Jin B. Kwon

Providing VCR functionality in staggered video broadcasting

A true video-on-demand (TVoD) system allows users to view any video program, at any time, and perform any VCR function, but its per-user video delivery cost is very expensive. A near video-on-demand (NVoD) system uses a more scalable approach by batching multiple clients to a shared stream of broadcasting videos. Staggered video broadcasting, one of the NVoD techniques, broadcasts multiple streams of the same video at staggered times, with one stream serving multiple clients. In order to provide subscribers with a high-quality VoD service, it is desirable to add VCR functionality, such as fast forward and fast rewind, but it is not easy to provide VCR functionality in NVoD, especially a video broadcasting system where no dedicated or interactive channel is available. We analyze the conditions necessary to provide VCR functions and then propose a reception schedule that satisfies these conditions, with minimal resource requirements. Since our proposed scheme receives video frames as units, it can rapidly sustain a changing VCR action pattern. It is demonstrated, by simulation, that the scheme provides VCR functionality consistently with minimal buffer space.

Distributed multimedia streaming over peer-to-peer networks

A peer-to-peer model is very useful in solving the server link bottleneck problem of a client-server model. In this work, we discuss the problems of distributing multimedia content over peer-to-peer network. We focus on two problems in peer-to-peer media content distribution systems. The first is the transmission scheduling of the media data for a multi-source streaming session. We present a sophisticated scheduling scheme, which results in minimum buffering delay. The second problem is on the fast distribution of media content in the peer-to-peer system that is self-growing. We propose a mechanism accelerating the speed at which the system’s streaming capacity increases.

Exploiting caching for realtime multimedia systems

Several caching schemes for realtime multimedia systems have been proposed but they only concentrate on the hit ratio as the performance metric, without providing any means to utilize the disk bandwidth gained through cache hits. A high cache hit ratio alone does not capture the effectiveness of a caching scheme for such systems unless hiccup-free service is guaranteed. We evaluate three caching schemes which make it possible to provide service to additional streams with the saved disk bandwidth. We first start with NIC in which no cached intervals are preempted or replaced unless they have finished their playback. Next, we propose PSIC, which preempts the cached intervals for more valuable intervals but still guarantees hiccup-free playback. Finally, we present SPT-L scheme which has a significantly greater overhead and the best hit ratio. However STP-L suffers from frequent starvation, which verifies our claim that the high cache hit ratio does not necessary result in better performance.

An admission control scheme for continuous media servers using caching

In continuous media servers, disk load can be reduced by using buffer cache. In order to utilize the saved disk bandwidth by caching, a continuous media server must employ an admission control scheme to decide whether a new client can be admitted for service without violating the requirements of clients already being serviced. Since deterministic admission control is based on the worst case assumption, it causes the wastage of the system resources. In this paper, we propose a statistical admission control scheme for a continuous media servers where caching is used to reduce disk load. This scheme improves disk utilization and allow more streams to be serviced with reasonable computational overhead, while maintaining near-deterministic service. The scheme, called short-sighted prediction admission control (SPAC), combines exact prediction through on-line simulation and statistical estimation using a probabilistic model of future disk load in order to reduce computation overhead.

Generalized data retrieval for pyramid-based periodic broadcasting of videos

A true video-on-demand system allows users to view any video program, at any time, and perform any VCR functions, but its per-user video delivery cost is too expensive to have commercial use. Periodic broadcasting (PB), which is a near video-on-demand technique, broadcasts videos repeatedly over broadcast channels. In this way, PB can service an unlimited number of clients simultaneously with a bounded service latency. We propose a data retrieval scheme, consisting of buffer management and data placement, for PB servers. Unlike existing schemes devised for a specific PB technique, our scheme can be adopted by general PB. Furthermore, it is devised considering variations in disk load induced by VBR-encoded videos.

VCR-oriented video broadcasting for near video-on-demand services

A true video-on-demand (TVOD) system lets users view any video program, at any time, and perform any VCR functions, but its per-user video delivery cost is too expensive. A near video-on-demand (NVOD) is a more scalable approach by batching multiple clients to a shared stream or broadcasting videos. In order to provide subscribers with high-quality VOD service, it is desirable to add VCR functionality such as fast forward, fast backward and so on, but the existing video broadcasting, protocols have a limitation on providing VCR functionality. In this paper, we present the conditions necessary to provide VCR functions and then propose a video broadcasting scheme suitable for providing them, which satisfies these conditions, with minimal resource requirements. Since our proposed scheme receives video frames as a unit it can keep up rapidly with a changing VCR action pattern. It is demonstrated that the scheme provides VCR functionality consistently with minimal buffer space through simulations.

