JP5842266B2 - Video distribution platform and video distribution method - Google Patents

Description

  The present invention relates to a video distribution platform, and more particularly, to a video distribution platform by a gateway adopting a building block method that facilitates function expansion such as video processing requested by a user.

  With the spread of the Internet in recent years, video distribution services have been developed, and video distribution platforms that can easily distribute individual videos such as Ustream (registered trademark) to an unspecified number of people are becoming popular. Such services have already been commercialized, and in fiscal 2010, the market size is expected to be 640 million yen for use in advertising and publicity, and 740 billion yen in 2020, 10 years later. Yes.

  On the other hand, changes have also occurred in recent video viewing modes. Research has also been conducted on personalized distribution that distributes video according to the environment and preferences of video viewers, and public viewing that shares the sense of reality by viewing the same video scene (for example, Non-Patent Documents 1 to 3). reference). Personalized distribution requires video quality conversion according to the wireless environment, addition of loss correction information such as forward error correction (FEC), and addition of various effect effects. Public viewing requires content information and There are various demands such as synchronization of video scenes and sharing of screen layout information.

  In addition, the present inventor has made a research presentation on scalable high-definition video distribution infrastructure (see, for example, Non-Patent Document 4), but details of gateway processing are awaiting future research.

Koichi Abe, Kaori Maeda, Hiroyuki Inoue, Toru Kondo, “Design and Implementation of a Video Distribution System that Enables Sharing of Viewing Experience”, Information Processing Society of Japan Research Report, 2010, Vol.2010-IOT-8, No. 34, pp.1-6 Hiroyuki Kasai, Hideyoshi Tominaga, “Scalable Gateway Method Based on Multiple Transcoder Stream Switching”, IEICE Transactions, 2003, Vol. J86-B, No. 10, pp. 2229-2244 GStreamer, [online], [searched August 11, 2011], Internet <URL: http://gstreamer.freedesktop.org/> Toru Kondo, “Study on scalable high-definition video distribution platform by building block method”, [online], [searched on November 1, 2011], Internet <URL: http://kaken.nii.ac.jp/pdf /2010/seika/mext/15401/20700066seika.pdf>

  Although the video distribution platform provides the content and viewing format requested by the viewer, many of the existing video distribution platforms such as Ustream are often distributed in a fixed format, and content processing as described above It is not easy to add new functions. Also, when the platform provides processing functions such as screen division and encoding format conversion, the same processing is often performed for each viewer's request in order to perform processing according to the viewer's individual request. As described above, the current video distribution platform is not flexible in response to a new processing request for content, and it is difficult to reuse the content, so the processing efficiency of the content is poor.

  In the future video distribution platform, it is desired that the platform can flexibly respond to various demands from viewers regarding video quality and video effects. The closer the distribution content is processed to the user side (edge) on the network, the worse the efficiency of content processing on the platform and the distribution efficiency of network traffic. In other words, by appropriately arranging content that can be commonly used after processing on the network, there is no need for individual processing at the edge, and there is a high possibility that network resources can be used for sharing.

  SUMMARY OF THE INVENTION An object of the present invention is to provide a video distribution platform that can easily perform function expansion related to content distribution and is excellent in content processing efficiency.

  In the video distribution platform according to the present invention, instead of centrally processing various processing requests, gateways for processing (converting) content are distributed in the platform, and the processing status of each gateway is managed. Thus, the processed output stream can be reused for a plurality of identical requests as a shared resource of the platform. The gateway is configured by blocking each function of the gateway so that a new processing request (function realization) is facilitated. In order to realize these, three elemental technologies of one-source multi-use, building block method, and inter-functional block linkage management are adopted.

  One-source multi-use means that one piece of information is used for various purposes. For example, it is possible to use a photographed video for a movie or a DVD. In this specification, it refers to performing various processing on one stream, converting it into a plurality of streams, and delivering the stream.

  The building block method indicates that when building a system, elements such as hardware and OS are disassembled and necessary functions are selected to build a building block. In the present specification, as shown in FIG. 1, one video processing function is set as one block, and a necessary block is connected according to a processing request. Thus, by sharing the interface of each block, the function expansion can be implemented independently in units of blocks, so that it is not necessary to reconstruct the function and can be expanded easily and quickly.

