Abstract:
An improved smallest load first(SLF) placement algorithm is proposed,which focuses on the sharp changes of popularity of time-shifted TV.With an objective of maximizing the number of replicas of each video,the placement constraints are changed,and dual-threshold dynamic adjustment is used as an assistant policy.As a result,the effect of popularity changes on the system load balance is reduced.Simulation results reveal that,with the proposed policy,the user request reject rate is reduced,and the system performance is improved,especially when the actual demand is not coincident with the expectation.