James Kok Konjaang

Cloud computing is an internet based computing, where large number of shared virtual servers are hooked to both public and private networks to provide scalable and multi-tenants Infrastructure as a Service (IaaS), Software as a Service (SaaS), and Platform as a Service (PaaS) to cloud users on pay-per-use basis. Many researchers in different scientific domains such as astronomy, physics, earth science and bio-informatics have used scientific workflow applications over the past decades to model various scientific computing problems in both parallel and distributed systems such as Cluster and Grid computing. However, the growing number of complex scientific computing systems and the fact that it is very expensive executing workflows in the traditional on-premises systems like clusters and grids makes it imperative to perform workflow task execution in cloud computing environments because of its flexibility and utility oriented pay-per-use pricing model. Workflow scheduling in IaaS cloud is a challenge due to their size and the growing number of complex scientific computing systems which often require more time to execute. Moreover, the problem of energy consumption has become one of the major concerns in clouds. These problems become more obvious when the workflow tasks to be scheduled are dependent on each other. In view of this, this study aimed at optimizing execution cost, energy and makespan for scheduling workflows as well as load balancing in IaaS cloud. To achieve this, novel heuristic algorithms that use the opportunities and challenges of cloud computing while taking into account the differences in VMs performance and instance acquisition delay in optimizing workflow task assignment will be proposed.