Controlled Buffer Sharing in Continuous Media Servers

Multimedia Systems, 2000

Excessive buffer requirement to handle continuous-media playbacks is an impediment to costeffective provisioning for on-line video retrieval. Given the skewed distribution of video popularity, it is expected that often there are concurrent playbacks of the same video file within a short time interval. This creates an opportunity to batch multiple requests and to service them with a single stream from the disk without violating the on-demand constraint. However, there is a need to keep data in memory between successive uses to do this. This leads to a buffer space trade-off between servicing a request in memory mode vs. servicing it in disk-mode. In this work, we develop a novel algorithm to minimize the buffer requirement to support a set of concurrent playbacks. One of the beauties of the proposed scheme is that it enables the server to dynamically adapt to the changing workload while minimizing the total buffer space requirement. Our algorithm makes a significant contribution in decreasing the total buffer requirement, especially when the user access pattern is biased in favor of a small set of files. The idea of the proposed scheme is modeled in detail using an analytical formulation, and optimality of the algorithm is proved. An analytical framework is developed so that the proposed scheme can be used in combination with various existing disk-scheduling strategies. Our simulation results confirm that under certain circumstances, it is much more resource efficient to support some of the playbacks in memory mode and subsequently the proposed scheme enables the server to minimize the overall buffer space requirement.

Information Systems, 1995

Buffer management in continuous-media systems is a frequently studied topic. One of the most interesting recent proposals is the idea of buffer sharing for concurrent streams. As analyzed in [6], by taking advantage of the temporal behaviour of concurrent streams, buffer sharing can lead to a 50% savings in total buffer space. In this paper, we study how to actually implement buffer sharing. To this end, we develop the CES Buffer Sharing scheme that is very efficient to implement, and that permits savings asymptotically very close to the ideal savings predicted by the analysis in (61. We show that the CES scheme can operate effectively under varying degrees of disk utilizations, and during transition periods when the number of concurrent streams changes. We also demonstrate how the scheme can be further improved, particularly for situations when the number of concurrent streams is small. In ongoing work, we will integrate the proposed scheme into a distributed continuous-media file system which is under development at the University of British Columbia.

Applied mathematical modelling, 2009

The video communication requires considerable large bandwidth - a scarce resource and thus it needs be used efficiently. It can be traded for buffer storage, ie for maintaining more buffer storage the data downloading rate (bandwidth) can be reduced. In order to transmitting the ...

Multimedia Systems, 1996

In this paper, we study the problem of how to maximize the throughput of a continuous-media system, given a xed amount of bu er space and disk bandwidth both pre-determined at design-time. Our approach is to maximize the utilizations of disk and bu ers. We propose doing so in two w ays. First, we analyze a scheme that allows multiple streams to share bu ers. Our analysis and preliminary simulation results indicate that bu er sharing could lead to as much as 50% reduction in total bu er requirement. Second, we d e v elop three prefetching strategies: SP, IP1 and IP2. As will be demonstrated by S P , straightforward prefetching is not e ective at all. In contrast, IP1 and IP2, which prefetch m o r e i n telligent l y t h a n d o e s S P , could be valuable in maximizing the e ective use of bu ers and disk. Our preliminary simulation results show that IP1 and IP2 could lead to a 40% improvement in throughput. keywords: bu er allocation, prefetching strategies, analysis of continuous-media or multimedia systems

Parallel Computing, 1998

During the past decade, the information technology has evolved to store and retrieve continuous media data types, e.g., audio and video clips. Unlike the traditional data types (e.g., text), continuous media consists of a sequence of quanta, either audio samples or video frames, that convey meaning when presented at a pre-specified rate. A challenging task when implementing these servers is to guarantee retrieval and delivery of a clip at its pre-specified rate. Continuous media servers are expected to play a major role in many application domains including library information systems, entertainment technology, educational applications, etc. The focus of this study is on the disk subsystem of a server and techniques that both enhance its performance and ensure its scalability. By scalable, we mean that the system can grow incrementally as the requirements of an application grows in order to avoid a degradation in performance. This study details data placement techniques across both (1) the zones of a disk in order to trade bandwidth for storage, and (2) multiple disks to enable the system to scale. In addition, we describe scheduling techniques to (I) trade memory for disk bandwidth and vice versa, and (2) merge multiple requests referencing a single clip in order to increase the number of simultaneous displays supported by the system. We present performance results from Mitra, an experimental prototype that realizes an implementation of the presented techniques. These results demonstrate the feasibility of a scalable server. 0 1998 Elsevier Science B.V.

