Efficient profit maximization in reliability concerned static vehicular cloud system

Modern vehicles are equipped with high-performance compute systems. These compute resources mostly stay idle as most of the time vehicles get parked in the parking lots. In this work, we propose to utilize the unused compute resources of the vehicles efficiently to enhance the computing power of regular cloud systems , which is termed as vehicular cloud . Unlike in traditional cloud computing resources, the vehicles or vehicular compute resources move in or out of the parking lot, which introduces dynamic nature of the available compute resources. This makes it challenging for the vehicular cloud to ensure reliability of execution of the user-submitted tasks. In this work, we propose an approach to maximize the profit of the vehicular cloud by ensuring the reliability of the vehicular cloud. We consider user-submitted tasks with execution time , deadline and revenue associated with it. Our approach classifies the tasks based on the deadline, and orders the tasks for task admission based on the expected profit of the task. We also perform the classification of available vehicular units based on the expected residency time of vehicles and use the same for allocating vehicular units for redundant execution of task to ensure higher reliability. As the task execution time has a direct impact on redundancy requirements to ensure higher reliability, we convert the longer tasks to a chain of shorter sub-tasks to reduce the redundancy requirement. Our experiments show that the proposed approach outperforms the state-of-the-art approach with a profit margin increasing up to 25 to 45 % in real-life scenarios.The codes and dataset for this work are available at our https://github.com/SuvarthiSarkar/Efficient-profit-maximization-in-reliability-concerned-static-vehicular-cloud-system.git GitHub repository. • Proposes a novel profit-maximization framework for vehicular cloud systems while ensuring high reliability, a combination not addressed in existing literature. • Enhances system reliability using MT99R (Mean Time to 99% Reliability), outperforming traditional MTTF-based (50% reliability) approaches. • Introduces a hybrid strategy combining checkpointing (task segmentation) and redundant sub-task execution to improve task completion success rates. • Incorporates intelligent scheduling strategies: classifying vehicles based on residency time, prioritizing tasks based on criticality, and overwriting least-progressing tasks to optimize resource usage.