Research Of Scalable Multimedia Disk Scheduling Algorithm

With the popularity of multimedia applications, related technologies( such as Heap Store, Video Compression, and High-speed Networks, etc.) has been further promotion. Both the disk capacity and the performance of main memory are much more higher than the disk performance. Therefore, the disk I / O performance becomes the bottleneck in lots of applications, especially in multimedia application environments. Because of a large number of I/O operations in multimedia applications, while the merits of the disk scheduling algorithm, largely determine the server's I/O performance and service quality.In the past years of research, domestic and foreign experts proposed many effective disk scheduling algorithms such as FCFS, SCAN, etc.. These scheduling algorithms are to improve the disk I/O performance by optimizing seek time, but not considered the request's deadline requirements, so can not meet the requirements of real-time multimedia applications. Then, they proposed many real-time disk scheduling algorithms such as EDF, SCAN-EDF, etc. However, the limitations in the environments of them make they can not be used generally. Therefore, the design of a multimedia disk scheduling algorithm which can ensure the real time requirements and quality of service (QoS), while maximizing the disk utilization , is of great significance.A new multimedia disk scheduling algorithm which based on SFC theory, termed MSSDS(Multi-Staged Space Disk Scheduling), is presented in this paper. According to the multiple QoS requirements of user requests, such as real-time deadline and user priority, this research has worked on the following aspects:(1) Model (D+2)-dimensional space.D dimensions represent the different types of priorities, one dimension represents deadline, another dimension represents cylinder position.(2) Model three stages of space-filling curves.The first stage takes the D types of priorities as input and outputs a one-dimensional value, using a D-dimensional space-filling curve (MSSDS-SFC1) that aims to minimize priority inversion. The second stage takes the output of the first stage and the disk request deadline as input, and applies MSSDS-SFC2 that aims to minimize deadline violations. Similarly, the third stage takes the output of the second stage and the cylinder position of the disk request, and applies MSSDS-SFC3 that optimizes the seek time. (3) Performance Analysis.To evaluate the performance (priority inversion, deadline missing, disk utilization etc.) of the MSSDS multimedia disk scheduler, we perform several simulation experiments using a simulator that models the Pana Viss server.(4) Practical Considerations and Applications.We apply the MSSDS disk scheduler to the design of a scheduler for a real-time file system that will be used in the NewsByte50.The Previous disk schedule algorithms just work on the simple QoS requirements of dick requests. MSSDS, which combines different priorities and QoS requirements, proposes a positive direction for the research.