期刊名称：Journal of King Saud University @?C Computer and Information Sciences
印刷版ISSN：1319-1578
出版年度：2022
卷号：34
期号：4
页码：1498-1510
语种：English
出版社：Elsevier
摘要：The Internet of Things (IoT) is a model where devices and applications interact together to provide smart services in different contexts like smart homes, or smart vehicles. One of the challenges of IoT is the diversity of vendors who provide the IoT devices and the IoT applications. Such diversity leads to reduced utilization of the provided computing resources within IoT computing platforms, due to the different hardware requirements of each vendor for his developed applications, and their corresponding devices. Thus, the same computing resources could be duplicated in the same IoT platform, with minor variations, to fit different vendors needs. Such duplication revsults in extra costs on the application vendors, and reduced utilization of those computing resources. Another consequence of such diversity is the lack of a common interface for communication across IoT devices and applications from different vendors. Such common interface should present a level of abstraction that would hide both the vendor-specific and the IoT platform-specific implementation details, to provide seamless integration of new IoT devices and applications within an IoT platform. This paper presents a novel architecture for an IoT computing platform, namely Hub-OS, for managing devices and applications from different vendors with efficient utilization of computing resources. IoT devices and applications could be easily integrated without the need to modify either Hub-OS software components, or the vendor-specific applications and devices. Furthermore, Hub-OS could host IoT applications that demand real-time processing if needed. To evaluate Hub-OS, we implemented one case study for a smart vehicle, and conducted a quantitative evaluation using varied configurations. Our quantitative evaluation shows that Hub-OS can integrate applications and devices from different vendors, with reduced latency, reduced CPU usage, efficient IoT application startup and migration times.
