H. Vin, A. Goyal, and P. Goyal. Algorithms for designing large-scale multimedia servers. Computer Communications, 18(3):192--203, March 1995.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....only if the resource available at the server are sufficient to meet its resource requirements. Admission control algorithms can provide either deterministic or statistical guarantees, depending on whether they reserve resources based on the worst case load or a probability distribution of the load [43, 42]. Regardless of the nature of guarantees provided by admission control algorithms, designing such algorithms for the server push architecture is simple the sequential and periodic nature of data retrieval enables the server to accurately predict the data rate requirements of each client and ....

H. M. Vin, A. Goyal, and P. Goyal. Algorithms for Designing Large-Scale Multimedia Servers. Computer Communications, 18(3):192--203, March 1995.

....Related Work Most approaches to admission control consider the requests of single media streams. The resource requirements are prespecified by the media request in terms of constant rate or little rate deviations [RVT96] and calculated by stochastic [NMW97] VGGG94b] or deterministic approaches [VGG95] ORSN96] Most concepts providing stochastic service guarantees assume stochastic retrieval time from storage systems which we do not consider. For example, VGGG94b] exploit the variation in access times from disk. In the following, we analyse strategies that support interactions and evaluate ....

Harrick M. Vin, Alok Goyal, and Pawan Goyal. Algorithms for designing large-scale multimedia servers. Computer Communications, March 1995.

....to admission control consider the requests of single media streams. The resource requirements are prespecified by the media request in terms of constant rate or little rate deviations [16] The available system resources are calculated by stochastic [13] 24] or deterministic approaches [23], 14] Based on the knowledge about the already reserved and freely available resources, it is possible to reject requests in case of server overloads. Most concepts providing stochastical service guarantees assume stochastical retrieval time from storage system whichwe do not consider. For ....

Harrick M. Vin, Alok Goyal, and Pawan Goyal. Algorithms for designing large-scale multimedia servers. Computer Communications,March 1995.

....to the client. Admission control mechanisms for pure server push architectures assume that during a presentation the data rate consumed by the client will be nearly constant [RVT96] The available system resources are calculated by stochastic [NMW97] VGGG94] or deterministic approaches [VGG95] ORSN96] Based on the knowledge about the already reserved and freely available resources, it is possible to reject requests in case of server overloads. In the following, we discuss some of the strategies that consider interactive applications. A priori reservation. To guarantee a given QoS ....

Harrick M. Vin, Alok Goyal, and Pawan Goyal. Algorithms for designing largescale multimedia servers. Computer Communications, March 1995.

....that a server can support simultaneously. Streams may have different bandwidth reservations due to the difference in the desired QoS or bandwidth estimation. The server utilization of the deterministic admission control algorithms is much lower than the predictive or the statistical algorithms [14]. As a result, a server cannot support more streams in the deterministic case compared to the predictive or the statistical schemes. This is because of the fact that the total number of streams that the server can support simultaneously is proportional to the server utilization. The difference ....

H. M. Vin, A. Goyal, and P. Goyal, "Algorithms for Designing Large-Scale Multimedia Servers," Computer Communications, vol. 18, pp. 192--203, March 1995.

....i.e. response time and probability of missing deadlines. Before proceeding, we briefly survey related works and highlight the main contributions of this paper, where appropriate. There is a multitude of work on the design of continuous media servers, some of which (and by no means all) include [1, 7, 14, 18, 3, 4, 19], where the authors mostly focus on data layout and retrieval and delivery techniques which facilitate maintaining of continuity in data delivery while providing either deterministic or statistical QoS guarantees. Scheduling of mixed workloads has not received as much attention. Although it was ....

H.M. Vin, A. Goyal, and P. Goyal. Algorithms for Designing Large-Scale Multimedia Servers. In Computer Communciations, March 1995.

....8.34 msec Figure 1: The characteristics of Seagate Barracuda 2 disk. In order to increase performance, schemes for reducing the impact of latency, as well as solutions for increasing bandwidth must be devised. Clever storage allocation schemes [2, 12, 3] as well as novel disk scheduling schemes [1, 11, 7, 13, 9] must be devised to reduce or totally eliminate latency so that buffering requirements can be reduced while bandwidth is utilized effectively. Storage techniques based on multiple disks such as replication and striping must be employed to increase the bandwidth. In this paper, we first present a ....

....seconds to approximately 26 seconds. In this system, the worst case delay is 26 seconds and the number of frames skipped is 780 (26 seconds of video at 30 frames sec) 4 Related Work A number of storage schemes for continuous retrieval of video and audio data have been proposed in the literature [5, 11, 12, 7, 3, 8, 13, 9]. Among these, 5, 11, 7, 13, 9] address the problem of satisfying multiple concurrent requests for the retrieval of multimedia objects residing on a disk. These schemes are similar in spirit to the simple scheme that we presented in Section 2. In each of the schemes, concurrent requests are ....

