Class Note for CMPSCI 677at UMass(4)
Class Note for CMPSCI 677at UMass(4)
Popular in Course
Popular in Department
This 16 page Class Notes was uploaded by an elite notetaker on Friday February 6, 2015. The Class Notes belongs to a course at University of Massachusetts taught by a professor in Fall. Since its upload, it has received 16 views.
Reviews for Class Note for CMPSCI 677at UMass(4)
Report this Material
What is Karma?
Karma is the currency of StudySoup.
You can buy or earn more Karma at anytime and redeem it for class notes, study guides, flashcards, and more!
Date Created: 02/06/15
L DEPARTMENT OF COMPUTER SCIENCE UMASS AMHERST Multimedia Servers I Multimedia digital audio video images Streaming audio and video Very different characteristics from textual and numeric files Need different techniques for managing multimedia data Video sequence of images played out at a constant rate Digital video is often stored in compressed format 6772 OPERATING SYSTEMS Lecture 25 Page 1 DEPARTMENT OF COMPUTER SCIENCE UMASS AMHERST Need For Video Compression I 0 Large data rate and storage capacity requirement Satellite 180x180 km2 600 MBimage imagery 30 m2 resolution NTSC video 30 framess 30 MBytess 640x480 pixels 3 bytespixel 0 Compression algorithms exploit Spatial redundancy ie correlation between neighboring pixels Spectral redundancy ie correlation between different frequency spectrum Temporal redundancy ie correlation between successive frames L 6772 OPERATING SYSTEMS Lecture 25 Page 2 DEPARTMENT OF COMPUTER SCIENCE UMASS AMHERST Requirements for Compression Algorithms I 0 Objectives Minimize the complexity of the encoding and decoding process Ensure a good quality of decoded images Achieve high compression ratios 0 Other general requirements Independence of specific size and frame rate Support various data rates L 6772 OPERATING SYSTEMS Lecture 25 Page 3 DEPARTMENT OF COMPUTER SCIENCE UMASS AMHERST Classification of Compression Algorithms I o Lossless com pression Reconstructed image is mathematically equivalent to the original image ie reconstruction is perfect Drawback achieves only a modest level of compression about a factor of 5 o Lossy compression Reconstructed image demonstrates degradation in the quality of the image gt the techniques are irreversible Advantage achieves very high degree of compression compression ratios up to 200 Objective maximize the degree of compression while maintaining the quality of the image to be virtually lossless L 6772 OPERATING SYSTEMS Lecture 25 Page 4 DEPARTMENT OF COMPUTER SCIENCE UMASS AMHERST MPEG An Overview I 0 Two categories intraframe and interframe encoding o Contrasting requirements delicate balance between intra and interframe encoding Need for high compression gt only intra frame encoding is not sufficient Need for random access gt best satisfied by intra frame encoding 0 Overview of the MPEG algorithm DCT based compression for the reduction of spatial redundancy similar to JPEG Block based motion compensation for exploiting the temporal redundancy gtxlt Motion compensation using both causal predictive coding and noncausal interpolative coding predictors g 6772 OPERATING SYSTEMS Lecture 25 Page 5 DEPARTMENT OF COMPUTER SCIENCE UMASS AMHERST Exploiting Temporal Redundancy I 0 Three types of frames in MPEG I fra mes Forward Prediction gtxlt Intra coded frames provide access points for random access yield moderate compression P frames gtxlt Predicted frames are encoded with reference to a previous or P frame B frames gtxlt Bidirectional frames encoded using the previous and the next IP frame gtxlt Achieves maximum compression Bidirectional Prediction L 6772 OPERATING SYSTEMS Lecture 25 Page 6 DEPARTMENT OF COMPUTER SCIENCE UMASS AMHERST Multimedia Storage Servers I o Digitally stores heterogeneous data objects consisting of audio video imagery textual and numeric data on extremely high capacity storage devices o Fundamental differences in data type characteristics and requirements Besteffort service for text vs realtime for video Small