VM Auto-Scaling for Workflows in Hybrid Cloud Computing

Appearance of Science Clouds enables scientists to facilitate large-scale scientific computational experiments over cloud environment. Many task computing (MTC) in computational science needs to certificate stable executions of applications even in rapid changes of vital status of physical resources and supports high performance resources in a long period. Auto-scaling approach on virtual machines (VM) increases efficient cloud resources management for the computational problem solving environment. Diverse auto-scaling methods which provide useful resource management presently are being debated and studied. However, most of the auto-scaling methods are just easily considered in performance metrics or execution deadline in specific workloads but not in various patterns of workflow. We propose an auto-scaling method, guaranteeing the execution of various patterns of workflow within deadline time in hybrid cloud environment. The experimental results show the method works dynamically and acceptably on hybrid cloud resources for various workflow patterns having random workload dependency.