There has been an increasing trend in designing the Video-on-Demand systems using parallel server-based architectures to meet the client demands for online streaming of videos. We use a distributed architecture for effective Buffer management which efficiently utilizes various server resources like the disk Bandwidth, memory utilization and provides clients with VCR like functionalities. In this paper we proposed a unique On-line dynamic buffer replacement policy that does not base its results solely on a single parameter such as video popularity but calculates a cumulative popularity based on several other factors and hence chooses the video with the least popularity thus calculated. It is necessary to batch requests to minimize the bandwidth requirement, to reduce I/O demand, improve throughput and increase the number of customers served by using less number of streams. We proposed a Modified Adaptive Batching Policy for batching requests arriving at a server. This is different fr...

Proceedings 23rd Euromicro Conference New Frontiers of Information Technology - Short Contributions -, 1997

Memory management is a key issue when designing cost-effective video-on-demand servers. State-of-the-art techniques, like double-buffering, allocate buffers in a per-stream basis and require huge amounts of memory. In this paper, we propose a buffering policy, namely Single Pair of Buffers, that reduces dramatically server memory requirements by reserving a pair of buffers per storage device. By considering in detail disk and network interaction we have also identified the particular conditions under which this policy can be successfully applied to engineer Video-On-Demand servers. Reduction factors of two orders of magnitude compared to the double-buffering approach can be obtained. Current disk and network parameters make this technique feasible.

Telecommunication …, 2009

In this contribution, we investigate the performance of the output buffer of an 'ondemand' video streaming server. The server maintains a local database of stored video clips and movies which can be streamed to the users upon request. We assume that the stored video is encoded in a scalable way, which means that the data streams contain a base layer ensuring a minimum of guaranteed quality and a stack of additional enhancement layers progressively improving the quality of the video. For the purpose of performance analysis, we assume that a video stream is split up in logical units called frames. Every frame consists of a number of packets, each containing information of one layer only. When the output buffer gets congested, one may choose to drop the transmission of some of the layers in a frame, thus reducing the frame transmission time and expediting the restoration of the buffer size to normal levels. A discrete-time finite capacity queueing model with buffer size dependent transmission times is proposed. Using a probability generating function approach, we focus on the characteristics of idle and busy periods. We obtain performance measures such as the frame loss ratio and the average frame transmission time. The latter measure relates to the quality of the video stream. We conclude with some numerical examples, including a realistic case study.

Proceedings of the 2012 ACM workshop on Capacity sharing - CSWS '12, 2012

Today, content replication methods are common ways of reducing the network and servers load. Present content replication solutions have different problems, including the need for pre-planning and management, and they are ineffective in case of sudden traffic spikes. In spite of these problems, content replication methods are more popular today than ever, simply because of an increasing need for load reduction. In this paper, we propose a shared buffering model that, unlike current proxy-based content replication methods, is native to the network and can be used to alleviate the stress of sudden traffic spikes on servers and the network. We outline the characteristics of a new transport protocol that uses the shared buffers to offload the server work to the network or reduce the pressure on the overloaded links.

2003

Currently, for the display of synthetic and archived multimedia movies and clips, buffers are managed at the server end (push mode), at the client end (pull mode), or in the push-pull mode to avoid excessive server side disk access time, avoid hot spots for multiple disk accesses for the same media objects, and/or to smoothen the effect of variations in transmission delay. However, the current schemes suffer from the lack of static analysis and the coordination between server and the clients to reduce jitters. In this paper, we describe a server coordinated predictive buffer management scheme for mobile devices that integrates static analysis of frame based synthetic movies and server directed client side management to reduce the retransmission of reusable multimedia objects over the Internet. Performance evaluation shows that the scheme reduces the jitter significantly while maintaining the needed QoS, and the scheme outperforms popular rendering formats.