Date of Award


Document Type


Degree Name

Doctor of Philosophy (PhD)


Electrical and Computer Engineering (Holcomb Dept. of)

Committee Member

Professor Haiying Shen, Committee Chair

Committee Member

Professor Richard R. Brooks

Committee Member

Professor Adam Hoover

Committee Member

Professor James (Zijun) Wang


Cloud computing has proved to be an effective infrastructure to host various applications and provide reliable and stable services. Content delivery and storage are two main services provided by the cloud. A high-performance cloud can reduce the cost of both cloud providers and customers, while providing high application performance to cloud clients. Thus, the performance of such cloud-based services is closely related to three issues. First, when delivering contents from the cloud to users or transferring contents between cloud datacenters, it is important to reduce the payment costs and transmission time. Second, when transferring contents between cloud datacenters, it is important to reduce the payment costs to the internet service providers (ISPs). Third, when storing contents in the datacenters, it is crucial to reduce the file read latency and power consumption of the datacenters. In this dissertation, we study how to effectively deliver and store contents on the cloud, with a focus on cloud gaming and video streaming services. In particular, we aim to address three problems. i) Cost-efficient cloud computing system to support thin-client Massively Multiplayer Online Game (MMOG): how to achieve high Quality of Service (QoS) in cloud gaming and reduce the cloud bandwidth consumption; ii) Cost-efficient inter-datacenter video scheduling: how to reduce the bandwidth payment cost by fully utilizing link bandwidth when cloud providers transfer videos between datacenters; iii) Energy-efficient adaptive file replication: how to adapt to time-varying file popularities to achieve a good tradeoff between data availability and efficiency, as well as reduce the power consumption of the datacenters. In this dissertation, we propose methods to solve each of aforementioned challenges on the cloud. As a result, we build a cloud system that has a cost-efficient system to support cloud clients, an inter-datacenter video scheduling algorithm for video transmission on the cloud and an adaptive file replication algorithm for cloud storage system. As a result, the cloud system not only benefits the cloud providers in reducing the cloud cost, but also benefits the cloud customers in reducing their payment cost and improving high cloud application performance (i.e., user experience). Finally, we conducted extensive experiments on many testbeds, including PeerSim, PlanetLab, EC2 and a real-world cluster, which demonstrate the efficiency and effectiveness of our proposed methods. In our future work, we will further study how to further improve user experience in receiving contents and reduce the cost due to content transfer.