Broadcast Based Code Dissemination Scheme for Duty Cycle Based Wireless Sensor Networks

Ting Shu, Wei Liu, Tian Wang, Qingyong Deng, Ming Zhao, Neal N. Xiong, Xiong Li, Anfeng Liu

Abstract

In duty-cycled Wireless Sensor Networks (WSNs), nodes alternate between active and sleep states to conserve energy, presenting a significant challenge for remote software updates. Traditional flooding mechanisms often result in "broadcast storms" or failed deliveries due to misaligned wake-up schedules. This paper presents PDLB (Proportion to Duty Cycle Length-Based broadcast), a novel dissemination scheme that optimizes energy efficiency and ensures 100% code coverage. PDLB integrates three core mechanisms: Number of Selected Slots for Broadcast (NSSB), Improved Counter-based Priority Broadcasting (ICPB), and Timeout Active Request Code (TARC). Simulation results demonstrate that PDLB achieves superior performance in terms of transmission delay and energy consumption compared to existing flooding methods, particularly when the number of broadcast slots is optimized to half the duty cycle length.

Introduction

The proliferation of the Internet of Things (IoT) has led to the widespread deployment of Wireless Sensor Networks (WSNs) for various monitoring and control applications. Modern WSNs are increasingly "software-defined," allowing for remote updates to modify functionality or patch security vulnerabilities. However, the primary constraint in these networks remains energy. To maximize operational lifetime, nodes typically employ duty cycling, where they remain in a low-power sleep state for the majority of the time, waking up only briefly to communicate.

Disseminating large code updates across a duty-cycled network is inherently difficult. If a source node broadcasts an update while its neighbors are asleep, the message is lost. Conversely, aggressive rebroadcasting leads to the "broadcast storm" problem, characterized by redundant packets, excessive collisions, and rapid battery depletion. Existing dissemination protocols often assume continuous node availability, making them inefficient for duty-cycled environments. This paper proposes the PDLB scheme to address these challenges by balancing proactive "push" broadcasts with reactive "pull" requests.

The PDLB Scheme

The PDLB scheme is designed to maximize coverage while minimizing the energy overhead associated with redundant transmissions. It consists of three integrated components.

1. Number of Selected Slots for Broadcast (NSSB)

In a duty-cycled network, a single broadcast is unlikely to reach all neighbors. NSSB addresses this by requiring a node to select $k$ specific slots within a duty cycle to rebroadcast the code. This increases the probability that neighbors with different wake-up offsets will be active during at least one transmission. The selection of $k$ is a critical parameter; the authors identify a "sweet spot" where $k$ is approximately half the total number of slots in a duty cycle, balancing delivery success against energy cost.

2. Improved Counter-based Priority Broadcasting (ICPB)

To mitigate the broadcast storm, PDLB utilizes a priority-aware suppression mechanism. When a node receives a new code update, it does not rebroadcast immediately. Instead, it waits for a Random Access Delay (RAD). The RAD is calculated based on two factors:

  • Distance from the Sink: Nodes further from the source are given higher priority (shorter delays) to facilitate rapid outward propagation.
  • Battery Level: Nodes with higher remaining energy are prioritized to perform broadcasts, effectively balancing the energy load across the network and extending the overall lifetime.

During the RAD, the node listens to the channel. If it hears the same code update from a threshold number of neighbors ($C_h$), it assumes the local area is sufficiently covered and cancels its own scheduled broadcast.

3. Timeout Active Request Code (TARC)

TARC serves as a safety net for nodes in sparse regions or those that missed updates due to collisions. If a node does not receive the update within a predefined timeout period, it switches from a passive listener to an active requester. The node identifies the wake-up schedules of its neighbors and sends a targeted request for the code. This "pull" mechanism ensures 100% network coverage without requiring the entire network to continue broadcasting indefinitely.

