A fault-tolerant resource reservation model in cloud computing
Sheikh Umar Mushtaq, Sophiya Sheikh · 2023
Making scheduling fault proof is one of the critical challenges in any dynamic environment like the cloud. Fault tolerance makes the system more efficient and thereby continues the processing of any task till completion without any disruption. This paper proposes Advance Reservation Fault Tolerance Model (ARFTM), which minimizes the Makespan in combination with making the system fault-proof. ARFTM uses a resource reservation strategy and reserves the virtual machines (VMs) in advance for the certainly predicted timeslot. In case of fault, the system provides an alternative VM to the affected task for successful task execution. Here, the alternative VM is chosen based on the previous workload history. The proposed ARFTM was compared with traditional MCT on various parameters like Makespan, flowtime, and average VM utilization. The results show that the proposed ARFTM overpasses the compared approach.