A multi-Agent based model for task scheduling in cloud-fog computing platform

The Internet-of-Things (IoT) is a network of physical objects that communicate and exchange data between them over the Internet. These objects, or things, generate a massive amount of data that require powerful analytics approaches and data storage platforms - this is where cloud computing may well come into play. The classic cloud computing paradigm, however, faces several serious issues, due to centralization, such as low performance, latency issues, security issues, bandwidth obstacle, etc.. The Fog (Edge) computing paradigm has been proposed to improve the performance and to address other weaknesses, by providing IoT data processing and storage capabilities locally at the edges (IoT devices), instead of sending all the data to the centralized cloud. Whether in cloud or in fog computing, with billions of devices and a huge number of simultaneous requests, managing limited resources becomes essential. One of the important steps in resource management is task scheduling. In this paper, we consider task scheduling in a cloud-fog computing platform, we propose a multi-agent based model that aims at serving the most important task first, taking into consideration the task priority, its wait time, its status and the resources required to complete it successfully. Furthermore, in the scheduling process, we first propose an update to the priority value of the task, while taking into consideration its dependencies to other tasks and their priorities. Simulation results show that our proposed model can lead to better resource utilization and improve performance.