Thesis (M. Phil.)--University of Hong Kong, 2006. / Title proper from title frame. Also available in printed format.

Thesis (M. Phil.)--University of Hong Kong, 2007. / Title proper from title frame. Also available in printed format.

Thesis (M.S.)--Youngstown State University, 2009. / Includes bibliographical references (leaf 42). Also available via the World Wide Web in PDF format.

Thesis (Ph. D.)--University of Chicago, Dept. of Computer Science, December 2003. / Includes bibliographical references. Also available on the Internet.

Includes bibliographical references (p. 74-77).

published_or_final_version / Electrical and Electronic Engineering / Doctoral / Doctor of Philosophy

Peer-to-peer (P2P) video streaming has become a very popular and cost-effective solution to provide video-on-demand (VoD) streaming service to a large group of Internet users. Recent advances of layered video coding are shown to be promising in addressing the problem of receiver heterogeneity in P2P streaming networks. The key idea is to encode a raw video into multiple non-overlapped layers. Peers in the network can enjoy different streaming qualities by subscribing to different number of layers according to their specific network bandwidth. In this thesis, a systematic and comprehensive study of layered P2P VoD streaming networks is conducted. We first investigate the overlay construction problem that concerns the strategy of forming a logical overlay network for peers to exchange data pieces. We propose a neighbor relationship management algorithm to maintain a stable yet diversified neighbor set to facilitate adding or dropping a video layer, i.e. layer adaptation. Subsequently, an incentive scheme is designed to differentiate neighbors’ contributions at different layers, and a peer is allowed to adjust its number of neighbors according to its average service response time. We then develop an efficient layer adaptation algorithm for adjusting the number of video layers a peer subscribes to. We use a sliding buffer window for video piece scheduling, and make use of the time difference between the playback point and the starting point of the sliding window to initiate the process of adding or dropping a layer. A probing period is also introduced to ensure the adding decision is indeed correct. We next focus on designing piece scheduling algorithm for video data exchange among peers. At each peer, a utility function for calculating the relative importance of each missing video data piece is designed. In piece scheduling, a peer prefers a piece with higher utility value. We show that our utility based piece scheduling algorithm yields better video quality than conventional approaches. Last but not least, we study the request peer selection problem for identifying the most suitable neighbor for a peer to make a piece request. Two algorithms are designed, smallest service response time first (SSF) and closest playback point first (CPF). SSF ensures peers with larger uplink bandwidths to serve more requests, and CPF encourages a newly joined peer to contribute its uplink bandwidth as quickly as possible. We show that there exists a sweet spot by properly combining the two together. Although our algorithms are presented sequentially in addressing each of the problems above, they do interact with each other. Throughout our study, a comprehensive packet-level simulator is used to closely examine and analyze such interactions. With that, our algorithms are judicially designed and fine-tuned for best performance. To the best of our knowledge, our in-house developed simulator is the most comprehensive packet-level simulator for layered P2P VoD streaming. / published_or_final_version / Electrical and Electronic Engineering / Doctoral / Doctor of Philosophy

Peer-to-Peer (P2P) technology has been developed rapidly during the past few years. Due to its superiorities on robustness and scalability, P2P technology has been viewed as a promising networking technology and many studies have been done on how to improve P2P technology. P2P file distribution, as a major application of P2P technology, has also been studied a lot. The amount of time required for all peers to get the file has been considered as a major optimization metric, which we refer as the file distribution time. Researchers have proposed protocols to minimize the file distribution time for different cases. However, most of the existing works are based on the single-file scenario. On the other hand, studies show that in a file sharing application, users may download multiple files at the same time. In this thesis, we analyze the file distribution time for the distribution of multiple files in both wired and wireless networks. We develop explicit expressions for lower bound of time needed to distribute multiple files in a heterogeneous P2P fluid model. Unlike the single-file scenario, we demonstrate that the theoretical lower bound in multi-file scenario is not always achievable. With a comprehensive consideration of all the configurations, we develop algorithms to partition the bandwidth of all the peers for a particular file such that the file distribution time is optimal. / published_or_final_version / Electrical and Electronic Engineering / Master / Master of Philosophy

