Hierarchical Model and Sensitivity Analysis for a Cloud-Based VoD Streaming Service

Cloud computing environments provide storage capacity, processing power, and other computational resources in a flexible way, enabling fast adaptation to highly dynamic workloads. Multimedia services, such as video streaming, are examples of applications that can use cloud computing to leverage their provisioning capacity. This way, it is possible to offer a large variety of multimedia content in many formats, so the users will be able to watch videos as they wish, with a proper resolution and quality, according to his preferences and connection speed. Private infrastructures for Video on Demand (VoD) and live video streaming are especially useful for e-learning on large corporations, universities, and governments. Analytical models are effective tools to evaluate the availability of software, hardware, and other computational resources. In this paper, we study a VoD service hosted in a private cloud computing environment. We present availability models considering the VoD streaming server components that are necessary for viewers access. Hierarchical modeling techniques are used to deal with the complexity of representing such system. Sensitivity analysis is used to determine the parameters that cause greatest impact on the availability, identifying which components require attention when attempting to achieve increased availability in a system. The proposed models are useful for planning private cloud infrastructures for VoD services.