An efficient hybrid scheduling and virtual machine framework for load balancing in cloud computing architectures

The proposed work makes several key contributions. First, a hybrid task scheduling strategy is developed using an Artificial Neural Network combined with a Fuzzy Inference System (ANFIS). Tasks are prioritized based on multiple criteria, after which an optimal virtual machine (VM) is allocated. Second, the VM topology is restructured into a Peer-to-Peer (P2P) configuration to simplify and enhance load balancing. Finally, load balancing is achieved through VM migration guided by the Zebra Optimization Algorithm (ZOA). The approach aims to reduce VM overload without increasing queuing or processing time. Performance is evaluated using the CloudSim simulator, considering metrics such as latency, execution time, migration time, number of migrated VMs, load level, and communication overhead. Simulation results demonstrate significant improvements across these parameters. The proposed framework, referred to as HTS-P2P (Hybrid Task Scheduling with Peer-to-Peer virtual machine interaction), integrates intelligent scheduling, decentralized load sharing, and optimized virtual machine migration to improve cloud resource utilization.