Proxy servers have been used to cache web objects to alleviate the load of the web servers and to reduce network congestion on the Internet. In this paper, a central video server is connected to a proxy server via wide area networks (WANs) and the proxy server can reach many clients via local area networks (LANs). We assume a video can be either entirely or partially cached in the proxy to reduce WAN bandwidth consumption. Since the storage space and the sustained disk I/O bandwidth are limited resources in the proxy, how to efficiently utilize these resources to maximize the WAN bandwidth reduction is an important issue. We design a progressive video caching policy in which each video can be cached at several levels corresponding to cached data sizes and required WAN bandwidths. For a video, the proxy server determines to cache a smaller amount of data at a lower level or to gradually accumulate more data to reach a higher level. The proposed progressive caching policy allows the proxy to adjust caching amount for each video based on its resource condition and the user access pattern. We investigate the scenarios in which the access pattern is priorly known or unknown and the effectiveness of the caching policy is evaluated.

Due to the high bandwidth requirement and rate variability of compressed video, delivering video across wide area networks (WANs) is a challenging issue. Proxy servers have been used to reduce network congestion and improve client access time on the Internet by caching passing data. We investigate ways to store or stage partial video in proxy servers to reduce the network bandwidth requirement over WAN. A client needs to access a portion of the video from a proxy server over a local area network (LAN) and the rest from a central server across a WAN. Therefore, client buffer requirement and video synchronization are to be considered. We study the tradeoffs between client buffer, storage requirement on the proxy server, and bandwidth requirement over WAN. Given a video delivery rate for the WAN, we propose several frame staging selection algorithms to determine the video frames to be stored in the proxy server. A scheme called chunk algorithm, which partitions a video into different segments (chunks of frames) with alternating chunks stored in the proxy server, is shown to offer the best tradeoff. We also investigate an efficient way to utilize client buffer when the combination of video streams from WAN and LAN is considered.

The advances in storage, host I/O, and networks entice a variety of on-demand services. To effectively use these resources, we propose a Scalable and Extensible Architecture for Multimedia Entertainment (SESAME) to explore the modularity and scalability of a Karaoke-on-demand (KOD) service. Karaoke popularizes MTV-style entertainment to let viewers sing along. It is shown that hierarchical architecture for KOD scales well and is cost-effective in the network environment where latency is not a major concern. The modularity and scalability are achieved by: 1) distributing Karaoke data at different levels of storage based on popularity to leverage the cost of near-line and on-line storages, and alleviate hot spot access for popular titles, 2) modularizing the server design as well as service hierarchy, and 3) reducing the concurrent accesses to a server by expanding service levels such that an inexpensive computer can be a server. We illustrate live applications in a Karaoke house and in a campus-wide setting. The SESAME approach is being extended to some other areas of interests such as distance learning.

An effective network-based learning environment provides synchronous and asynchronous education to learners who are physically, remote from or opt to telecommute to the base location of instruction. The environment provides content creation, user registration, course material storage, multiple delivery means, and end-user viewing tools. An intriguing aspect of the environment is how to model a lecture in terms of pre-, in-, and post-class information. We regard a lecture model collectively as an activity consisting of three elements: lecture materials, lecture presentation, and lecture interaction. We derive the synchronization specification for the media objects and define strong-weak networked hyperlink (SWNH) to represent the lecture structure and inter/intra media object reference in the distributed network environment. The specification and representation exhibit an object-oriented modular design. The presentation of the lecture embody both the event-driven and timeline-based presentation synchronization where the browsing and authoring tools developed earlier fit right into the learning environment, e.g. FLIPS and HQT by the authors, or are immediately available through other providers, e.g. Web browsers.