Method and apparatus for instant playback of a movie title

• US 20080282036A1
• Filed: 03/12/2007
• Published: 11/13/2008
• Est. Priority Date: 03/09/2005
• Status: Active Grant

• Current Assignee: NBC Universal Media LLC (Comcast Corporation)
• Original Assignee: Vvond, LLC (Comcast Corporation)
• Inventors: Ganesan, Prasanna
• Granted Patent: US 7,698,451 B2
• Time in Patent Office: Days
• US Class Current: 711/129
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 7/01 Probabilistic graphical models, e.g. probabilistic networks
• CPC Class Codes: H04L 65/60 Network streaming of media packets
• CPC Class Codes: H04L 67/1091 Interfacing with client-server systems or between P2P systems
• CPC Class Codes: H04N 21/4331 Caching operations, e.g. of an advertisement for later insertion during playback
• CPC Class Codes: H04N 21/47202 for requesting content on demand, e.g. video on demand
• CPC Class Codes: H04N 21/632 using a connection between clients on a wide area network, e.g. setting up a peer-to-peer communication via Internet for retrieving video segments from the hard-disk of other client devices broadcast-related systems characterised by transmission among terminal devices H04H60/80; distributed application using peer-to-peer [P2P] networks H04L67/104
• CPC Class Codes: H04N 21/8456 by decomposing the content in the time domain, e.g. in time segments
• CPC Class Codes: H04N 7/17336 Handling of requests in head-ends

Abstract

Techniques for fragmenting a file or a collection of media data are disclosed. According one aspect of the techniques, a file pertaining to a title is fragmented into a header and several tails or segments. The header is a continuous portion of the file while the segments are respective parts of the remaining portion of the file. The header is seeded substantially in all boxes, and none, one or more of the segments are distributed in each of the boxes in service. When a title is ordered, the header is instantly played back while the segments, if not locally available, are continuously fetched respectively from other boxes that have the segments.

Claims

1. A method for fragmenting a file pertaining to a title, the method comprising:

dividing the file into a sequence of data blocks, the sequence of data blocks partitioned into a first portion and a second portion, where each of the data blocks includes at least several Kbytes;

forming a header from the data blocks in the first portion, the data blocks in the header being consecutive, a data size of the header predetermined to ensure that a playback of the title proceeds without jitters; and

forming N segments, each of the N segments including some of the data blocks in the second portion, the data blocks in each of the N segments are nonconsecutive, wherein N is a finite integer greater than 1.

2. The method as recited in claim 1, wherein the file is a collection of audio and video data together with auxiliary data, if there is such auxiliary data.

3. The method as recited in claim 2, wherein the header is locally cached in each of boxes in service as a residing object, and M of the N segments are distributed as distributed objects in each of the boxes, wherein M varies between 0 and N from one box to another.

4. The method as recited in claim 3, wherein the data size of the header is so determined that, when the title is played back, the header lasts just long enough to start receiving the M segments from one or more of the boxes.

5. The method as recited in claim 4, wherein the data size of the header is extended to include some extra time for synchronizing the M segments being respectively fetched from the one or some of the boxes.

6. The method as recited in claim 3, wherein the data size of the header is determined to last a certain period.

7. The method as recited in claim 6, wherein when the certain period ends, some data from the M segments being fetched from one or more of the boxes has already arrived.

8. The method as recited in claim 3, wherein the data size of the header is dynamically defined to take care of possible abnormal conditions of a network, and be at least D*X+Max (over all times t>

=X) [byte offset for corresponding frame−

D*t], where it is assumed that (a) an ordering box is downloading data at a fixed rate D and (b) a downloading suffers from some interruptions that result in the box not receiving any data at all for a total period of no more than X seconds.

9. The method as recited in claim 3, wherein the data size of the header is statistically determined in a region to have enough data to take care of irregularities in downloading the M segments from one or more of the boxes.

10. The method as recited in claim 9, wherein the header includes at least three sections of data, each section specifically provided to deal with a particular irregularity.

11. The method as recited in claim 10, wherein the three sections include an A/V data buffer, a downstream jitter buffer, and an upstream jitter buffer, where the A/V data buffer is defined to last X seconds, the downstream jitter buffer is defined to last Y seconds, and the upstream jitter buffer is defined to last Z seconds, and wherein the A/V data buffer is enough before some data from the M segments becomes available to succeed the header, Y is statistically determined to ensure that there are enough data to overcome possible jitters in downloading, and Z is statistically determined to ensure that there are enough data to overcome possible jitters due to a sudden loss of the required uploading bandwidth.

12. The method as recited in claim 3, wherein the data size of the header is dynamically determined with respect to a title and a box the header is to be cached.

13. The method as recited in claim 12, wherein the data size is predetermined to ensure that there is enough data to supplement a playback when an actual bitrate is higher than an average bitrate.

14. The method as recited in claim 12, wherein the header includes at least an A/V data buffer, a downstream jitter buffer, and an upstream jitter buffer, the A/V data buffer includes enough data to supplement a playback when an actual bitrate is higher than an average bitrate, the downstream jitter buffer and the upstream jitter buffer include additional data that supplement the average bitrate when an already allocated bandwidth for downloading or uploading becomes narrower.

15. The method as recited in claim 1, wherein a FIFO data buffer is provided and fed with the header first followed by received data from the M segments being continuously fetched from the one or more of the boxes together with data from N-M segments locally cached.

16. The method as recited in claim 15, wherein, when the title is ordered for playback, received data from the M segments being continuously fetched from the one or more of the boxes together with data from N-M segments are multiplexed into the data buffer.

17. The method as recited in claim 16, wherein the multiplexed data in the data buffer is part of the second portion of the data blocks.

18. The method as recited in claim 1, wherein, when the title is executed, received data from L of the M distributed objects being continuously fetched from one or more of the boxes together with data from N-M segments are multiplexed into a data buffer, and wherein 1≦

L≦

M.

19. The method as recited in claim 18, wherein one of the L distributed objects pertains to audio data in a preferred language, and the one of the L segments and the N-L segments are sufficient to render a playback of the title.

20. The method as recited in claim 1, wherein the forming of the N segments comprises decimating simultaneously the second portion of the data blocks into the N segments.