Wing-Fai Poon

Adaptive batching scheme for multicast video-on-demand systems

In this paper, an adaptive algorithm is developed for providing true video on demand (VoD) services in a multicast environment. In conventional batching schemes, the batching time of the system is fixed and the performance of such static schemes is highly dependent on the selection of the batching time. If the batching time is wrongly estimated, the performance of the system will be greatly degraded. Our algorithm tries to dynamically find the optimal batching time by the newly updated arrival rate so as to minimize the bandwidth requirement. The results show that the system performance of the adaptive approach is always better than the static scheme in terms of total bandwidth requirement and customer reneging probability, especially in using long batching time for the high arrival rate.

Design of multicast delivery for providing VCR functionality in interactive video-on-demand systems

Multicast delivery is one of the solutions to reduce the cost in a large video-on-demand (VoD) system. However, multicast transmission makes much more difficult the implementation of interactive functions for individual users and introduces start-up delays for the users, which contradicts the idea of on-demand services. In this paper, we first try to explore and evaluate the performance of different multicast VoD systems. A new scheme called single-rate multicast double-rate unicast (SRMDRU) is then developed to minimize the system resources for supporting full VCR functionality in a multicast VoD system. This scheme also allows multicast systems to support true VoD services so customers can be served as soon as the system receives the requests. Computer simulations show that the multicast systems using the SRMDRU scheme perform much better than other multicast systems in terms of system blocking probabilities.

Performance Study of Large-Scale Video Streaming Services in Highly Heterogeneous Environment

To support large-scale Video-on-Demand (VoD) services in a heterogeneous network environment, either a replication or layering approach can be deployed to adapt the client bandwidth requirements. With the aid of the broadcasting and caching techniques, it has been proved that the overall performance of the system can be enhanced. In this paper, we explore the impact on the broadcasting schemes coupled with proxy caching and develop an analytical model to evaluate the system performance in a highly heterogeneous network environment. We develop guidelines for resources allocation, transmission strategies as well as caching schemes under different system configurations. The model can assist system designers to study various design options as well as perform system dimensioning. Moreover, a systematic comparison between replication and layering is performed. From the results, it can be seen that the system performance of layering is better than that of replication when the environment is highly heterogeneous even if the layering overhead is higher than 25%. In addition, it is found that the system blocking probability can be further reduced by exploring the broadcast capability of the network if the proxy server cannot store all the popular videos.

Determination of efficient transmission scheme for video-on-demand (VoD) services

In providing video-on-demand services in multicast environment, a broadcasting technique like skyscraper is very efficient for serving a very popular video in terms of resource requirement. However, a multicasting method like patching will achieve better performance if the video is not popular enough. Since there may not be a clear distinction between very popular and popular videos, an analytical model is developed in this paper to select an appropriate delivery scheme for the videos according to both the popularity of videos and the customer reneging behavior. The result shows that not only is the overall bandwidth of the system minimized, but a preset quality of service is also provided for each customer when using our proposed scheme.

Provision of continuous VCR functions in interactive broadcast VoD systems

Video broadcasting is one of the feasible solutions to implement a large-scale video-on-demand (VoD) system. Nevertheless, it is still an open issue for the provision of continuous VCR functions in a delay insensitive broadcast VoD system. In this paper, we propose to jointly optimize an active buffer management scheme with contingency channels to support the VCR functions in an efficient protocol called partitioned broadcasting. We develop a greedy channel management scheme by exploiting the property of the broadcasting protocol such that the system bandwidth capacity can be fully utilized. Incorporating the channel management scheme with the partitioned video broadcast, the VoD system not only provides delay insensitive services but also handles all the interactive requests. Extensive simulation results demonstrate that the partitioned broadcasting system outperforms the traditional system based on the staggered broadcasting protocols. It is found that 20 broadcasting channels and 10 contingency channels are sufficient to support on average 720 customers for a single video with less that one second start-up delay and all types of VCR functions.

Video-on-Demand Systems With Cooperative Clients in Multicast Environment

Peer-to-peer (P2P) and multicast approaches are two common transmission strategies to provide scalable and cost efficient solutions for video-on-demand (VoD) services. The former approach requires small server resources and provides a negligible delay to users, but the bandwidth requirement inside the network will be rapidly increased when more customers join the video session. On the other hand, if the system simply uses a multicast scheme to deliver a video, customers will experience a noticeable delay before watching the video whereas the overall bandwidth requirement will not be significantly increased. In this paper, a new transmission policy denoted peer-to-peer batching (PPB) policy is proposed to efficiently deliver video data in a large scale VoD system by exploiting the multicast capability of the network and P2P paradigm. To avoid the disruption of services, the fault tolerance and recovery mechanism is also developed for PPB. In addition, in order to further improve the scalability of the system, a distributed PPB (DPPP) is also proposed. We then develop a mathematical model to evaluate the performance of each of our proposed policies analytically. Based on this model, system designers are able to study various design options as well as perform system dimensioning. Both analytical and simulation results show that the proposed scheme achieves better system performance compared with existing schemes.

A hybrid delivery strategy for a video-on-demand system with customer reneging behavior

A new delivery strategy is proposed for a video-on-demand system to allocate transmission resources according to the popularity of videos. A video is delivered to customers through one of three channels, namely broadcast, multicast and unicast, depending on whether the video is very hot, hot or cold, respectively. An algorithm is developed to select the appropriate delivery channel for the videos so that not only the overall bandwidth of the system is minimized but also a preset quality of service is provided for each customer. An analytical model is also developed to evaluate the bandwidth requirement of the whole system with VCR functionality. The result shows that a significant bandwidth reduction can be achieved when using our proposed delivery strategy.

Peer-to-Peer Batching Policy for Video-on-Demand System

In this paper, an enhanced peer-to-peer batching policy (EPPB) is developed for providing true video-on-demand (VoD) services in multicast environment. It is accomplished by exploring the multicast delivery coupled with peer-to-peer (P2P) transmission strategy for video streaming such that gets benefit from both transmission schemes to efficiently deliver video data to the clients. The results show that this enhanced version performs better than the original version in terms of total bandwidth requirement up to 30%

Investigating the Performance of Hierarchical Video-on-Demand System in Heterogeneous Environment

To support large-scale video-on-demand (VoD) services, many transmission schemes such as data broadcasting and proxy caching have been developed to minimize the resources requirement and enhance the system scalability. However, most of the previous works mainly focused on providing the VoD services in a homogeneous environment. In this paper, we investigate a feasible solution for building an efficient hierarchical VoD system using proxy caching coupled with broadcasting and appropriate coding schemes in heterogeneous environment so as to minimize the system transmission cost. By developing a performance model in this paper, system designers can study how various techniques for VoD can be organized to maximize the efficiency of the system. From the results, it can be seen that the system performance can be greatly enhanced by efficiently coupling various techniques.

A hierarchical video-on-demand system with double-rate batching

Abstract In this paper, a two-tier hierarchical network architecture is exploited to provide true video-on-demand (VoD) services using double-rate (DR) batching in multicast environment. In the proposed system, new customers will be grouped together by the local server with the DR batching scheme and then be served by the multicast backbone network. Two scheduling policies are developed to serve customers at the local batching servers. The results show that the hierarchical VoD system can greatly reduce the resources requirement compared with the centralized approach. In addition, in order to secure the multicast channels for the new multicast groups, the backbone channel reservation algorithm that is similar to the look-ahead scheme [Multimedia Systems 3 (1995) 137] is proposed. If there are 100 videos and 500 backbone multicast channels, the system is able to support 0.2 requests/s with less than 7% admission blocking probability.