[Article contains additional citation context not shown here]

A. Goyal, H. M. Vin and P. Goyal. Algorithms for designing large-scale multimedia servers. Computer Communication, 1994.

.... very active research area [OBRS94, GarOS98, GhKS95] Multimedia storage servers deal with [GVKRR95] file management [AOG92] disk management [RangV93, CheL96, KwonCS97] I O scheduling [RedW93, CheKY93, GemmH94, KenS97, WolfYS97] buffer management [NgY96] network management, resource scheduling [VinGG95, DanSS96, GarOS98], admission control and QOS guarantees [VinGGG94] Anderson [And91] defines the problem as a metascheduling problem of continuous media resources. CHIMP [CandPS96] is a multimedia authoring and presentation system that deals with authoring, retrieval, and scheduling of presentations. Nerjes et al. ....

Vin, H., Goyal, A., Goyal, P., "Algorithms for Designing Large-Scale Multimedia Servers", Computer Communications, March 1995.

No context found.

H. M. Vin, A. Goyal, and P. Goyal. Algorithms for Designing Large-Scale Multimedia Servers. Computer Communications, 18(3):192--203, March 1995.

....deterministic and statistical admission control algorithms is that the latter leads to better utilization of resources than the former at the expense of weaker guarantees. Due to space constraints, we do not discuss admission control algorithms in this paper. A detailed discussion may be found in [22, 31, 32]. 6 The Data Type Specific Layer The data type specific layer uses the mechanisms provided by the data type independent layer to implement data type specific policies. The data type specific layer consists of a set of modules, each of which implements policies optimized for a particular data ....

H. M. Vin, A. Goyal, and P. Goyal. Algorithms for Designing Large-Scale Multimedia Servers. Computer Communications, 18(3):192-- 203, March 1995.

....tolerate infrequent losses [3] ffl Measurement based admission control algorithms use past measurements of bit rate and disk access times of media streams as an indicator of future resource requirements. They achieve the highest disk utilization at the expense of providing the weakest guarantees [2]. These admission control algorithms span an entire spectrum and achieve varying server utilization while providing different levels of guarantees. In environments with varying user requests, a multimedia server must support some or all of the above admission control algorithms to service its ....

H. M. Vin, A. Goyal, and P. Goyal. Algorithms for Designing Large-Scale Multimedia Servers. Computer Communications, 18(3):192--203, March 1995.

No context found.

H. M. Vin, A. Goyal, and P. Goyal, "Algorithms for designing large-scale multimedia servers," Computer Communications 18, pp. 192--203, March 1995.

....playback, the number of media units accessed for each stream during a round must be large enough to meet their respective playback rate requirement. Furthermore, the service time (i.e. the total time spent in retrieving media units during a round) should not exceed the duration of the round [10, 11]. To effectively utilize a disk array, and hence to maximize the number of clients that can be serviced simultaneously, a multimedia server must interleave the storage of each media stream among the disks in the array. The unit of data interleaving, referred to as a media bock, denotes the maximum ....

....the quality of service being provided to clients) yielded by such read ahead policies are at the expense of increase in buffer space requirement. However, by judiciously choosing the set of blocks to be read ahead in underflow rounds, the increase in the buffer space requirement can be minimized [11]. 3 Variable Size Block Placement Policy Consider a multimedia server that interleaves the storage of video streams on a disk array in terms of variable size blocks, each containing fixed number of frames. Hence, retrieving f i frames for each client during a round will require the server to ....

H. M. Vin, A. Goyal, and P. Goyal. Algorithms for Designing Large-Scale Multimedia Servers. To appear in Computer Communications, March 1995.

.... on disk in terms of media blocks (each containing several media units) the server can: 1) exploit the sequentiality of video playback to determine the set of media blocks to be accessed in a round; 2) batch all of these requests; and (3) employ disk scheduling algorithms (e.g. SCAN, greedy [9]) to minimize the seek time and rotational latency incurred during the retrieval. Determination of the round duration R is governed by the following tradeoffs. Increasing the duration of the round increases the number of media units, and hence the number of media blocks, accessed during a round. ....

H. M.Vin,A. Goyal, and P. Goyal. Algorithms for Designing Large-Scale Multimedia Servers. Computer Communications, 18(3):192--203, March 1995.

No context found.

H. Vin, A. Goyal, and P. Goyal. Algorithms for designing large-scale multimedia servers. Computer Communications, 18(3):192--203, March 1995.

No context found.

H. Vin, A. Goyal, and P. Goyal. Algorithms for designing large-scale multimedia servers. Computer Communications, 18(3):192--203, March 1995.

No context found.

H. M. Vin, A. Goyal and P. Goyal, "Algorithms for Designing Large-Scale Multimedia Servers", Computer Communications, Vol. 18, No. 3, Pages 192-203, March 1995

No context found.

Harrick M. Vin, Alok Goyal, Pawan Goyal, Algorithms for Designing Large-Scale Multimedia Servers, Computer Communications, March 1995.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC