Techniques for Optimizing Cloud Footprint

Infrastructure as a Service (IaaS) has emerged as a popular service model in the context of cloud computing. Examples of IaaS vendors include, but not limited to, Amazon's EC2, Rack space, GoGrid and the Google Compute Engine. Use of IaaS obviates the need for set up and maintenance of infrastructure and thereby boosts product development agility - a key in increasingly competitive landscape. However, the use of IaaS is much more expensive compared to use of an in-house data enter. This calls for development of techniques to minimize the cost overhead associated with the use of an IaaS without sacrificing its various benefits, e.g., elasticity of the cloud. In this paper we present novel techniques to optimize operational efficiency in the cloud. Specifically, we present three techniques targeted to different production scenarios. The techniques have been deployed in production and resulted in up to 50% reduction in operational costs for the target Netflix applications.