WO2010025614A1

WO2010025614A1 - Method and stream media system thereof for media distribution and position by partitioned storage

Info

Publication number: WO2010025614A1
Application number: PCT/CN2009/000853
Authority: WO
Inventors: 周磊; 孙福清; 程宁
Original assignee: 中兴通讯股份有限公司
Priority date: 2008-09-05
Filing date: 2009-07-30
Publication date: 2010-03-11
Also published as: CN101420603A; CN101420603B

Abstract

The invention discloses a method and a stream media system thereof for media distribution and position by partitioned storage. A central node periodically inserts time tag information into original code streams; the code streams with the same channel identifiers and the same time tag information are named and constitute partitioned files for the partitioned storage; the partitioned files are distributed to each edge node from the central node; the node to which a user belongs positions the partitioned file using the partitioned storage according to a signaling; and a plurality of partitioned files are smoothly switched by the time tag information. As the time tag is periodically inserted into the live code streams and the partitioned files are named by combining a channel identifier, the invention enables the uniform storage of the partitioned files when creating a live channel, and the seamless splicing of the adjacent partitioned files with the time tag, thereby achieving the uniform storage of Time Shift television(TSTV) service and True Video On Demand(TVOD) service, greatly reducing the requirements of the whole system on storage space and storage resources and improving the reliability and the stability of the service of the stream media system.

Description

Method for realizing media distribution and positioning by slice storage and streaming media system thereof

The present invention relates to the field of control methods and systems for streaming media in multimedia communication, and more particularly to a method for realizing media distribution and positioning by fragment storage and a streaming media system thereof. Background technique

Personalized interactive TV or Interactive IPTV (IPTV) is a broadband cable network that integrates Internet, multimedia, communication and other technologies to provide users with digital A new technology for a variety of interactive services, including television. It can adapt well to the rapid development of today's networks and make full use of multiple network resources.

IPTV is different from traditional analog cable TV and from classic digital TV. Because traditional analog TV and classic digital TV have the characteristics of frequency division, timing, one-way broadcasting, etc. Although the classic digital TV has many technological innovations compared to analog TV, it is only a change of signal form, and does not touch the media. The way content is disseminated.

IPTV uses a computer to perform functions such as receiving video-on-demand programs, video broadcasting, and surfing the Internet. It uses high-efficiency video compression technology to achieve a video streaming bandwidth of 800Kb/s, which is close to the 3Mb/s video streaming bandwidth required by DVD. This will enable video services in the future, especially Internet live broadcast, long-distance true video on demand, program source production, etc., have a strong advantage, is a new technical concept. The IPTV service is regarded as the main growth point of the future by the telecom industry due to its interactivity and content richness. In China, where the interactive entertainment boom is prevalent, huge market demand space will lead to rapid industrial growth.

Nowadays, many IPTVs will mention two important functions in their solutions: TVOD (True Video On Demand, TVOD) and Time Shift TV (TSTV). When implementing the functions of TSTV and TVOD, in order to improve the response speed of users accessing the website, the content distribution network CDN (CDN) technology can be used to save the total storage space of all CDN nodes in the entire system. At the same time, we must balance the playback efficiency of users on-demand and the stability of scheduling between nodes. When creating a live channel, if you continue to save the two functional applications Saving and taking different policies to provide services will inevitably result in a large waste of storage space, and will also increase the overhead of storage I/O resources occupied by the system.

Therefore, the prior art needs to be improved and developed. Summary of the invention

The problem to be solved by the present invention is to provide a method for realizing media distribution and positioning by using slice storage and a streaming media system thereof, so that time-shifted television-type services and television program-on-demand services in the interactive network television system can be uniformly stored. Achieve the entire system to reduce the storage space and storage resources requirements.

The technical solution of the present invention is as follows:

A method for fragment storage of a streaming media system to implement media distribution, the method comprising the following steps:

A1, time stamp information is periodically entered by the central node in the original code stream;

Bl. Name the code streams with the same channel identifier and unified time label information, and form a fragment file for fragment storage;

C1. Copy the fragment file from the central node to each edge node to implement media distribution. Further, in the method, the step A1 further includes: the central node changing the storage duration of the fragment file by controlling the length of the time stamp interval.

