QCM2R: A QoS-aware cross-layered multichannel multisink routing protocol for stream based wireless sensor networks

Abstract

Unlike the scalar data (such as temperature, pressure and humidity), the vector data (such as image, audio and video) necessitates more stringent Quality of Service (QoS) requirements in terms of bandwidth, delay, reliability and information security. These QoS requirements can be hardly achieved in a proper fashion by using a single channel for wireless communication. However, multichannel methodology may assist in accomplishing these QoS requirements by making possible parallel communication, enhancing throughput/delivery ratio, reducing transmission delay and countering jamming attacks. Furthermore, enabling data gathering at multiple points (i.e. multisink approach) may improve QoS by handling congestion, avoiding single point of failure issue and making possible load balancing between the available routes towards the corresponding destinations. To achieve reliable communication in stream based multichannel Wireless Sensor Networks (WSNs), this work proposes a novel QoS-aware Cross-layered Multichannel Multisink Routing protocol (QCM2R) for WSNs. For substantiating the performance of QCM2R protocol, the simulations are performed in NS-2 demonstrating the performance superiority of the proposed QCM2R protocol against the counterpart in terms of network lifetime, reliability, delay and throughput.

Introduction

Wireless sensor networks are the collection of small battery-enabled inexpensive devices that are used for monitoring the surrounding environment. They may service a variety of emerging smart applications (Rehan et al., 2018) and deal with natural calamities (Castillo-Effer et al., 2004), ecological observation (Szewczyk et al., 2004), structural monitoring (Xu et al., 2004), reconnaissance activities (Yick et al., 2005), surveillance (Ciuonzo and Rossi, 2018) and so on. At the inception of WSNs, sensor devices have basic hardware capability that includes a small memory, sensing/processing/transmission units and a small battery for keeping them alive (Akyildiz et al., 2002) (Rehan et al., 2016). Being mainly focused on maximizing their lifespan, they were reckoning with low data rate communication (e.g. temperature, pressure and humidity measurements of the terrain of interest). The technological advancements have enhanced the capability of sensor nodes and enabled them to deal with large volumes of data. Consequently, Wireless Multimedia Sensor Networks (WMSNs) (Yigitel et al., 2011) (Akyildiz et al., 2007) have evolved. The multimedia sensor nodes are equipped with cameras and microphones and are capable of handling vector data pertaining to image, audio and video surveillance of the target region.

Upon sensing an event in the area of interest, a sensor node forwards it towards data gathering point (aka sink node) in a hop-by-hop manner. When the information is received at sink node, then the network information system may analyze the corresponding event for appropriate decision making. Normally, for transmitting the scalar data, there does not require any strict QoS requirements than for transmitting the vector data. It is due to the fact that vector data is delay sensitive and requires high bandwidth and reliability for meaningful information delivery. Such stringent QoS requirements are very challenging to meet using the conventional strategies such as single-channel and single-sink/path based routing methodologies.

A vast majority of conventional WSNs use single channel for routing the scalar and vector data across the network (e.g. AntNet (Di Caro and Dorigo, 1998) and TPGF (Shu et al., 2010) etc). Since single channel strategy may allocate the same channel to neighboring nodes, therefore it may not afford parallel transmission. Consequently, the nodes in the interference range may wait for their turn to perform communication or contend for the single channel for data transmission, which may enhance network delay. Furthermore, the nodes transmitting simultaneously (on the single channel) in the interference range may contribute to network interference and inevitably network reliability is adversely affected. Besides that, the network operation may even be halted when the single operational frequency suffers from jamming. With the advancements in technology, the researchers have realized that multichannel methodology may provide a reasonable solution to the above mentioned issues and therefore it is better than single channel approach (Rehan et al., 2017). Moreover, multichannel methodology may effectively enhance the performance of WSNs by ensuring parallel communication, high throughput, less delay and fresh data delivery (Rehan et al., 2017) (Soua and Minet, 2015). The multichannel WSNs may suffer from some quality-related issues such as link stability problem (Soua and Minet, 2015), channel switching overhead, interference, broadcast support and so on (Rehan et al., 2017) (Soua and Minet, 2015), however such issues may be handled by appropriately designing the multichannel routing protocol.

Traditional WSNs have one data gathering point called sink (aka destination). When either a relevant node (e.g. on-path node) in the vicinity of a sink or sink itself suffers from any critical issue, then network performance may be degraded e.g. an increase in data volume may bottleneck the centralized sink node (Murugan and Pathan, 2016) because it may cause congestion around sink node. Another issue, may be single point of failure of a related node or even sink, occurring due to energy depletion or channel jamming issues. Such issues may adversely affect the performance of WSNs and unfavorably impact QoS oriented data delivery. A reasonable solution to the above mentioned issues is to employ multisink approach that may increase data gathering points. The multisink approach may avoid the congestion problem near sink and circumvent the single point of failure issue in WSNs (Erman et al., 2009). Furthermore, multisink approach performs load distribution across sensor network and achieves energy balancing around sink node (Eslaminejad and Razak, 2012) and therefore network lifetime is extended (El-Fouly et al., 2018). Since multisink methodology enhances network throughput (El-Fouly et al., 2018), henceforth it is suitable for high data rate applications such as stream based communication in WSNs. However, finding the cost effective end-to-end path between source and sink(s) is one of the main issues to be coped with in multisink WSNs (Eslaminejad and Razak, 2012).

It is clear from the above discussion that both multichannel and multisink methodologies may improve the performance of WSNs. Therefore, to get the benefits of both worlds, this work proposes a reactive QoS-aware Cross-layered Multichannel Multisink Routing protocol (QCM2R) that manifests clearly the following main properties:

• The proposed protocol establishes QoS-aware paths between source and sinks upon the occurrence of an event in the region of interest.

• The proposed protocol selects the best QoS-aware path at a particular epoch for end-to-end data delivery. Furthermore, it may dynamically shift the path for ensuring load balancing between the available routes during the communication session.

• The proposed protocol executes distributed channel assignment jointly with routing. The channel allocation may be performed in a manner to possibly perform two-hop channel orthogonality.

• The proposed protocol administers a channel refresh mechanism for reassigning wireless channels to the path(s) experiencing a bad channel event on one or more links.

The rest of this paper is organized as follows. The Section II discusses literature review and motivation. In Section III, proposed network model is outlined. Section IV elaborates the proposed solution in a detailed manner. In Section V, performance evaluation of the proposed work is delineated. In the end, future scope and directions are discussed in Section VI while a brief conclusion of the research project is outlined in Section VII. Furthermore for convenience, a list of relevant acronyms and symbols are described in Table 1, Table 2 respectively.