Tag Archives: Bandwidth-Latency-Minimization algorithm

Ph.D. Student Xinjie Guan selected for GHC Research Scholarship

Selected from a competitive pool of over 1000 applicants, SCE Computer Science Ph.D. student Xinjie Guan has been awarded an ISOC (the Internet Society) scholarship to present her research at the 2012 Grace Hopper Celebration of Women in Computing (GHC) being held October 3-6, 2012 in Baltimore, MD. Xinjie Guan’s poster session, “Push or Pull?: Toward Optimal Content Delivery using Cloud Storage”, will showcase the novel distributed algorithm, named Bandwidth-Latency-Minimization (BLM), which she developed under the direction of her faculty advisor, Dr. Baek-Young Choi. Real experiments as well as simulations have shown that the BLM significantly optimizes content delivery using cloud storage for video-over-IP applications. Academic achievement, potential in the field, need and thoughtful, creative, well-written essays were part of the selection criteria used by GHC to determine this year’s poster session participants. Congratulations Xinjie!


Abstract: Cloud computing and ‘Storage As A Service’ (SaaS) are experiencing a momentous popularity increase due to its flexible, and scalable access to resources. Especially, cloud storage is becoming an economical alternative to traditional content delivery networks (CDNs) such as Akamai and Limelight Networks for moderate-size content providers. Previous research on content distribution mainly focuses on reducing latency experienced by content customers. A few recent studies address the issue of bandwidth usage in CDNs, as the bandwidth consumption is an important issue due to its relevance to the cost of content providers. However, few works consider both bandwidth consumption and delay performance for the content providers that use cloud storage with limited budgets, which is the focus of this paper. We develop an efficient light-weight approximation algorithm toward the joint optimization problem of content placement. We also conduct the analysis of its theoretical complexities. The performance bound of the proposed approximation algorithm exhibits a much better worst case than those in previous studies. We further extend the approximate algorithm into a distributed version that allows it to promptly react to dynamic changes in users’ interests. The extensive results from both simulations and Planetlab experiments exhibit that the performance is near optimal for most of the practical conditions.

Significance: Video-over-IP applications are experiencing a momentous popularity increase via crowd-acceleration. Content delivery using distributed caching on cloud storage, such as Amazon Simple Storage Services, can alleviate high bandwidth demands of such applications and can significantly cut down the costs in building and maintaining servers comparing with traditional Content delivery networks. However, the latency experienced by content users and the cost of provisioning VoIP services including bandwidth and storage space are heavily depended on content placement and delivery strategies. Few prior researchers have considered saving bandwidth consumption together with latency performance in neither traditional CDNs nor cloud storage. Our work aims to optimize the content delivery using cloud storage.

Publications: A preliminary version of this work was published in International Conference on Communication (ICC) 2011. It has now been extended with a novel distributed algorithm for cloud storage; and the scheme has been extensively evaluated using Planetlab testbed.