Further, in the method, the fragment storage process of step B1 further includes one or more of the following operational mechanisms:

Rollback mechanism: Deletes the fragment file that is required for non-recorded programs beyond the time-lapse time range. Aging mechanism: Deletes fragment files that exceed the aging time range and do not need to be saved.

Archive mechanism: Save the shard file that needs to be saved beyond the aging time range.

Further, in the method, the media distribution process in the step C1 further includes one or more of the following operation strategies:

Hotspot push policy: The media location service system copies a segment file with a top access to program access from the central node to an edge node without the slice file through a file transfer protocol;

Manual distribution strategy: The operator creates a hotspot for the program through the background operation, and the media location service system triggers execution of the hotspot push;

Program pull-down policy: When a slice file is continuously clicked for more than a specified number of times, the media location service system copies the slice file from the central node to cause a trigger through a file transfer protocol. Edge node.

Further, in the method, the step C1 further includes the step of recording a program, where the universal media server node provides the media location service system with the abnormal fragment file when the recorded fragment file is abnormal. The channel identifier and time stamp information are used to obtain the sharing node information of the fragment file.

Further, in the method, the step C1 further includes a process of the self-recovery of the fragment file recording failure, and the process includes:

The edge node detects a fragment file with a storage error in the recording, notifies the media location service system of the channel identifier and the time label information corresponding to the fragment file, and requests to store the node information of the fragment file normally. ;

C12. The media location service system searches for available nodes according to the channel identifier and the time tag information, and feeds the found node information to the requested edge node;

C13. The edge node selects a node according to a program pull-down policy, and obtains, by using a file transfer protocol, a fragment file corresponding to the fragment file in which the error occurs.

A method for implementing media location by fragment storage of a streaming media system, the method comprising the following steps:

A2, through a media location service system, splicing a fragment file having time stamp information belonging to a program time range to a node capable of providing a complete program fragmentation;

B2. The node capable of providing the complete program fragmentation locates the fragment file when the fragment storage is performed according to the signaling including the channel identifier and the program time range;

C2: Smoothly switch multiple slice files by using time tag information in the original code stream to implement media playback.

Further, in the method, the node that can provide the complete program fragment is the node to which the user belongs, and the step C2 further includes the following common playback process:

C21. The channel identifier of the program and the program time range are fed back to the node to which the user belongs by using real-time streaming protocol signaling;

C22. The node to which the user belongs finds one or more fragment files of the channel identifier and the time stamp information corresponding to the program time range, and performs sequential playback;

C23. After the playback service ends, the system releases the occupied resources and data area.

Further, in the method, the step C2 further includes the following tandem playback process, where The process includes:

C31. If the node to which the user belongs cannot find or find the fragment file corresponding to the full channel identifier and the time label information, request the media location service system to find a node that can provide the fragment file completely;

C32. The media location service system searches for an available center node according to the channel identifier and the time label information, and feeds back the found node information to the node to which the user belongs;

C33. The node to which the user belongs connects the found node and initiates a real-time streaming protocol for media playback;

C34. After the playback service ends, the system releases the occupied resources and data area.

A media storage system for media distribution and media location, comprising a central node and at least one edge node, and a media location service system, wherein the media location service system is respectively connected to the central node and the edge node , among them,

The central node is configured to periodically input time tag information in the original code stream;

The edge node is configured to copy the fragment file from the central node to implement media distribution;

The media location service system is configured to provide sharing information of the central node or the edge node;

The central node and the edge node are further configured to name a code stream having the same channel identifier and unified time label information, and form a fragment file for fragment storage, and perform signaling positioning according to the channel identifier and the program time range. The shard file to the shard storage.

