Feedback Control Algorithms to Deploy and Scale Multiple Web Applications per Virtual Machine

Abstract

This paper presents feedback control algorithms to autonomously deploy and scale multiple web applications on a given Infrastructure as a Service cloud. The proposed algorithms provide automatic deployment and undeployment of applications and proportional-derivative scaling of the application server tier. The algorithms use utilization metrics as input and do not require a performance model of the application or the infrastructure dynamics. Moreover, our work supports deployment and scaling of multiple simultaneous applications per virtual machine (VM). This allows us to share VM resources among deployed applications, reducing the number of required VMs. The approach is demonstrated in a prototype implementation that has been deployed in the Amazon Elastic Compute Cloud.