Clustering Strategy Performance Analysis of Cloud DVR for Energy Efficiency

Abstract

Comcast's cloud digital video recorder (cDVR) is a new service available to its customers. In the Cablevision judgment, the main current legal interpretation allowing cloud DVR is based on a single copy. As a result, the cDVR data center's operating costs are very expensive. To save energy, an asynchronous service system with user classification based on cDVR use is used. CDVRs with comparable use schedules are built in one cluster in this system. If there have been no cDVR requests for a period of time, the cloud recording service on this cluster goes to sleep. When one or more cDVR requests arrive, they are delayed in queues until the cDVR service comes up. To evaluate the performance of this system, a 2-class Markov Geo vacation model is given in this article. In simulations and experiments, different scheduling strategies are compared. Distributed networking may help in situations where resource sharing is a problem or when greater fault tolerance is required. Higher degrees of anonymity are also well supported by distributed networking. Cloud computing is a kind of computing that uses thunder the conventional client/server computing paradigm, enterprises with fast growth and scalability requirements may find it difficult to manage their own distributed network