The method for realizing media distribution and positioning and the streaming media system provided by the invention provide a unified time stamp in the live stream and combined with the same channel identifier to name the fragment file, thereby realizing When the live channel is created, the unified storage of the fragment file, plus the seamless splicing of the adjacent fragment files of the time label, achieves the unified storage of the time-shifted TV service and the TV program on-demand service, and further through the media distribution strategy and the media positioning strategy. , greatly reducing the demand for storage space and storage resources of the entire system, and improving the reliability and stability of the streaming media system service. BRIEF abstract

Figure la is a CND networking and code stream diagram of the present invention;

Figure lb is a fragment file directory format of the present invention; Figure lc is a schematic diagram of the storage time period of the fragment of the present invention;

2 is a flowchart of self-recovery of a fragment file recording failure of the present invention;

3 is a flow chart showing the normal play of TVOD on demand of the TV program of the present invention;

4 is a flow chart of the TVOD on-demand TVOD tandem playback of the present invention. Preferred embodiment of the invention

The specific implementation manners of the media distribution, the positioning method, and the streaming media system thereof are described in detail below in conjunction with the illustrated drawings, that is, combined with the self-recovery process of the fragment file recording failure, the TVOD on-demand TVOD normal playback process, and The tandem playback process introduces a streaming media storage strategy, a media distribution strategy, and a media location strategy, and is mainly applied to an interactive network television service such as a television program on-demand or a time-shift television type of a computer or a set top box + a television or a mobile phone.

The method for realizing media distribution and positioning and the streaming media system thereof in the fragment storage of the present invention, the main core point is that media distribution and media positioning are implemented by streaming media fragment storage, that is, by using a unified time label and adding the same channel identifier to name the fragment. And the means for applying for sharing the node information to the media positioning system to save storage space and smooth switching play. The technologies such as TSTV, TVOD, and CDN are well known in the prior art, and are not described herein again.

First, we need to explain the streaming storage strategy and its three related operating mechanisms:

As shown in FIG. 1a, the whole system is used for simple networking between CDN nodes in IPTV, and includes a central node or two central nodes that are mutually backup (not shown in the drawing) and multiple edge nodes, through the media. The Location Service System (MLSS) is used to schedule and control all nodes. The central node converts the original live stream into a live stream that has entered the time stamp, and then distributes it to each edge node as the live stream of the edge node. The central node and edge node fragment files are recorded and stored in the same way, and are named and fragmented according to the specific rules of the time label.

For a single CDN node, a time stamp is periodically entered in the live stream, and the streaming media storage policy is to combine a code stream having the same channel identifier and a time label into a fragment file, and the fragment file may also be The slice directory file is named by the channel identifier and the time stamp, and the storage duration of each slice file is changed by controlling the interval of the time stamp.

As shown in Figure lb, CCTV1-T (20080514080000-20080514083000) can be regarded as the fragment file described above, which is a fragment catalog file, which contains multiple small points. The file file includes a live channel identifier such as CCTV1 and recorded time stamp information such as 20080514080000-20080514083000, and the code stream information of the live channel in the time stamp is also recorded in the content of the plurality of small slice files.

In order to keep the storage tempo of different CDN nodes consistent, it is stipulated that only the central node sends a unified time label to the live stream, and then distributes the live stream after the time stamp is distributed from the central node to each edge node, such that In this way, each node can be stored in a uniform beat, that is, all nodes are simultaneously stored. The same channel identifies the fragment file named by the time label stored by the next node. After being distributed to another node, if another node also has the same channel identifier and a storage file named by an adjacent time label, then the two The file can achieve seamless stitching of the playback stream. Since the storage is distributed to each node in the previous distribution, theoretical playback can be played from each edge node. However, the node to which the user belongs is preferred.

However, each node does not completely save all live programs of each channel indefinitely, because each channel has a rollback mechanism when it is stored in live broadcast, that is, when the channel storage rollback spans more than one span For a specific value, such as when the time shift time range is exceeded, the system will delete the oldest stored fragment file. In order to save storage space, the rollback duration span set by the central node must meet various possible requirements of the user, and the edge node's rollback duration span can be set to a smaller value because the edge node stores the score. When the slice file cannot meet the user's requirements, the rollback duration span can be obtained from the distribution node and the neighboring node through the distribution and positioning policies. For programs that need to be archived for a long time, the system will copy the file corresponding to the program and save it. The junk files or invalid files generated in the above operations can also be deleted by the aging policy.