  In order to enable flexible content reuse, in addition to the building block method, the gateway has an inter-functional block cooperation management mechanism as shown in FIG. Functional block linkage management refers to controlling and managing input / output of functional blocks. As a result, the functional blocks can be flexibly rearranged in response to a request from the viewer, the intermediate state of stream generation can be taken out, and multi-use can be performed in units of functional blocks, so that flexible content reuse is possible.

Based on the above, the video distribution platform according to one aspect of the present invention is arranged in each domain of the network, and performs video processing according to each user's request on the content of the same source, and after the video processing multicast delivery to the competent domain content, and a plurality of streams gateway for relaying transfer content without performing image processing in the case of distributing content between different domains, the plurality of streams gateway in response to a request from the user And a management server that issues a content transfer request to one of the plurality of stream gateways when delivering a content distribution request and a processing request to any one of the plurality of domains, Perform linked stream processing. Then, the stream gateway, when said management server receives a plurality of distribution requests and processing requests for content from determines module and its connection sequence necessary for image processing according to each processing request, the content of each distribution request If there is a module that matches the module determined to be necessary among the already activated modules being processed, the already activated module is reused, and if not, the module is newly activated, and these are reused and newly activated. The modules are linked in the determined order while sharing a part to newly generate the content related to each distribution request and processing request, while when the content transfer request is received from the management server, the content distribution source or a content received from another stream gateway, as it is without being subjected to image processing Transferred to another stream gateway al.

  According to the present invention, it is possible to easily perform function expansion related to content distribution. In addition, it becomes possible to respond quickly even when the same request is issued from multiple receiving clients, and it is possible to improve the efficiency of the entire platform by distributing the processing load of the gateway and reducing the distribution traffic of the same format stream. it can.

Conceptual diagram of the building block method Conceptual diagram of functional block linkage management 1 is a configuration diagram of a video distribution platform according to an embodiment of the present invention. Flow chart from user video distribution request to distribution Schematic diagram of stream gateway configuration model STGW processing flow chart

  DESCRIPTION OF EMBODIMENTS Hereinafter, embodiments for carrying out the present invention will be described with reference to the drawings. In addition, this invention is not limited to the following embodiment.

(overall structure)
FIG. 3 shows a configuration of a video distribution platform according to an embodiment of the present invention. The video distribution platform according to the present embodiment includes a delivery domain (DD) 10 that is an organizational unit such as an ISP (Internet Service Provider) to which the client belongs, and a viewer interface (VI) that is a viewing web application used when the client views. ) 12, a stream gateway (STGW) 14 that performs video processing processing, a content management server (MGSV) 16 that manages content and users, information on each STGW 14, and provides control according to user requests and viewing web applications Composed.

  The STGW 14 is arranged in units of DD, and the MGSV 16 can manage the information of each STGW 14 to request processing from the STGW 14 according to the user. Multiple locations are also assumed for load balancing. It is assumed that IP multicast (hereinafter referred to as multicast) can be used in the DD, and multicast is used to distribute a plurality of video and audio. Delivery between different DDs is performed by conversion to multicast or unicast.

  The MGSV 16 manages the delivery address and quality information as the content information, and transmits the content information and provides the viewing screen according to the request from the user. Further, selection of STGW 14 suitable for the user and load distribution are performed based on regular feedback information from STGW 14.

  VI12 can share screen layouts and receive multi-streams using software that plays back streams on a web browser. Since it is provided on a web browser, there is no need to install dedicated software on the viewing terminal itself.

(Video distribution flow)
FIG. 4 shows a flow from a video distribution request from a user to distribution. Table 1 shows information managed by the MGSV 16.

  The MGSV 16 manages content information and the distribution status between the STGWs 14 and STGWs 14. With these pieces of information, it is possible to reuse content and perform distributed processing according to the state of the STGW 14. Further, the MGSV 16 manages the content distributed between the STGWs 14 and the number of viewers for the content using Stream_info. When a video distribution stop request is made, a request to stop the process used for distribution is transmitted to each STGW 14 only when it is determined that there is no viewer at the distribution destination based on the Stream_info information.

  As shown in (5) of FIG. 4, before the content distribution is performed as necessary, a process for changing the platform state to a form suitable for the content distribution is performed in advance. For example, when it is desired to ensure delivery reachability between different DDs, encapsulation into unicast for ensuring multicast reachability is performed. When the machining function that the STGW 14 does not have is expanded, a necessary machining function is downloaded from the MGSV 16.

