A number of techniques have been developed for maximizing disk utilization in media servers, including disk arm scheduling and data placement ones. Instead, in this paper we focus on how to efficiently utilize the available memory. We present techniques for best memory use under different disk policies, and derive precise formulas for computing memory use. We show that with proper memory use, maximizing disk utilization does not necessarily lead to optimal throughput. In addition, we study the impact of data placement policies including disk partitioning and multiple disks. Finally, our analysis shows that maximizing disk utilization and disk striping incur high system costs, and are not advisable in a media server.

The performance and capacity of commodity computer systems have improved drastically in recent years. However, these systems still lack the support for real-time data access, which is required by an increasing number of emerging applications. In this paper we first present several important storagebound real-time applications and classify their Quality of Service (QoS) requirements. We then survey the representative work on disk management in the areas of IO scheduling, admission control, and data placement. Finally, we present our approach for providing disk QoS in commodity systems and present key empirical results from the micro-benchmark-based evaluation of our QoS-enhanced Linux kernel.

Allowing higher-priority requests to preempt ongoing disk IOs is of particular benefit to delay-sensitive multimedia and real-time systems. In this paper we propose Semi-preemptible IO, which divides an IO request into small temporal units of disk commands to enable preemptible disk access. We present main design strategies to allow preemption of each component of a disk access---seek, rotation, and data transfer. We analyze the performance and describe implementation challenges. Our evaluation shows that Semi-preemptible IO can substantially reduce IO waiting time with little loss in disk throughput. For example, expected waiting time for disk IOs in a video streaming system is reduced 2.1 times with the throughput loss of less than 6 percent.

In video-on-demand (VOD) systems, as the size of the buffer allocated to user requests increases, initial latency and memory requirements increase. Hence, the buffer size must be minimized. The existing static buffer allocation scheme, however, determines the buffer size based on the assumption that the system is in the fully loaded state. Thus, when the system is in a partially loaded state, the scheme allocates a buffer larger than necessary to a user request. This paper proposes a dynamic buffer allocation scheme that allocates to user requests buffers of the minimum size in a partially loaded state as well as in the fully loaded state. The inherent difficulty in determining the buffer size in the dynamic buffer allocation scheme is that the size of the buffer currently being allocated is dependent on the number of and the sizes of the buffers to be allocated in the next service period. We solve this problem by the predict-and-enforce strategy, where we predict the number and the sizes of future buers based on inertia assumptions and enforce these assumptions at runtime. Any violation of these assumptions is resolved by deferring service to the violating new user request until the assumptions are satisfied. Since the size of the current buffer is dependent on the sizes of the future buffers, the size is represented by a recurrence equation. We provide a solution to this equation, which can be computed at the system initialization time for runtime efficiency. We have performed extensive analysis and simulation. The results show that the dynamic buffer allocation scheme reduces initial latency (averaged over the number of user requests in service from one to the maximum capacity) to

Conventional wisdom holds that reducing disk latency leads to higher disk utilization, maximizing disk utilization leads to higher throughput, and employing a faster disk leads to better performance. All of this is true when building a conventional file or database system. In this paper we show that these principles can be misleading when applied to a media server. To design such a server, we propose a cost-based approach that focuses on the perstream costs. We give various examples to illustrate the design process. Keywords: multimedia, disk latency, memory utilization, per-stream cost. 1 Introduction Maximizing throughput is a common design objective for a media server. To improve throughput, two approaches have been used: reducing disk latency (i.e., seek overhead and rotational delay) and minimizing the required memory. To reduce disk latency we can either employ efficient disk scheduling [13, 14, 15] or enact intelligent data placement policies [9, 11]. Both methods effectively ...

In this study we present a scheme called two-dimensional BubbleUp (2DB) for managing parallel disks in a multimedia server. Its goal is to reduce initial latency for interactive multimedia applications, while balancing disk loads to maintain high throughput. The 2DB scheme consists of a data placement and a request scheduling policy. The data placement policy replicates frequently accessed data and places them cyclically throughout the disks. The request scheduling policy attempts to maintain free "service slots" in the immediate future. These slots can then be used to quickly service newly arrived requests. Through examples and simulation, we show that our scheme significantly reduces initial latency and maintains throughput comparable to that of the traditional schemes. Keywords: multimedia, data replication, initial latency, disk array. 1 Introduction Media servers are designed to provide large numbers of presentations in the form of audios, movies or news clips. These servers need...

Supporting preemptible disk access is essential for interactive multimedia applications that require short response time. In this study, we propose Virtual IO, an abstraction for disk IO, that transforms a nonpreemptible IO request into a preemptible one. In order to achieve its objective efficiently, Virtual IO uses disk profiling to obtain accurate and detailed knowledge about the disk. Upon implementation of Virtual IO, we show that not only does Virtual IO enable highly preemptible disk access, but it does so with little or no loss in disk throughput.

Conventional wisdom holds that reducing disk latency leads to higher disk utilization, maximizing disk utilization leads to higher throughput, employing a faster disk leads to better performance. All of this is true when building a conventional file or database system. In this paper we show that these principles can be misleading when applied to the design a media server. We examine a number of techniques that have been developed for maximizing disk bandwidth utilization in media servers, including disk arm scheduling and data placement ones. We show that, some disk latency reduction techniques can be counterproductive since memory, rather than disk, is the bottleneck in improving throughput. We present techniques for best memory use under different disk policies, and derive formulas for computing memory use. In addition, to account for the design tradeoffs between disk bandwidth and memory use, we propose a cost-based approach that focuses on minimizing the per-stream costs, including...

In this paper we do focus on the client side, presenting a combined memory-disk buffering algorithm that allows the client to dynamically and effectively deal with variable data rates and delays. We call this algorithm MEDIC, for MEmory-DIsk Cache. MEDIC manages the resources of the client's computer or set-top box. MEDIC carefully allocates its limited memory to competing tasks, i.e., to receiving new data from the network, to writing data to disk as memory fills up, to reading data from disk as needed, and to holding data for decoding and playback. Since data is concurrently written to the disk cache and read from the disk cache, MEDIC must also intelligently issue IOs to avoid undue conflicts.

Abstract—Allowing higher-priority requests to preempt ongoing disk IOs is of particular benefit to delay-sensitive and real-time systems. In this paper, we present Semi-preemptible IO, which divides disk IO requests into small temporal units of disk commands to improve the preemptibility of disk access. We first lay out main design strategies to allow preemption of each component of a disk access—seek, rotation, and data transfer, namely, seek-splitting, JIT-seek, and chunking. We then present the preemption mechanisms for single and multidisk systems—JIT-preemption and JIT-migration. The evaluation of our prototype system showed that Semi-preemptible IO substantially improved the preemptibility of disk access with little loss in disk throughput and that preemptive disk scheduling could improve the response time for high-priority interactive requests. Index Terms—Storage, preemptible disk access, preemptive disk scheduling, real-time, QoS, disk IO preemption. 1