Redundant VoD Streaming Service in a Private Cloud: Availability Modeling and Sensitivity Analysis

Abstract EN

For several years cloud computing has been generatingconsiderable debate and interest within IT corporations.

Since cloud computing environments provide storage and processingsystems that are adaptable, efficient, and straightforward,thereby enabling rapid infrastructure modifications to be madeaccording to constantly varying workloads, organizations ofevery size and type are migrating to web-based cloud supportedsolutions.

Due to the advantages of the pay-per-use model andscalability factors, current video on demand (VoD) streamingservices rely heavily on cloud infrastructures to offer alarge variety of multimedia content.

Recent well documentedfailure events in commercial VoD services have demonstratedthe fundamental importance of maintaining high availability incloud computing infrastructures, and hierarchical modeling hasproved to be a useful tool for evaluating the availability ofcomplex systems and services.

This paper presents an availabilitymodel for a video streaming service deployed in a privatecloud environment which includes redundancy mechanisms inthe infrastructure.

Differential sensitivity analysis was appliedto identify and rank the critical components of the systemwith respect to service availability.

The results demonstrate thatsuch a modeling strategy combined with differential sensitivityanalysis can be an attractive methodology for identifying whichcomponents should be supported with redundancy in order toconsciously increase system dependability.