RPOM—Rational Process Offloading for Improving the Resource Utilization of the Internet of Things

Abstract

The Internet of Things (IoT) interconnects diverse objects and service platforms for providing ubiquitous application support through diverse communication technologies. Quality of service and experience required application support is ensured through precise resource allocation and request scheduling in this platform. Considering the densely populated user scenario and service demand, this article introduces a Rational Process Offloading Method (RPOM) for reducing the service backlogs in IoT. This method distinguishes the scheduled and offloading of required application requests for preventing additional service delays. The decisions for offloading and time-sensitive application responses are performed using the state learning process. In this learning, the offloading and scheduling states are validated using current and previous state analysis for independent and congestion-less responses. The state is trained using the time and offloading demand observed between the request and response. However, the varying state modeling is used for determining a forecast-based service allocation, improving the utilization. The RPOM’s performance is validated using the metrics of resource utilization, offloading ratio, service delay, and backlogs