Experimental Methodology

The PDLB scheme was evaluated using extensive C++ simulations. The experimental setup modeled a circular network with a 100m radius. To test the scalability and robustness of the scheme, three node densities were simulated:

  • Sparse: 1,000 nodes
  • Normal: 2,500 nodes
  • Dense: 4,000 nodes

The energy model accounted for the costs of broadcasting ($E_s$), receiving ($E_r$), and active listening ($E_a$). Performance was measured against traditional flooding and existing duty-cycle-aware protocols, focusing on coverage, transmission delay, and total energy consumption.

Results and Analysis

The simulation results highlight several key advantages of the PDLB approach.

Coverage and Reliability

PDLB consistently achieved 100% code coverage across all tested densities. In contrast, traditional flooding methods adapted for duty cycles often failed to reach more than 40% of the nodes when restricted to a single broadcast slot. The combination of NSSB's multiple broadcast slots and TARC's request mechanism proved essential for reaching isolated or poorly connected nodes.

Energy Efficiency and the "Half-Cycle" Rule

A significant finding of the study is the relationship between the number of broadcast slots ($k$) and energy efficiency. Energy consumption is minimized when $k$ is approximately half the duty cycle length. This configuration provides enough "push" opportunities to cover most nodes quickly while avoiding the diminishing returns and high energy costs of broadcasting in every slot.

Impact of Network Density and Duty Cycle

As network density increased from 1,000 to 4,000 nodes, the transmission delay was reduced by 0.83% to 65.87%. Higher density increases the likelihood of successful transmissions in fewer cycles. However, the "cost of sleeping" remains significant: for every additional slot in a duty cycle, the number of broadcasts required increased by 1.54% to 33.33%, and the transmission delay grew by 2.27% to 70.04%.

Conclusion

The PDLB scheme provides a robust and energy-efficient solution for code dissemination in duty-cycled WSNs. By intelligently selecting broadcast slots, prioritizing healthy nodes for transmission, and implementing a request-based recovery mechanism, PDLB ensures total network updates with minimal overhead.

Future research will focus on adapting PDLB for mobile topologies, where neighbor relationships change dynamically. Additionally, exploring heterogeneous duty cycles—where nodes have varying sleep schedules based on local sensing requirements—remains an important area for further optimization.

Summary updated

Abstract

The technology of software-defined devices has great prospect in updating the programs of the sensing devices to enable the devices to adapt to the new environments and applications. However, to effectively spread the code to each node in the software-defined-based wireless sensor networks (WSNs) is still a challenging task. In this paper, a proportion to duty cycle length-based broadcast (PDLB) scheme is proposed to disseminate code in duty cycle-based WSNs in a fast and energy-efficient style. In PDLB scheme, the sink first initiates a code broadcast. Then, the nodes that have received code start counting and calculate the priority of broadcasting based on an improved counter-based priority broadcasting approach proposed in this paper. Next, the nodes with the highest priority select k slots to broadcast according to the length of duty cycle so that the code can be spread from the sink to the network periphery. Finally, the nodes that have not obtained the code wait for a period of timeout to broadcast to actively send code request; after receiving code request, the node that obtained the code initiates a code broadcast to cause the code requester to obtain the code. The PDLB scheme has been validated by comprehensive experiments that show the following regularities. The smaller the count threshold value Ch, the larger the number of slots selected for broadcasting that can guarantee a certain coverage ratio, while the code dissemination speed slows down at the same time. The longer the duty cycle is, the more the broadcast time is required, and the larger the transmission delay is. For each slot added in a working cycle, the number of broadcasts is increased by 1.54%-33.33% and the transmission delay is augmented by 2.27%-70.04%. The higher the node density is, the shorter the code spread speed is. As the network density increases, the transmission delay is reduced by 0.83%-65.87%. When the number of broadcasting slots is half of the total number of slots in a work cycle, the energy consumption reaches the lowest value.
BibTeX
@article{shu2019broadcast,
  author = {Ting Shu and Wei Liu and Tian Wang and Qingyong Deng and Ming Zhao and Neal N. Xiong and Xiong Li and Anfeng Liu},
  title = {{Broadcast Based Code Dissemination Scheme for Duty Cycle Based Wireless Sensor Networks}},
  journal = {IEEE Access},
  year = {2019},
  doi = {10.1109/access.2019.2932006},
}
Figure