital in WSNs. A dominating set (DS) of G(V,E) is a subset D ⊂ V such as each node in $V-D$ is adjacent to at least one node in D. A connected dominating set (CDS) is a DS that induces a connected subgraph of G. A virtual backbone of a WSN can be formed by a CDS [7] where finding minimum CDS is NP-hard [8]. Each node in a CDS is called dominator where the nodes out of CDS are called dominatee. A capacitated CDS (CapCDS) is a CDS in which dominators have a capacity value which is the number of nodes it can dominate. Although a CDS has an important role as a virtual backbone in WSNs, improperly assigning dominatees causes that some dominators have to cover more dominatees than their capacity to service. Therefore, constructing a CapCDS is vital to prolonging the network lifetime in WSNs.

In this paper, we propose the first distributed self-stabilizing algorithm for CapCDS construction in WSNs. By using an unfair distributed scheduler, it stabilizes at most ($\frac{n^2}{3}+2n$) moves where n is the node count. The remaining of this paper is organized as follows. Section 2 represents the background which consists of the system model and the formal definitions. We discuss the related work in Section 3. The proposed algorithm is explained in Section 4. Section 5 is devoted to theoretical analysis. The performance evaluation that includes the results of experimental testbeds and simulations are given in Section 6. Finally, conclusions are presented in Section 7.