For various reasons, the node that the user belongs to saves the score file is not complete, and can not find individual fragments. Then you need to find a node that can be played completely. This node may be other edge nodes or central nodes. . If the found node is another edge node, the streaming media playing process is a "normal playing process"; if the found node is a central node, the streaming media playing process is "tapping playback process".

As shown in Figure lc, the figure illustrates the rollback mechanism, the aging mechanism, and the archiving mechanism. The rollback time can be generally defined as the time shift time, and the fragment file exceeding the time shift time range. If it is not the fragment file required for the TVOD recording program, the fragment file will be deleted; the fragment exceeding the aging time range File, if it is not a program that needs to be saved for a long time, the fragment file will be deleted. The mechanism can clear the expired T VOD recording program fragments and various garbage fragment files; and the fragment files exceeding the aging time range, if they need to be saved for a long time, will not be deleted and will be saved, and saved. Go to other locations. By deleting the fragment file in the one-dimensional time, the storage capacity of the system node can be reduced, the problem of repeated storage of the TSTV and TVOD services is better solved, and the total storage space of all CDN nodes is saved. In addition, the program performance of the program provided to the user is improved, and the stability of scheduling between nodes is enhanced.

Secondly, the fragmentation file fails to record the self-recovery process to illustrate the media distribution strategy:

In the system of simple networking between CDN nodes in IPTV, the self-recovery process for the failure of the fragment file recording of a certain node X is as shown in FIG. 2, which specifically includes the following steps:

Step S202, the edge node X detects that a certain fragment file has an abnormality in the recording, and has a storage error;

Step S204, the edge node X notifies the media location service system MLSS of the channel identifier and the time label information corresponding to the abnormal fragment file, and requests the MLSS to store other node information of the fragment file normally;

Step S206, the MLSS finds the available node Yz according to the channel identifier and the time tag information.

( ζ = 1, 2, ... η ), feeding back the found node information to the requested edge node X;

Step S208, the edge node X selects a node Yz according to the program pull-down policy, and uses a file transfer protocol FTP (File Transfer Protocal, hereinafter referred to as FTP), from the node Υζ (ζ=1, 2, ... η) Pull down the required shard file to get the shard file corresponding to the storage error.

In order to save storage space, only the central node stores all the partial files needed to provide the service, so that when the edge node does not exist but some of the fragment files existing in the central node are or may be frequently served, media distribution is required. The strategy copies these fragments from the central node to the edge node, which specifically includes hotspot push, program pulldown, and manual distribution strategies:

The media location service system MLSS counts all the nodes according to specific rules, and accesses the top-ranked program, that is, the fragment file where the hotspot is located, and the hotspot file that is not on the edge node will be served through the file transfer protocol FTP service. The shard file is copied from the central node to the edge node. At this time, the copying of the hotspot file is triggered by the media location service system MLSS, which is a so-called hotspot push policy.

If the operator artificially forms a hotspot through the background operation, then the hot push strategy Similarly, triggered by the media location service system MLSS, copying the slice file from the central node to the edge node is called a manual distribution policy.

The program pull-down policy means that, for a specific edge node, when the on-demand TVOD is connected multiple times in a specified number of times, the media location service system MLSS pulls the fragment file to the node through the FTP service. . The copying action of the shard file is triggered by the edge node itself, as in step 208 above.

