Towards Autonomic Cloud Configuration and Deployment Environments

Cloud configuration deployment and management is still largely a manual task carried out by system administrators. Introducing autonomy in Cloud management would entail, amongst other things, the ability for the Cloud manager to automatically scale up or down the number and type of deployed images/virtual machines, to meet SLAs for performance etc. In this paper we present auto Ju Ju, a prototype autonomic cloud manager built on top of Juju, a Cloud service orchestration and deployment manager for the Ubuntu Linux OS. Auto Ju Ju makes autonomous decisions about when to scale Cloud services horizontally (by adding or removing instances) and vertically (by deploying different architectural components) to improve performance. We show how this autonomic Cloud manager can self configure and optimise a Cloud deployment.