Providing VCR Functionality in Staggered Video Broadcasting

A true video-on-demand(TVOD) system lets users view any video program, at any time, and perform any VCR functions, but its per-user video delivery cost is too expensive. A near video-ondemand( NVOD) is a more scalable approach by batching multiple clients to a shared stream or broadcasting videos. Staggered video broadcasting, one of NVOD techniques, broadcasts multiple streams of the same video at staggered times, with one stream serving multiple clients. In order to provide subscribers with a high-quality VOD service, it is desirable to add VCR functionality such as fast forward and fast backward, but it is not easy to provide VCR functionality in NVOD, especially video broadcasting system where any dedicated or interaction channel is not available.In this paper, we analyze the conditions necessary to provide VCR functions and then propose a reception schedule which satisfies these conditions, with minimal resource requirements. Since our proposed scheme receives video frames as a unit it can keep up rapidly with a changing VCR action pattern. It is demonstrated that the scheme provide VCR functionality consistently through simulations.

Adaptive channel management scheme for video broadcasting services

Traditional unicast VOD systems require huge amounts of the server networks bandwidth, so near-VOD systems using broadcast protocols have been proposed to reduce the bandwidth requirement. Such systems have not been able to allocate channels dynamically to adapt to environments where the popularity distribution of videos changes, because of the essential inflexibility of existing broadcasting protocols. We address dynamic channel allocation for video broadcasting services, and propose a channel management technique that can adapt well to the dynamic environment, using a flexible broadcasting scheme presented in our previous work (2005). Our channel management scheme decides how many channels should be allocated to each video to maximize the system throughput, based on the videos popularity and reneging probability. The proposed scheme has been verified using extensive simulation.

Adjustable broadcast protocol for large-scale near-video-on-demand systems

Video-on-demand (VOD) is a service allowing customers to select video programs from a central server for viewing on a television or a computer screen. Traditional unicast VOD systems require huge amounts of the server networks bandwidth, so near-VOD systems using broadcast protocols have been proposed to reduce the bandwidth requirement. Although NVOD systems can service many users with less network bandwidth by broadcasting popular videos, the popularity of any given video is likely to vary widely as time goes on. However, the broadcast protocols cannot adapt to dynamic environments where the popularity distribution of videos changes, because of their fixed broadcasting schedule. We introduce a dynamic broadcast protocol called Adjustable Broadcast (AB) based on on-line scheduling, which can adjust the number of channels being used on the fly. AB also exploits various prefetching strategies in the broadcast schedule to reduce the required bandwidth. And, we propose a dynamic channel allocation scheme using AB that minimizes the average viewers waiting time. Using extensive simulation, we have demonstrated the effectiveness of AB and the adaptability of our dynamic channel allocation.

A Statistical Admission Control Scheme for Continuous Media Servers Using Caching

In continuous media servers, disk load can be reduced by using buffer cache. In order to utilize the saved disk bandwidth by caching, a continuous media server must employ an admission control scheme to decide whether a new client can be admitted for service without violating the requirements of clients already being serviced. A scheme providing deterministic QoS guarantees in servers using caching has already been proposed. Since, however, deterministic admission control is based on the worst case assumption, it causes the wastage of the system resources. If we can exactly predict the future available disk bandwidth, both high disk utilization and hiccup-free service are achievable. However, as the caching effect is not analytically determined, it is difficult to predict the disk load without substantial computation overhead. In this paper, we propose a statistical admission control scheme for continuous media servers where caching is used to reduce disk load. This scheme improves disk utilization and allows more streams to be serviced while maintaining near-deterministic service. The scheme, called Shortsighted Prediction Admission Control (SPAC), combines exact prediction through on-line simulation and statistical estimation using a probabilistic model of future disk load in order to reduce computation overhead. It thereby exploits the variation in disk load induced by VBR-encoded objects and the decrease in client load by caching. Through trace-driven simulations, it is demonstrated that the scheme provides near-deterministic QoS and keeps disk utilization high.