Response Mechanism And Performance Analysis For Video-on-Demand System

VOD (Video-on-Demand) system is an important applied technique of multimedia.A typical VOD system is composed of servers, high-speed network and terminal. The relationship between response mechanism and performance in Video-on-Demand system is analy- zed by queueing theory and Stochastic Petri Nets.On the one hand, we set up some queue models to analyze the primary guide lines Firstly, according to queueing theory, we conducts the qualitive analysis to some of guide lines in the video service system with unicasting and analyzes some states of the video service system in different scenarioes .We consider the average number of customers'requests surviving during the cycle and the video server throughput as the main object to study. In this paper, we provide the connection between the average number of customers'requests surviving during the cycle and service load, customer patience factor. At the same time, we state the effection that service load and customer patience factor act on the video server throughput. By analyzing the results of simulation, we find this is also an approach to save the video server's resource efficiently and improve the video server's performance. Secondly, according to queueing theory, we conducts the qualitive analysis to some of guide lines in the video service system on batching and analyzes some states of the video service system in different scenarioes. We consider the average number of customers'requests surviving during the cycle and the video server throughput as the main object to study. In this paper, we provide the connection between the video server throughput, the average number of customers'requests surviving during the cycle and threshold value on batching, service load, customer patience factor. At the same time, we state the effection that threshold value on batching, service load and customer patience factor act on the consumer delay. By analyzing the results of simulation, we find this is also an approach to save the video server's resource efficiently and improve the video server's performance.On the other hand, A Stochastic Petri Nets model of the Video-on-Demand system is constructed and the performance of the system is analyzed by it. It is shown that the performance of the system is relative to the selection of response mechanism. If the rate of the"Respond to serve"transition is fast, the average delay of the serve will be less, the average defection of the customers will be lower and the average throughput of the system will be high. By analyzing some of performance guide lines, we find that there are some direction signification to improve the performance of system.In the end, we summarize the paper and put forward the direction that need to be further studied.