Blockchain Based Deep Pyramidal Residual Fractional Maxout Network for Workflow Scheduling in Cloud Computing

Cloud services are adapted to several problems of real-time computing. Users access diverse cloud services to execute numerous tasks. Additionally, it is essential to consider diverse quality of service (QoS) measures to enhance service performance. The impact of blockchain approaches helps to improve QoS metrics. In this work, the blockchain-based deep pyramidal residual fractional maxout network (Blockchain_DPRFMN) is devised for workflow scheduling in cloud computing. First, a cloud with a blockchain system model is simulated and then workflow scheduling is performed based upon multi-objective parameters such as virtual machine (VM) parameters and task parameters. Central processing unit (CPU) and bandwidth utilization are the VM parameters; whereas, task parameters include task priority, earliest finish time, task length, earliest start time, actual task running time, predicted energy, trust, makespan equivalent of total cost, memory capacity, reliability, and resource utilization. The deep feedforward network (DFFN) is utilized to predict energy. Finally, task scheduling is accomplished by Blockchain_DPRFMN, which is designed by integrating deep pyramidal residual network (PyramidNet) and deep maxout network (DMN) with fractional calculus (FC). Moreover, Blockchain_DPRFMN has attained a maximal residual energy of 0.769J, a minimal makespan of 0.580, resource utilization of 0.627, and latency of 0.485 sec.