DVSP: Dynamic Virtual Sensor Provisioning in Sensor–Cloud-Based Internet of Things

Virtual sensor provisioning is an essential process in sensor-cloud-based Internet of Things (IoT), and it is responsible for the efficient utilization of physical resources in the system. However, the existing schemes for virtual sensor provisioning do not provide an optimal solution while considering overall demand of multiple users/services. As a result, redundant sensor nodes are provisioned, which leads to increased energy consumption and reduced network lifetime. In this paper, we present a dynamic virtual sensor provisioning scheme for sensor-cloud-based IoT applications to maintain the energy efficiency of the deployed physical sensor nodes while maintaining the quality of service (QoS) of the service requests. We model the interaction between the cloud service provider and the sensor owners using the single-leader multifollower Stackelberg game. The players of the game exploit the spatial correlation among the on-field sensor nodes, and consequently, the oligopoly created between the players is dynamically updated. We show the existence of a Stackelberg-Nash-Cournot equilibrium in the game. We evaluated the performance of the proposed scheme through extensive simulations. The results depict improvement in the energy efficiency of the nodes as well as increase in the lifetime of the deployed on-fields sensors in the proposed scheme compared to benchmark schemes. We also plot the average number of QoS violations in each iteration for the user requests.

I. Introduction

Internet of Things (IoT) is a futuristic paradigm which enables connectivity between any type of devices [1]–[3]. In this regard, the sensor-cloud framework is envisioned to provide a scalable architecture to manage this ecosystem of enormous number of sensing devices [4]. The sensor-cloud offers a collaboration of the service provides [such as cloud service provider (CSP) and sensor owners (SOs)] and the users. The CSP and SOs provide services while gaining economical benefit in terms of price charged to end users. A real-life example of such framework is weather service. A service provider, for example AccuWeather (https://www.accuweather.com/), provides weather services to end users by utilizing the cloud services (such as Amazon Web Services) and the weather stations deployed by National Oceanic Atmospheric Administration (NOAA) or private TV stations. Thus, NOAA is one SO in this example. The advent of sensor-cloud framework empowers various application domains providing numerous advantages compared to the traditional wireless sensor network (WSN)-based infrastructure [5]–[9]. It enhances the real-time information processing and storage with the cloud-based framework where the on-field nodes are deployed covering a vast geographical area. The sensor-cloud architecture facilitates dynamic access and resource management of the physical sensory resources by providing a virtualized interface between the end users and the sensory resources. In this infrastructure, using the technique of virtual sensor provisioning, these physical sensors are accessed by end users using various different services offered by the CSP. Therefore, based on all these features, the sensor-cloud system provides a multiuser multiapplication environment for designing decision support systems. Few potential applications of the sensor-cloud architecture to name in different domains are health-care, precision agriculture, environmental monitoring, and military.