Search

With the increasing popularity of cloud computing, there is an increased demand for cloud resources in cloud. It has become even more urgent to find solutions to improve resource utilization. From the perspective of a cloud consumer, a cloud application processes a large information flow involving user actions that access resources, but little work has so far been devoted to research from the perspective of the interaction between the user and the cloud application. In this paper, we analyze the interaction in detail, and propose a general mathematical interaction model to formulate the challenge pertaining to storage resource allocation as an optimization problem, focusing on minimizing both the user's cost and server's consumption. A potential response mechanism is then designed based on the interaction model. Furthermore, the proposed model is used to explore strategies when multiple users access the same file simultaneously. Additionally, an improved queuing system, namely M/G/ ∞ queue with standby, is introduced. Finally, an evaluation is presented to verify the interaction model.

A novel Internet content distribution service Offline downloading service has emerged recently and attracted a large number of users all over the world. Unlike traditional Internet content distribution systems such as Bittorrent (BT), Offline downloading service uses dedicated cloud unities (Cache cloud) to store files what users wish to download and provide uploading, in order to guarantee file availability and enhance data transfer rate. Besides, with the help of tracker server, P2P (peer-to-peer) swarms constructed by users benefit to save server-side (Cache cloud) bandwidth consumption of the novel service. In this paper, we conduct a two-month measurement study for evaluating the effectiveness of the novel Offline downloading service on improving user experience. Our study shows that the bandwidth resource of Cache cloud is expected to be insufficient in the near future. The relevant influential factors are found to be file popularity, time period and peer connection status. Guided by these findings, we develop a flexible bandwidth allocation strategy for Cache cloud, in propose of saving the Cache cloud bandwidth without affecting service quality by maximizing the cooperation between cloud-acceleration and peer-assistance.