Fault Identification in Deadline Sensitive Applications in Cloud Computing

Abstract

Portability is setting the stage for how individuals lap up online media content. Cloud computing has stirred incredible enthusiasm from both scholarly world and in the technical arena quite recently. Because of Service Level Agreement (SLA), the cloud service providers need to ensure the unwavering quality of their frameworks. Nonetheless, erratic equipment or programming disappointments are generally inescapable. For realtime undertakings, failure of the entire system may make them miss their due dates. Thus a robust and a fault tolerant system can effectively evade such cases. The fault tolerant system is generally used to acknowledge adaptation to internal failure by copying an errand into two duplicates a primary one and a secondary piece. The excess presented by reinforcement duplicates brings about additional overhead for cloud frameworks. A media handling model is proposed to empower dynamic scheduling synthesis and cross-pipeline assignment partaking in the cloud for adaptable online processing. The challenge in using cloud administrations for on-request video transcoding, be that as it may, is to keep up a vigorous QoS for watchers and cost-proficiency for spilling administration providers. The model enables application designers to stretch out its ability to empower certain customization for running live channels. In this paper we have discussed the issues that are prevalent in the fault tolerance, elaborated on the two algorithms that we have used to identify the faults.