readwrites for text vs large readwrites for video L 6772 OPERATING SYSTEMS Lecture 25 Page 7 DEPARTMENT OF COMPUTER SCIENCE UMASS AMHERST 0 Techniques for efficiently managing video data Placement techniques Fault tolerance issues Scheduling retrieval and admission control IO stream sharing buffering batching caching 0 Methodology What are the fundamental issues How to address these issues Theory How to instantiate the solutions Practice L 6772 OPERATING SYSTEMS Lecture 25 Page 8 L DEPARTMENT OF COMPUTER SCIENCE UMASS AMHERST Terminology Actuator Disk fundamentals Seek time Rotational latency Transfer rate Platter Scheduling algorithms FCFS 39 SCAN SSTF SATF 03125quot 6772 OPERATING SYSTEMS Lecture 25 Page 9 L DEPARTMENT OF COMPUTER SCIENCE UMASS AMHERST Terminology Cont d I 6772 OPERATING SYSTEMS 0 Disk arrays o Striping Interleave the storage of each media stream among disks Stripe unit maximum amount of logically contiguous data that is stored on a single disk Degree of striping Number of disks across which a media stream is striped o Redundant and nonredundant disk arrays Lecture 25 Page 10 DEPARTMENT OF COMPUTER SCIENCE UMASS AMHERST Video Storage Server Fundamentals I 0 Data transfer rate of disks gtgt data rate requirement of isolated video streams gt designing singleuser video servers is straightforward 0 Server stores digitized video streams on an array of disks 0 Clients can request the retrieval of video streams for realtime playback 0 Two possible server architectures Client pull Server push L 6772 OPERATING SYSTEMS Lecture 25 Page DEPARTMENT OF COMPUTER SCIENCE UMASS AMHERST Clientpull Architecture I 0 Server retrieves data only in response to an explicit request from client 0 Used in conventional file system to provide besteffort service 0 Adapting clientpull architecture for video clients ensure playback continuity by Determining the playback instant of a frame Estimating response time for each request Issuing a read request accordingly 0 Response time a function of the system load gt varies widely over time gt estimation is nontrivial L CMPSCI 677 OPERATING SYSTEMS Lecture 25 Page 12 L DEPARTMENT OF COMPUTER SCIENCE UMASS AMHERST Serverpush Architecture I 6772 OPERATING SYSTEMS o Periodicity of video playback gt service clients in periodic rounds 0 Round retrieve a fixed number of frames for each media stream 0 Continuous retrieval gt total service time must not exceed the playback duration of frames retrieved during a round Lecture 25 Page 13 L DEPARTMENT OF COMPUTER SCIENCE UMASS AMHERST Efficient Placement on Disk Arrays I Diskl Disk 2 U w 1 Disk 4 WZ o Stripe video streams on disk arrays in terms of blocks or stripe units 0 Two paramaters stripe unit size and degree of striping o Stripe unit size block size use large 128512 KB block size Large block size reduces disk seek and rotational latency overheads 6772 OPERATING SYSTEMS Lecture 25 Page 14 L DEPARTMENT OF COMPUTER SCIENCE UMASS AMHERST Retrieval Techniques I Need to retrieve 30 frame in each second 0 Streaming media data imposes realtime constraints on retrieval Client or server buffering can provide some leeway but still need guarantees clients accessing a server Employ admission control algorithms 6772 OPERATING SYSTEMS 0 Performance guarantees on retrieval gt need to limit the number of Lecture 25 Page 15 DEPARTMENT OF COMPUTER SCIENCE UMASS AMHERST Admission Control I 0 Server push retrieval retrieve f frames in each periodic round R 0 Continuous playback requirements retrieval time of f1 f2 fk frames for all k clients should not exceed R o Admission control test Estimate resoure needs of new client time to retrieve fz frames Verify if total resource needs 3 capacity total retrieval time 3 R If so admit else deny g 6772 OPERATING SYSTEMS Lecture 25 Page
Are you sure you want to buy this material for
You're already Subscribed!
Looks like you've already subscribed to StudySoup, you won't need to purchase another subscription to get this material. To access this material simply click 'View Full Document'