TOTO: An Intelligent Task Scheduling Framework for IoT Applications in Multi-Cloud Environment

Abstract

Internet of Things (IoT) applications and the increasing demand for cloud-based services have created major challenges in task scheduling for energy efficiency and carbon awareness in multi-cloud environments. This study proposes an Energy and Carbon Emission Aware Task Scheduling framework for IoT applications based on Three-on-Three Optimizer (TOTO) to achieve efficient sustainable resource allocation and execution. This proposed model is implemented and tested in WorkflowSim simulation environment with DigitalOcean Cloud workload. The TOTO-based scheduler dynamically optimizes the allocation of tasks by reducing energy consumption, carbon emissions, and at the same time enhancing the performance metrics of the system, like Service Level Agreement (SLA) Violations, Quality of Service (QoS), Cost, and Throughput. The experimental results show that the proposed TOTO approach outperforms conventional optimization methods such as the Genetic Algorithm (GA), Particle Swarm Optimization (PSO), and Deep Reinforcement Learning (DRL). The proposed method reduced the SLA Violations by 31.4% and the execution Cost by 28.7% and improved the QoS by 24.9% and Throughput by 33.6% than the existing methods. Moreover, the TOTO scheduler was able to effectively minimize overall energy consumption and carbon emissions by optimizing workload distribution among several cloud resources. The results demonstrate the potential of the proposed solution for supporting sustainable, reliable and high-performance execution of IoT applications in modern multicloud platforms.