Finally, the media positioning strategy is described by the two playback processes of the TV program on-demand TVOD: Taking the TV program on-demand TVOD normal playback process as an example, as shown in FIG. 3, the following steps are included: Step S302, the user operates the electronic program guide EPG through the remote controller ( The EPG feeds the recorded program information of the on-demand to the set top box STB (Set Top Box); in step S304, the set-top box STB uses the Real Time Streaming Protocol (RTSP) signaling to mark the channel of the recorded program and The time range is fed back to the node to which the user belongs; Step S306, the node to which the user belongs determines the one or more fragment files corresponding to the electronic program guide EPG according to the channel identifier and the program time range;

Step S308, the node to which the user belongs sends the requested media stream to the set top box STB, and when the program corresponds to multiple fragment files, each edge node automatically completes the play switching between the fragment files according to the streaming media storage policy. Process

Step S310, the end of the recorded program service provided by the end, the system releases the occupied resources and the data area.

Time-shifted TV The normal playback process of TSTV is similar.

For example, the TV program on-demand TVOD tandem playback process, as shown in FIG. 4, includes the following steps: Step S402, which is the same as step S302;

Step S404, which is the same as step S304;

Step S406, if the edge node X cannot find or find the fragment file corresponding to the full-on-demand program according to the channel identifier and the time label information, then request the media location service system MLSS to provide the node that can provide the fragment file completely;

Step S408, the MLSS finds one or more central nodes that can provide services according to the channel identifier and time tag information provided by the edge node, and returns the node information that can provide the service to the edge node to which the user belongs;

Step S410, the edge node to which the user belongs initiates a live stream to the central node capable of providing the service. Transport protocol RTSP tandem process;

Step S412, the edge node to which the user belongs starts to provide services to the user;

Step S414, the operation of the user actually acts on the central node that can provide the service by the tandem; step S416, which is the same as S308;

Steps S418, S420 and S422 are the end of steps S414, S412 and S410, respectively. Time-shifted TV The TSTV's tandem playback process is similar.

When the stored file is recorded, if the fragment file recorded by the USS node of a common media server is unsuccessful, the media location service system MLSS is requested to obtain the shared node information that can provide the fragment file. As long as the corresponding channel and time stamp information is provided, the fragmented file with unsuccessful recording can be obtained from the sharing node.

When the user places a program on TVOD or TSTV, the set-top box STB (Set Top Box) needs to include channel and program time range information in the on-demand signaling. Since the fragment file contains the recorded time information, the node to which the user belongs can be provided according to the set-top box. The channel and program time range, locate the required fragment file. If the required fragment file does not exist on the node or does not exist on the node, it needs to be connected to the MLSS through the media location service system. The center node or edge node of the fragment file, that is, the "television playback process" or "normal playback process".

It is worth noting that, regardless of the playback process of the situation, once determined, the media is played entirely from the determined node. In addition, the program time range of the set-top box on-demand may include multiple fragment files, and the time label information can smoothly switch the playback of multiple fragment files.

The specific implementation manners of the above streaming media storage policy, media distribution policy, and media location strategy are embodied in the process of recording file failure failure self-recovery, the TV program on-demand TVOD normal playback process, and the tandem playback process, for fragment storage, media distribution. Methods such as media positioning and the like can be applied to various possible solutions according to actual applications, which are well known to those skilled in the art and will not be described herein.

The method for implementing media distribution and positioning and the streaming media system provided by the fragment storage provided by the specific embodiment of the present invention, because a unified time stamp is regularly entered in the live stream, and the same channel identifier is used to name the fragment file. It realizes the unified storage of the fragment files when creating the live channel, and the seamless splicing of the adjacent fragment files of the time label, which realizes the unified storage of the time-shifted TV-type business and the TV program-on-demand service, and further through the media distribution strategy and media. The positioning strategy greatly reduces the storage system and storage resource requirements of the entire system, and improves the reliability and stability of the streaming media system service. It should be understood that those skilled in the art can modify or change according to the description of the above scheme, such as the naming method of the fragment file, etc., and all such improvements and transformations should belong to the appended claims. protected range. Industrial applicability

The method for realizing media distribution and positioning and the streaming media system provided by the invention provide a unified time stamp in the live stream and combined with the same channel identifier to name the fragment file, thereby realizing When the live channel is created, the unified storage of the fragment file, plus the seamless splicing of the adjacent fragment files of the time label, achieves the unified storage of the time-shifted TV service and the TV program on-demand service, and further through the media distribution strategy and the media positioning strategy. , greatly reducing the demand for storage space and storage resources of the entire system, and improving the reliability and stability of the streaming media system service.

