Intelligent classification and structure aware resource estimation for effective execution of workflows in the cloud

Abstract

Cloud computing has contributed a lot for the recent developments in information technology. It is very beneficial for small and medium sized organizations, as cloud resources are available at a cheaper cost than traditional computing methods. The ease of resource provisioning and releasing policies has created a comfortable environment for users to consume the cloud resources and pay only for what is being used. Due to the inherent features of cloud, workflow scheduling in the cloud is becoming popular and got the attention of scientists and researchers. Even it is existing for more than a decade still there is scope for innovations. Particularly, more focus is required to select a suitable cloud service provider, identifying an appropriate type of resource and applying a suitable algorithm for executing workflows in the cloud.This thesis presents an improved way for scheduling workflows in the cloud.The first step towards workflow scheduling in cloud is, representing the given workflow in a format convenient for latest schedulers to understand. The textual representation of workflows proposed in this work, simplifies the process of workflow scheduling. The textual format is able to represent even large workflows, without affecting the task dependencies. The use of critical path algorithm for scheduling scientific workflows reduces the overall execution time by 22%, when compared with other dominant scheduling algorithms. newline