Optimal Resource Allocation for Brokers in Media Cloud
Due to the rapid increases in the population of mobile social users, providing the users with satisfied multimedia services has become an important issue. Media cloud has been shown to be an efficient solution to resolve the above issue, by allowing mobile social users to connect to it through a group of distributed brokers. However, as the resource (like bandwidth, servers, computing power, etc.) in media cloud is limited, how to allocate resource among media cloud with brokers becomes a challenge. Media cloud can determine the price of the resource and a broker can decide whether it will pay the price for the resource when there is an incoming multimedia task (simplified as task). A broker can collect the revenues from the mobile social users by providing the multimedia services. Since resource is limited, the price will generally go up as the resource becomes more and more consumed. Therefore, in this paper, by assuming that accepting each task a broker can get a reward (by collecting revenues from mobile social users like online ads, etc.) and it needs pay some price (to the media cloud) for each task in the network, we concentrate on the optimization problems of when to admit or reject a task for a broker in order to achieve the maximum total discounted expected reward for any initial state. By establishing a discounted Continuous-Time Markov Decision Process (CTMDP) model, we verify that the optimal policies for admitting tasks are state-related control limit policies. Our numerical results with explanations in both tables and diagrams are consistent with our theoretic results.
KeywordsMedia cloud Cloud broker Mobile social user Resource allocation Optimal control policy Continuous-Time Markov Decision Process (CTMDP)
- 5.Qiu, X., Wu, C., Li, H., Li, Z., Lau, F.: Federated private clouds via brokers marketplace: a Stackelberg-game perspective. In: IEEE 7th International Conference on Cloud Computing, pp. 296–303, June 2014Google Scholar
- 10.Ross, S.M.: Stochastic Process. Wiley, New York (1983)Google Scholar