Extending CloudSim to simulate sensor networks. (January 2023)

• Record Type: Journal Article
• Authors: Habaebi, Mohamed Hadi; Merrad, Yaçine; Islam, Md Rafiqul; Elsheikh, Elfatih A A; Sliman, F M; Mesri, Mokhtaria
• Is Part Of: Simulation. Volume 99:Number 1(2023)
• Page Start: 3
• Page End: 22
• Publication Date: 2023-01
• Subjects: Cloud computing -- CloudSim -- modeling -- scheduling policies -- SensorCloudSim -- simulation platform; Computer simulation -- Periodicals; 003.3
• DOI: 10.1177/00375497221105530
• Languages: English
• ISSNs: 0037-5497

Abstract

With the enormous growth of sensing devices tending to the use of Internet of everything, data aggregated by these devices are the biggest data streams generated in the history of IT. Thus, aggregating such data in the cloud for leveraging powerful cloud computing processing and storage is essential, and it eventually led to the emergence of Sensor-Cloud concept. This has allowed aggregation of the sensors' data to the cloud for further processing, storage, and visualization. Furthermore, virtualization makes the sensors accessible to other end-user applications that require such data. All of these features are expected to be provided by the Sensor-Cloud invisibly, without the end-user application developer being aware of the sensor location or hardware specifications. For these reasons, a simulation platform where Sensor-Cloud infrastructure agents and components may be modeled, scheduling policies defined, and execution time assessed is essential to assure performance and quality of service. The aim of this study is to develop such a platform by enhancing CloudSim, the most well-known and powerful simulation tool for cloud computing. A user-friendly Java Script Swing-based graphical user interface (GUI) has been carefully designed and implemented for this purpose. The user can then utilize the specific interface to define the Cloudlet type as well as the scheduling on a single virtual machine. Finally, a simulation study is carried out on the platform to demonstrate its efficiency and accuracy. We were able to fully model the needed scenarios and acquire real-time results, displaying good accuracy in terms of application response time with a mean absolute percentage error (MAPE) of 3.37%, demonstrating the increased proposed platform's proper operation.