System architecture for a large scale video on demand service
Computer Networks and ISDN Systems - Driving applications for future networks
Optimization of the grouped sweeping scheduling (GSS) with heterogeneous multimedia streams
MULTIMEDIA '93 Proceedings of the first ACM international conference on Multimedia
I/O issues in a multimedia system
Computer
Issues in storage and retrieval of multimedia data
Multimedia Systems - Special issue on multimedia database systems
Scheduling Algorithms for Multiprogramming in a Hard-Real-Time Environment
Journal of the ACM (JACM)
Making a Cost-Effective Video Server
IEEE MultiMedia
A fast algorithm for single processor scheduling
SFCS '78 Proceedings of the 19th Annual Symposium on Foundations of Computer Science
Designing an on-demand multimedia service
IEEE Communications Magazine
IEEE Communications Magazine
Performance model of interactive video-on-demand systems
IEEE Journal on Selected Areas in Communications
Designing a multiuser HDTV storage server
IEEE Journal on Selected Areas in Communications
A shared buffer architecture for interactive VOD servers
Cluster Computing
A Fibre Channel Dimensioning for a Multimedia System with Deterministic QoS
NETWORKING '02 Proceedings of the Second International IFIP-TC6 Networking Conference on Networking Technologies, Services, and Protocols; Performance of Computer and Communication Networks; and Mobile and Wireless Communications
Hi-index | 0.24 |
Video-on-demand (VOD) servers need to be efficiently designed in order to support a large number of users viewing the same or different videos at different rates. In this paper, we propose a simple quasi-static retrieval scheme that may be used for this purpose when interactive disk-array based video servers are used. We assume presence of no buffer at the server and hence, all retrieved segments are immediately transmitted to the appropriate users. Such a retrieval scheme employs a macro-slot assigner and a disk scheduler. Macro-slots define a duration of time in a service round within which one video segment needs to be retrieved for a user. Having assigned macro-slots to users, the disk scheduler uses the non-preemptive earliest deadline first (N-EDF) algorithm to determine a valid schedule for segment retrieval. In order to increase the number of users that can be served by the system, the number of disk seeks in a service round needs to be decreased. This is done in the Scan-N-EDF algorithm by appropriately grouping retrievals and using the scan algorithm within each group. For a given video segment size, the amount of buffer required at each user's set-top box is reduced to two video segments. Guarantees are provided for the avoidance of video starvation as well as buffer overflow at each user's set-top box.