Computer Networks 134 (2018) 66–77 Contents lists available at ScienceDirect Computer Networks journal homepage: www.elsevier.com/locate/comnet Towards minimum-delay and energy-efficient flooding in low-duty-cycle wireless sensor networks Long Cheng a,∗ , Jianwei Niu b , Chengwen Luo c , Lei Shu a,d , Linghe Kong e , Zhiwei Zhao f , Yu Gu g a College of Engineering, Nanjing Agricultural University, China b Beihang University, Beijing, China c Shenzhen University, China d School of Engineering, University of Lincoln, UK e Shanghai Jiao Tong University, China f University of Electric Science and Technology of China, China g Watson Health Cloud, IBM Watson Health, USA a r t i c l e i n f o Article history: Received 15 June 2017 Revised 30 November 2017 Accepted 15 January 2018 Available online 3 February 2018 Keywords: Wireless sensor networks Low-duty-cycle Flooding Minimum-delay a b s t r a c t Wireless sensor networks (WSNs) play a very important role in realizing Internet of Things (IoT). In many WSN applications, flooding is a fundamental network service for remote network configuration, diagnosis or disseminating code updates. Despite a plethora of research on flooding problem in the literature, there has been very limited research on flooding tree construction in asynchronous low-duty-cycle WSNs. In this paper, we focus our investigation on minimum-delay and energy-efficient flooding tree construction considering the duty-cycle operation and unreliable wireless links. We show the existence of the latency- energy trade-off in flooding. We formulate the problem as a undetermined-delay-constrained minimum spanning tree (UDC-MST) problem, where the delay constraint is known a posteriori. Due to the NP- completeness of the UDC-MST problem, we design a distributed Minimum-Delay Energy-efficient flooding Tree (MDET) algorithm to construct an energy optimal tree with flooding delay bounding. Through exten- sive simulations, we demonstrate that MDET achieves a comparable delivery latency with the minimum- delay flooding, and incurs only 10% more transmission cost than the lower bound, which yields a good balance between flooding delay and energy efficiency. © 2018 Elsevier B.V. All rights reserved. 1. Introduction Wireless sensor networks (WSNs) are important elements for realizing the Internet of Things (IoT), which are composed of tiny wireless sensing devices equipped with data processing and com- munication capabilities [1]. WSNs offer several advantages over traditional wired industrial monitoring and control systems in- cluding extended network coverage, easy and fast installation, re- silience against single node failure and cost effective maintenance. On the contrary, traditional wired sensing and automation systems normally require expensive communication cables to be installed and regularly maintained [2]. In many WSN applications, e.g., fac- tory automation, industrial process monitoring and control, and plant monitoring, flooding is a fundamental network service for re- mote network configuration, diagnosis or disseminating code up- ∗ Corresponding author. E-mail address: chengl@vt.edu (L. Cheng). dates. The development of effective flooding protocol is hence a key research topic in this area. During flooding (or network wide broadcasting), messages from a root node are disseminated to the whole network via multi-hop communication. Since sensor nodes are usually energy constrained for WSN sustainable monitoring and surveillance applications, they normally operate at a very-low- duty-cycle (e.g., 1% or less) to ensure the service continuity [3]. Existing flooding protocols [4] utilize the broadcast nature of radio transmission to improve the delivery ratio and reduce trans- mission redundancy, i.e., a single transmission can be heard by multiple neighbors within the sender’s radio range. However, in an asynchronous low-duty-cycle WSN, neighboring nodes do not al- ways wake up at the same time. Flooding is essentially achieved through a number of unicasts [3,5], and thus more transmissions are required to ensure the flooding coverage than conventional wireless networks. On the other hand, sensor nodes are subject to radio frequency interference. For example in harsh industrial environments, highly caustic or corrosive environments, high humidity levels, vibrations, https://doi.org/10.1016/j.comnet.2018.01.012 1389-1286/© 2018 Elsevier B.V. All rights reserved.