Electrical computers and digital data processing systems: input/ – Input/output data processing – Input/output data buffering
Reexamination Certificate
1999-10-06
2002-08-20
Gaffin, Jeffrey (Department: 2182)
Electrical computers and digital data processing systems: input/
Input/output data processing
Input/output data buffering
C710S005000, C711S158000, C711S114000, C709S241000
Reexamination Certificate
active
06438630
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
This invention relates to computer data storage and server systems, and more particularly to digital video/audio storage and playback systems supporting multiple continuous media streams.
2. Description of the Related Art
Multimedia or video server systems are used in a variety of applications for the storage and playback of video, audio or other multimedia data streams. For example, multimedia servers may be used for broadcast, cable, satellite or telco solutions to distribute multimedia information to clients or consumers. Professional broadcasters and associated service providers, such as networks and affiliates or cable providers, may employ digital video servers to support high bandwidth multimedia broadcast applications including multi-channel program playout, ad insertion, and digital content management. Other applications for multimedia server systems may include computer-based training in which multimedia training materials or lectures may be stored on the server system accessed by students over a network or the internet.
Video archiving, browsing and retrieval is another multimedia server application. Various movies may be stored by the server and distributed to users upon request. Video-on-demand or video delivery systems may enable a plurality of users or viewers to selectively watch movies or other audio/video sequences which are stored on one or more video servers or media servers. The video servers may be connected through data transfer channels, such as a broadcast cable system, satellite broadcast system or the internet, to the plurality of users or subscribers. The video servers may store a plurality of movies or other audio/video sequences, and each user can select one or more movies from the video servers for viewing. Each user may include a television or other viewing device, as well as associated decoding logic, for selecting and viewing desired movies. When a user selects a movie, the selected movie may be transferred on one of the data transfer channels to the viewing device of the respective user. Multimedia servers are also found in webcasting applications in which entertainment may be multicast on the internet to different subscribers. Multimedia servers are found in numerous other applications as well.
To meet the demands of many different applications and users, it is desirable for a multimedia server system to provide flexibility and extensibility. Two important requirements for a multimedia server system are storage space and filesystem bandwidth. Multimedia data, such as full-motion digital video, requires a large amount of storage and data transfer bandwidth. Thus, multimedia systems use various types of video compression algorithms to reduce the amount of necessary storage and data transfer bandwidth. In general, different video compression methods exist for still graphic images and for full-motion video. Video compression methods for still graphic images or single video frames may be intraframe compression methods, and compression methods for motion video may be interframe compression methods.
Examples of video data compression for still graphic images are RLE (Run-Length Encoding) and JPEG (Joint Photographic Experts Group) compression. Although JPEG compression was originally designed for the compression of still images rather than video, JPEG compression is used in some motion video applications. Most video compression algorithms are designed to compress full motion video. Examples of video compression techniques are MPEG (Moving Pictures Experts Group), MPEG-2, DVI (Digital Video Interactive) and Indeo, among others.
Even with the use of compression techniques, multimedia applications may still require extremely large amounts of storage. For example, two hours of video encoded at. 1 Mb per second may require roughly one gigabyte (1 GB) of storage. A system supporting numerous different content may require up to several terabytes (TB) of storage. The server system must also be able to provide enough bandwidth for the various users to access selected multimedia content without overloading the storage system. For example, to support 100 simultaneous subscribers viewing multimedia content encoded at 1 Mb per second, a server may need to support a bandwidth in excess of 100 Mb per second when allowing for overhead. If enough bandwidth is not available, then some requests may have to be denied, or the play quality may suffer (video may run too slowly or may appear “jerky”). To meet such storage and bandwidth needs, a multimedia server may utilize one or more RAID storage systems. In a RAID system, for a given multimedia file, blocks of multimedia data may be stored across multiple hard disk units. The blocks may be read out or transferred to the communication network and transmitted or broadcast to the user or users. At the receiving end the blocks may be decoded for user viewing on a display device.
The use of hard disk drives to store information magnetically is well known. A single hard disk drive typically includes several rotating disks or platters upon whose surfaces information may be written or read by a read/write head. Within a hard disk drive the platters rotate together and all read/write heads typically are moved simultaneously to access different platter positions. A platter typically is formatted into concentric tracks that define collectively, platter-to-platter, a family of concentric cylinders, and into sectors that represent a portion of a track. A controller associated with the hard disk unit determines which cylinder, read/write head and sector is to be accessed for reading or writing.
The platters may also be considered as being divided into zones. Since they are physically larger, tracks in zones at the outer platter parameter contain more sectors than tracks in zones near the rotational axis of the platter. Therefore, assuming the platters rotate with a constant velocity, the data bandwidth available from the outer most zones is greater than the data bandwidth available from the innermost zones. Even with modern hard disk drives, there can be a 2-1 variation between worst case and average case disk transfer bandwidth due to sectors/track variations between outer and inner zones.
Many multimedia applications require continuous media streams in which data streams must be delivered at a specified and possibly time-varying data rates and with a specified uniformity of that delivery rate. To support continuous media streams, prior video streaming systems have been based on a constant time, variable data storage scheduler. Constant time, variable data schedulers deliver a stream at a particular bit rate by accessing the disk with a frequency that is common with all streams and varying the amount of data fetched at each period to meet the required data rate of the stream. The admission control problem for constant time, variable data disk schedulers is fairly straight forward. A new stream may be admitted if the disk can sustain the required aggregate data rate over the fixed access. period.
FIG. 1
illustrates a constant time/variable data rate dependent data placement/scheduling mechanism For simplicity only two streams are illustrated. Data blocks
100
are accessed at a fixed period common to all streams. The size of the data blocks
100
is varied between different streams according to the rate required for a particular stream. As shown in
FIG. 1
, stream one has a larger block size and thus a higher data rate than stream two. When streams are recorded to disk using a constant/time variable data scheduling mechanism, the placement of data on the disks is rate dependent.
Constant time schedulers may have some deficiencies however. These deficiencies may include significant internal fragmentation of storage. Internal fragments occur because storage is allocated in multiples of some fixed-size block. A minimum block size is the size of the device sector, but typically the blocks may be larger to minimize the size of free block lists. Internal fragmentation may occur in
Conley Rose & Tayon PC
Elamin Abdelmoniem
Gaffin Jeffrey
Kowert Robert C.
Sun Microsystems Inc.
LandOfFree
Scheduling storage accesses for multiple continuous media... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Scheduling storage accesses for multiple continuous media..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Scheduling storage accesses for multiple continuous media... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2972733