《Auto-Scaling to Minimize Cost and Meet Application Deadlines in Cloud Workflows》.pdf

Auto-Scaling to Minimize Cost and Meet Application Deadlines in Cloud Workflows Ming Mao Marty Humphrey Department of Computer Science Department of Computer Science University of Virginia University of Virginia Charlottesville, VA 22904 Charlottesville, VA 22904 ming@ humphrey@ ABSTRACT applications and high-memory machines for databases. An auto- A goal in cloud computing is to allocate (and thus pay for) only scaling framework must recognize that these different instance those cloud resources that are truly needed. To date, cloud types are often not priced linearly according to their processing practitioners have pursued schedule-based (e.g., time-of-day) and power. Finally, it should not ignore the practical considerations of rule-based mechanisms to attempt to automate this matching real cloud deployments, such as that additional computing between computing requirements and computing resources. capacity such as VMs can generally be acquired at any time but However, most of these “auto-scaling” mechanisms only support could take several minutes to be ready to use [2]. Also, it cannot simple resource utilization indicators and do not specifically ignore that current cloud