Security-Aware Scheduling of Multiple Scientific Workflows in Cloud

With the advent of cloud computing benefits of unlimited computing resources at low cost and flexibility due to virtualization technology, has attracted attention for executing complex scientific processes. The scientific workflows comprise of interdependent and heterogenous tasks. The cloud provides different instances of VMs (computing, memory, storage) and different security services (such as authentication, integrity, confidentiality) for secure processing to this complex workflow. It becomes challenging to select the right combination of VM instance type to optimize overall cost and processing duration (makespan) and selecting the right combination of security levels to keep the risk rate within permissible limits. In this work, we propose OptReUse- a workflow scheduling algorithm for efficient utilization of VMs. OptReUse combined with evolutionary algorithms gives lower cost of scheduling compared to prior art. We demonstrate the benefits of scheduling multiple workflows by OptReUse algorithm. We propose an enhanced security model for accurate estimation of risk and study the impact of different security levels on cost and makespan.