An Effective Virtual Machine Allocation in Federated Cloud by PARAMR-DNN Technique

To advance the Federated Cloud (FC) services, efficient resource management in an FC is essential. Therefore, ParaMR-DNN-based effective Virtual Machine (VM) allocation in FC is proposed. Initially, the cloud users send the request to the cloud server. Then, the features are extracted and passed as input to the Cloud Broker (CB), which performs Load Balancing (LB) by employing the Fuzzy Inverse Multiquadric (FIM) approach. Later, by employing the Siberian Tiger Weighted Policy Optimization (STWPO) technique, the features from the Cloud Service Provider (CSP) and user requests are utilized to select the optimum CSP. Additionally, to find the best VM path within CSP, the shortest VM path selection is executed. After that, it is subjected to the Paralyzed Multi-Resolution Deep Neural Network (ParaMR-DNN) classifier, which efficiently allocates the resources to the VM in FC. Finally, the performance evaluation showed that a high accuracy of 98.54% for Resource Allocation (RA) and less execution cost of 675 ms for CSP selection were obtained by the proposed framework. Therefore, the experimental outcomes revealed that the proposed model outperformed the prevailing works.