(STGW configuration model)
FIG. 5 schematically shows a configuration model of the STGW 14. A functional block for performing processing on the video is realized by a module activated by the STGW 14. The STGW 14 includes a controller that controls the module, a database (DB) that manages the module status, an INPUT / OUTPUT module that performs video acquisition / distribution to the network side, and a processing module group that performs video processing. Each module has an interface for receiving a control message from the controller and for transmitting / receiving data, which are defined as Control-InterFace (C-IF) and Stream-InterFace (S-IF), respectively. The C-IF and S-IF are interfaces that facilitate inter-block management, and socket communication instead of pipes may be used in the local host to reduce input / output processing load. This makes the interface of each module common.

Specifically, the module is a video compositing process that brings together a plurality of contents input from another module activated by a stream gateway that activates the module or a content distribution source or another stream gateway , Video segmentation processing that divides input content into multiple blocks and outputs at least one of them, quality conversion processing that converts the frame rate of input content, forward error correction for packet loss of input content A loss correction process for correcting by retransmission control is performed.

(STGW processing example)
Next, a processing example of the STGW 14 on the premise of transmission / reception of an RTP (Real-time Transport Protocol) multicast stream will be described. FIG. 6 shows the flow of processing from receiving a request from MGSV 16 to responding to it. Table 2 shows information managed by the STGW 14. It is assumed that content is distributed from the user terminal Src_host belonging to the jurisdiction domain of STGW-C to the user terminal Dst_host belonging to the jurisdiction domain of STGW-A via STGW-B.

  The controller of the STGW 14 (STGW-A) determines a module necessary for processing according to a request message from the MGSV 16. At this time, the order of combination with modules necessary for processing requirements is also defined in advance. Each module state is managed in the Module_state_info table in the DB, and it is determined whether start / stop processing is necessary. Multi-use processing is performed as video distribution processing, and usable identifier information to be allocated to the S-IF is searched from the Port_info table, and the C-IF is transmitted to each module to create the S-IF. Each module that has received the request generates and deletes the S-IF and responds to the transmission. After receiving the response to the transmission from all the modules, the controller updates each DB and notifies the MGSV 16 of the content reception waiting port.

  When the MGSV 16 receives the notification of the content reception waiting port from the STGW-A, the MGSV 16 requests the content transfer with the content_id, the address of the STGW-A, and the reception waiting port information to the upstream STGW-B. In response to the transfer request, STGW-B prepares for network transparency processing and notifies MGSV 16 of the content reception waiting port. When the MGSV 16 receives the notification of the content reception waiting port from the STGW-B, the MGSV 16 requests the content transfer to the upstream STGW-C with the content_id, the STGW-B address, and the reception waiting port information. In response to the transfer request, STGW-C prepares to transfer the content uploaded from Src_host to STGW-B and returns a response to MGSV 16. When receiving a response from STGW-C, MGSV 16 updates the Content_info table and the Stream_info table and notifies Dst_host of the distribution destination address of the content related to the request. As a result, content that has undergone user-desired video processing is distributed from Src_host to Dst_host.

(Function block linkage management process)
Management of cooperation between functional blocks is performed using the S-IF of each module. The STGW 14 manages content_id, which is a stream identifier received in the Process_info table in the DB, and the S-IF identifier if_id of each module that is processing the stream identifier. Further, the number of stream processes used by each S-IF is managed in the If_info table. This makes it possible to control the input / output of the module used for each content being processed, and to generate, delete, and link S-IFs according to the processing.

  Next, multi-use processing using inter-functional block linkage management will be described with reference to FIG. The processing is shown in the following list. For the purpose of explanation, a TRANS process performed using three modules (modules 1, 2, and 3 in FIG. 6) is taken as an example, and information on usage examples is described in the figure. Multi-use is performed based on the S-IF of each module. Therefore, it is possible to determine whether it is necessary to delete the S-IF by providing dummy information in each S-IF and managing the number of times of multi-use.

  (1) Search if_id used for the stream to be processed from the Process_info table. In this example, if_id = 1, 2, 3, 4, 5, 6 corresponds.

  (2) Obtain information about if_id from the If_info table based on the searched if_id, and extract module_name and option. For example, when the module performs CODEC processing, the frame rate and resolution are various, and therefore the frame rate and resolution can be specified as options.

  (3) The information in (2), the module name required for the processing, and the option included in the message from MGSV 16 are matched. In this example, matching is established until if_id = 4.

  (4) If matching is established, the if_id information is stored and the dummy information is incremented by one.