As the Internet develops and technology evolves, multimedia traffic dominates the Internet. Traditional client-server based video streaming solutions are expensive and suffering from the scalability issue. Inspired by the immense success and efficiency of BitTorrent in distributing file contents to large groups of users, the peer-to-peer (P2P) paradigm has been adopted to provide streaming services including both P2P video-on-demand (VoD) and P2P live streaming. Among the design challenges, one open yet fundamental question is, how to identify the most suitable neighbor for a peer to make a piece request. We refer this problem as the request peer selection problem. Properly allocating the piece request to potential providers is essential to the system and would help to balance the load at each peer. A well designed request peer selection algorithm increases the requestor’s probability to download the desired piece timely which leads to enhanced video quality and improved playback continuity. This thesis aims to enhance the overall system performance by proposing optimized request peer selection algorithms for both P2P VoD and P2P live streaming systems. In P2P VoD, participating peers start playback from the very beginning of a stored video. The playback-points of peers, as well as the amount of video contents/pieces the peers cached, depend on when they join the video session, or their viewing ages. The content buffered by younger peers is a subset of older peers, thus collaborative piece exchange among peers is undermined due to the unbalanced supply and demand. To address this issue, a playback-point based request peer selection algorithm: closest playback-point first (CPF) is proposed. Specifically, when a peer requests a particular video piece, among the set of potential providers, a request is sent to the peer that has the smallest playback-point difference with itself. With CPF, peers with similar content availability are loosely grouped into clusters to maximize the utilization of individual peer’s upload capacity. Extensive packet level simulations show that with CPF, the video playback quality is enhanced and the VoD server load is significantly reduced. Unlike P2P VoD, P2P live streaming systems have to meet real-time playback constraints, which makes it even more challenging. To better regulate the network traffic and to balance the load among peers, a service response time (SRT) based request peer selection algorithm is proposed. In particular, a peer in the network estimates the service response time (SRT) between itself and each neighboring peer. SRT is measured from when a data piece request is sent until the requested piece arrives. When a peer makes a piece request, the neighbor with smaller SRT and fewer data pieces would be favored. The rationale is: smaller SRT implies excess serving capacity and fewer data pieces suggests fewer potential piece requests received. Extensive packet level simulations show that the traffic load in the network is better balanced, the streaming server load is reduced, and the overall quality of service, measured by playback continuity, startup delay etc., is improved. / published_or_final_version / Electrical and Electronic Engineering / Master / Master of Philosophy

Peer-to-peer (P2P) paradigm, for its scalability and low cost management, is widely used in today’s network. Based on the typical designs for request/response services, a lot of efforts have been made to support publish-subscribe services in P2P networks. Gossip-based publish-subscribe system, which is commonly used in unstructured P2P networks, can provide great flexibility in query language and does not require special efforts on maintaining topology. The purpose of our work is to investigate effective and efficient mechanisms to build gossip-based publish-subscribe systems in unstructured P2P networks. Specifically, the probabilistic bi-quorum system (PBQS), for its assurance in effectiveness, becomes the object of our study. Uniform sampling is a fundamental tool to construct PBQS. By adopting uniform sampling, PBQS provides a bound on the likelihood that data messages will find a copy of the subscription. A random walk of length O(log n) is commonly used to gain a uniform sample on an expander graph of size n. To obtain a multitude of uniform samples thus requires an equivalent number of random walks of length O(log n) each. A number of works have relied on the Chernoff bound to analytically reduce the overhead needed to obtain a multitude of uniform samples. Besides, researchers have also shown that it is not necessary to replicate both data and query on uniformly chosen nodes. Alternatively, BubbleStorm performs controlled flooding on a constructed overlay to build PBQS. BubbleStorm does not require nodes forming a bubble to be uniformly chosen at random, and the probabilistic bound computed by BubbleStorm is different from uniform sampling based PBQS. In this thesis, we first show that the Chernoff bound on the statistical properties of samples collected from a random walk does not help in selecting uniformly random nodes. We then re-examine the role of uniform sampling in PBQS, and found that when multiple data answer a single subscription, it is sufficient and necessary for each data to be distributed uniformly at random. Looking into BubbleStorm, we examine more closely the probabilistic bound provided by this system. We found that, unlike uniform sampling based PBQS, the bubble intersection in BubbleStorm is distance dependent. Given a specific pair of publisher-subscriber, the data may never find the subscription. We further investigate the topology construction and found that re-creating topology prior to each controlled flooding or keeping topology with high degree of churn can help alleviate the distance dependency problem. We arrive at the conclusion that BubbleStorm construction is equivalent to caching of random walks. We show that re-using this cache to obtain samples over time leads to degradation of uniformity of the samples. We evaluate topology re-wiring as a simple method to keep the cache fresh, thereby benefiting from the low latency of controlled flooding without degrading the uniformity of samples over time. / published_or_final_version / Electrical and Electronic Engineering / Master / Master of Philosophy