Sensorcloudsim : an application scheduling & a sensor cloud infrastructure simulator

Abstract

With the enormous growth of IOT sensing devices that is tending to the Internet of everything, the data aggregated by these devices are the biggest data streams generated in the history of the IT. For this reason, aggregating those data to the cloud for leveraging the powerful cloud computing processing and storage is essential. This led to the emergence of the sensor cloud concept which is not only a mere aggregation of the sensors’ data to the cloud for further processing, storage and visualization. But also, a virtualization of the sensors, making them accessible to different end users’ applications, which are requesting their data. This without the end user application developer being aware of the sensor localization and hardware. So, the revolution brought by the sensor cloud paradigm is that any end user can develop an application that requests any sensor deployed anywhere, by any organization, without knowing the sensor API, localization or hardware, as long as the sensor is registered to the sensor cloud infrastructure. From the above mentioned it is obvious that the scheduling of the applications requesting the sensors is of main concern, and that optimizing the scheduling policy would definitely lead to the optimizing of the quality of the sensing as a service to the end users. However, most of the time, the scheduling policies implanted by the sensor cloud platform provider are hidden from the users and leveraging the platform is not free. This makes experimenting scheduling policies on a real platform either not possible or really cost inefficient. Here is where simulation plays a big role in research. Having a simulation platform where sensor cloud infrastructures agents and components can be modeled, scheduling policies configured, and execution time assessed would be crucial to researcher trying to develop the infrastructure performance and quality of service. In this work we will be discussing the development of such a platform, by extending cloudsim, a famous cloud computing simulation platform, and we will try to come up with an optimum quantum time in a time-sharing scheduling policy by experimenting on the simulator.