  (5) If matching is not established, the stored if_id is determined to be a multi-use S-IF, and the multi-use processing ends. In this example, when if_id = 5, the option content in the DB is different from the option content of the message, so it is determined that it is not established.

  The above processing is repeated for the number of if_id acquired in (1) or until matching is not established. The failure of matching is determined by matching whether the option content is the same as the required module name. This means whether the processing content of the processing unit is the same, and it is possible to detect the S-IF up to the processing level similar to the processing request.

(Comparison with existing system)
Table 3 shows a comparison result between the video distribution platform according to the present invention and the existing system. For comparison, the video distribution system of Non-Patent Document 1, the scalable gateway of Non-Patent Document 2, and the Gstreamer of Non-Patent Document 3 are denoted by Abe, SG, and Gst, respectively.

  Since SG and Gst are single gateways, the distributed arrangement of gateways cannot be compared. However, the present invention and the system of Abe et al. Share multiple processing functions not only within the gateway but also throughout the platform by using multiple distributed arrangements. Can be achieved. In addition, the present invention achieves ease of stream reuse and function expansion, which is not found in the Abe system.

  The video distribution platform according to the present invention can easily expand the functions related to content distribution, and is excellent in content processing efficiency, so that the same video scene can be viewed simultaneously on a plurality of user terminals. It is useful for such as.

14 STGW (Stream Gateway)
16 MGSV (Management Server)

Claims (7)

When placed in each domain of the network, video processing according to each user's request is performed on the content of the same source, and the content after the video processing is multicast distributed within the jurisdiction domain, and the content is distributed between different domains Has multiple stream gateways that relay and transfer content without video processing ,
Issues a distribution request and processing requests for content to any of said plurality of streams gateway in response to a request from the user, in the case of distributing content between different domains of the content in any other of the plurality of streams gateways A management server that issues a transfer request,
The stream gateway, when said management server plurality receiving a delivery request and processing requests for content from determines module and its connection sequence necessary for image processing according to each machining request, processes the content according to the distribution request If there is a module that matches the module determined to be necessary among the already activated modules, the already activated module is reused, and if not, a new module is activated, and these reused and newly activated modules are When a content transfer request is received from the management server while newly generating content related to each distribution request and processing request by linking in the determined order while sharing a part, as yet another of the content received from the stream gateway, without performing image processing Video distribution platform, characterized in that to transfer the stream gateway. The video distribution platform according to claim 1,
The module performs video composition processing for combining a plurality of contents input from another module activated by a stream gateway that activates the module or a content distribution source or another stream gateway. A video distribution platform characterized by this. The video distribution platform according to claim 1,
The module divides the content input from another module activated by the stream gateway that activates the module or the content distributor or another stream gateway into a plurality of blocks and outputs at least one of them A video distribution platform that performs video segmentation processing. The video distribution platform according to claim 1,
The module performs a quality conversion process for converting a frame rate of content input from another module activated by a stream gateway that activates the module, a content distribution source, or another stream gateway. A video distribution platform featuring The video distribution platform according to claim 1,
The module corrects the packet loss of the content input from another module activated by the stream gateway that activates the module or the content distribution source or another stream gateway by forward error correction or retransmission control . A video distribution platform characterized by processing. When placed in each domain of the network, video processing according to each user's request is performed on the content of the same source, and the content after the video processing is multicast distributed within the jurisdiction domain, and the content is distributed between different domains Includes a plurality of stream gateways that relay and transfer content without video processing, and a content distribution request and processing request to any of the plurality of stream gateways in response to a request from a user, and content between different domains And a management server that issues a content transfer request to another one of the plurality of stream gateways,
When the stream gateway receives a plurality of content distribution requests and processing requests from the management server, determining a module necessary for video processing and a connection order thereof according to each processing request;
The stream gateway reuses the already-activated module if there is a module that matches the module determined to be necessary among already-activated modules that are processing the content related to each distribution request, and if not, a new module The step of starting
The stream gateway newly generates a content related to each distribution request and processing request by connecting the reused and newly activated modules in the determined order while sharing a part ;
When the stream gateway receives a content transfer request from the management server, the content received from the content distribution source or another stream gateway is transferred to another stream gateway as it is without video processing ; A video distribution method comprising: The video distribution platform according to any one of claims 1 to 5,
When the management server manages the content distributed between the stream gateways and the number of viewers corresponding to the content, and determines that there is no viewer at the content distribution destination when receiving a video distribution stop request from the user Sends a request to each stream gateway to stop processing used for content distribution
A video distribution platform characterized by this.