Smart Virtualized Platform for Cloud Management Systems

点击翻译按钮获取中文摘要

In this strategic project, we plan to investigate a novel design of a comprehensive framework for smart cloud management that offers carrier-grade quality guarantees. The main idea is to innovatively utilize different optimization frameworks to address on-demand elasticity of the virtualized platform while considering the characteristics of the existing workloads (e.g., VNFs), components relations, computational constraints, QoS requirements, and service requests' characteristics. Moreover, we aim to employ various data analytics and Machine Learning techniques to accurately model the behaviors of the VNF workload and predict resource utilization and workload demands. Resource utilization and workload behaviors patterns will be identified. For example, features such as workload spikes, frequent changes in the application traffic/state/volume, services' QoS requirements, and interdependencies between components to handle a workload (i.e., VNF) will be explored.