Decision Support Tool for IoT Service Providers: Cost-Performance Optimization for IoT-based Sensor-Actor Systems

Abstract

The Internet of Things (IoT) refers to the uniquely identifiable objects (things) and their virtual representations in an Internet-like structure. IoT has appeared on the stage, interconnecting a variety of physical objects over the Internet, enabling the objects to communicate and cooperate with each other to achieve predefined goals. It is predicted to be an integral component of the Future Internet (FI)

therefore, IoT should be seamlessly and smoothly integrated with other FI integrated services. Nevertheless, IoT devices are location dependent, and expensive to develop and deploy, because the IoT supporting infrastructure such as computing power, storage and networks are resource constrained. To fulfill these shortcomings, another recent phenomenon named Cloud computing can be the most promising and cost-effective solution. Indeed, Cloud offers relatively cheap, ubiquitous, unlimited and elastic solution for the supporting infrastructure. Therefore, to connect, manage and track the IoT-based devices and provide feasible access to a set of multitude computing resources, many IoT Service Providers (IoTSP) may utilize Clouds to offer their users certain services. Aiming to offer services to globally distributed users, an IoTSP will deploy its Virtual Machines (VMs) on multi Clouds that is consisting of various Cloud Service Providers (CSPs) to have satisfactory coverage and performance for the users. Integration of IoT with the multi Clouds raises new challenges among which the economic challenges are from the most critical factors for success of the IoTSP business. One of the major problems in this context is to minimize the overall cost of the IoT system while keeping satisfactory level of performance in order for the business to be profitable. To this end, the IoTSP has to maintain its IoT-devices cost-optimally and to place its VMs on the available CSPs cost-optimally. In other words, the problem for IoTSPs is finding the cost optimum placement for their IoT devices and VMs. This dissertation addresses the problem by proposing a decision support tool for IoTSPs to find their cost-optimum devices and VM placement on multi Clouds. Sound system architecture is designed for the tool to carry out the tasks. The tool comprises a heuristic algorithm for IoT device placement, a cost estimation module as well as VM placement optimization algorithm. The cost estimation module estimates the total infrastructure costs for any given VM placements considering multi Cloud environment. Applying the proposed optimization algorithms on the estimated cost and the expected performance returns the cost-optimum IoT device and VM placement for the IoTSP. The proposed decision support tool is examined by several simulation scenarios and the results demonstrate the working of the tool.