Claims

Claim

A method for implementing media distribution by fragment storage of a streaming media system, the method comprising the steps of:

C1. Copy the fragment file from the central node to each edge node to implement media distribution.

2. The method according to claim 1, wherein the step A1 further comprises:

The central node changes the storage duration of the fragment file by controlling the length of the time stamp interval.

3. The method according to claim 2, wherein the fragment storage in the step B1 comprises one or more of the following operational mechanisms:

Rollback mechanism: Deletes the fragment file that is required for non-recorded programs beyond the time-lapse time range; aging mechanism: deletes the fragment file that exceeds the aging time range and does not need to be saved; and the archiving mechanism: saves the aging time range and needs to be saved Fragment file.

4. The method according to claim 1, wherein the media distribution in the step C1 comprises one or more of the following operational strategies:

Hotspot push strategy: The media location service system copies the top file of the program access ranking from the central node to the edge node of the slice file without the program access ranking by the file transfer protocol;

Manual distribution strategy: The operator creates a hotspot for the program through a background operation, and the media location service system triggers execution of the hotspot push;

Program pull-down policy: When a slice file is continuously clicked for more than a specified number of times, the media location service system triggers the copying of the slice file that is continuously requested by more than a specified number of times from the central node by using a file transfer protocol. Edge node.

5. The method according to claim 1, wherein the step C1 further comprises the step of recording a program, wherein

When the recorded fragment file is abnormal, the universal media server node provides the media location service system with the channel identifier and the time label information of the abnormal fragment file to obtain the shared node information of the abnormal fragment file.

The method according to claim 5, wherein the step C1 further includes a process of failing to recover from the fragment file recording, the process comprising:

The edge node detects a fragment file with a storage error in the recording, notifies the media location service system of the channel identifier and the time label information corresponding to the fragment file stored in the error, and requests to store the storage normally. Node information of the wrong fragment file;

C12. The media location service system searches for available nodes according to the channel identifier and the time tag information, and feeds back the information of the found node to the requested edge node;

C13. The edge node selects an available node according to a program pull-down policy, and obtains, from the available node, a fragment file corresponding to the fragment file in which the error occurred, by using a file transfer protocol.

7. A method for implementing media location by fragment storage of a streaming media system, the method comprising:

B2, the node capable of providing the complete program fragmenting is located in the fragment file when the fragment storage is performed according to the signaling including the channel identifier and the program time range;

C2: The edge node to which the user belongs uses the time tag information in the original code stream to smoothly switch the plurality of slice files to implement media playback.

The method according to claim 7, wherein the node capable of providing a complete program fragment is a node to which the user belongs, and the step C2 includes a normal play process, and the normal play process includes:

C22. The node to which the user belongs finds one or more fragment files of the channel identifier and the time range corresponding to the time stamp information, and performs sequential playback;

The method of claim 8, wherein the step C2 further comprises a tandem playback process, and the tandem playback process comprises:

C31. If the node to which the user belongs cannot find or find the fragment file corresponding to the channel identifier and the time label information, request the media location service system to search for the fragment corresponding to the channel identifier and the time label information. The node of the file;

C32. The media positioning service system searches for available media according to channel identifier and time tag information. a heart node, and feeding back the found node information to the node to which the user belongs;

C33. The node to which the user belongs connects to the found node and initiates a real-time streaming protocol for media playback;

10. A streaming media system for implementing media distribution and media location, comprising a central node, at least one edge node, and a media location service system, wherein the media location service system and the central node and the at least An edge node communication connection, wherein

The media location service system is configured to provide sharing information of the central node or the at least one edge node;

The central node and the at least one edge node are further configured to name a code stream having the same channel identifier and unified time label information, and form a fragment file for fragment storage, and according to a letter including a channel identifier and a program time range. Let the shard file be located when the shard is stored.