JP2006129514A - Packet data copy distributing method and apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a data transmission system and transmission apparatus for effectively utilizing bands of a transmission line in the data transmission system which transmits an IP packet by dividing a meaningful data constitution such as a video frame or slice and discards the packet in transmission network congestion. <P>SOLUTION: Flow identifier data of hierarchical program data to be processed when congestion occurs, and control code data for starting or stopping discarding packet data are held and if congestion occurs, for data having flow identifier data indicating the held hierarchical program data, a selective transmission unit 3 starts or stops discarding the packet data based on the control code data for starting/stopping discarding of the packet data. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to distribution, transfer, and reception / reproduction of a program including video and audio data, and more particularly to a program transmission / reception system including a plurality of streams.

  As data transmission in an ATM (Asynchronous Transfer Mode) network, there is a best effort UBR (Unspecified Bit Rate) service. As a transmission control method for improving the throughput during UBR transmission, “Study on Selected Cell Discarding Control Method with Dynamic Control of Two Thresholds” (Technical Report of IEICE, Vol.99, No.621) 55 to 60 (2000.02)), an EPD (Early Packet Discard) control system is known. The UBR service is a best effort transmission service in an ATM network, and discarding occurs in units of ATM cells when the transmission network is congested. When discarding occurs in units of ATM cells, transmission of ATM cells that should have become invalid after cell discard is performed even though all packet data before segmentation including the discarded cells becomes invalid. The bandwidth is wasted. Therefore, in the event of congestion, the EPD control method does not discard in units of cells but in units of packet data.

On the other hand, as a transmission method of video data in an IP (Internet Protocol) network, “encoding transmission control method in video multicast” (Journal of the Institute of Image Information and Television Engineers, Vol. 52, No. 6, pages 863 to 870 (1998-6)). )It has been known. This method describes a control method for transmitting hierarchically encoded video program data from a video server to a client using a plurality of channels (multicast addresses). In this method, the I, P, and B frame data defined by the MPEG (Moving Picture Experts Group) encoding method, the low-frequency component and the high-frequency component of the spatial frequency component in each frame are defined as video data hierarchies. A total of six types of encoded data of components are distributed from the server to the client using a plurality of channels. In the I, P, B frame data defined by the MPEG encoding method, I frame data is necessary to reproduce P frame data, and I frame, P to reproduce B frame data. Frame data is required, and as a result, importance exists in the order of I frame, P frame, and B frame. Also, the spatial frequency component in each frame has an importance relationship such that the low frequency component representing the basic contour is more important than the high frequency component representing the details of the video. In the coded transmission control method in this video multicast, the six types of hierarchically coded data are distributed from the video server to the client using six channels.

IEICE technical report, Vol.99, No.621, pp. 55-60 (2000.02) The Journal of the Institute of Image Information and Television Engineers, Vol.52, No.6, 863-870 (1998-6)

  When data transmission is performed in an IP (Internet Protocol) network that is one of the best effort networks, discarding is performed in units of packets when congestion occurs in the transmission network.

  For example, when the data to be transmitted is video data, a part of the data constituting the video frame or slice is discarded. When a part of data constituting a video frame or a slice is discarded, a problem similar to the phenomenon described in the item of ATM transmission in the prior art occurs. That is, there is a problem that, even though the video frame or slice data including the discarded packet is invalidated as a whole, the invalidated packet is continuously transmitted even after the packet is discarded, and the bandwidth of the transmission path is wasted. .

  In addition, when packet discarding is started in the middle of transmission of a video frame or slice data, there is a problem that the video data that has already been transmitted is invalidated. When transmission is started, there is a problem in that invalid data is transmitted until the start of the next video frame or slice. Further, when data is distributed by multicast, there is a problem that if copying is started from the middle of video frame data and multicast distribution is performed, invalid data is distributed to the beginning of the next frame data.

  For example, in the hierarchical video program data transmission composed of a plurality of streams described in the section of the prior art, each layer of data is packetized into a plurality of packets (herein referred to as “packets”) and transferred. The In this case, since the frame can be reproduced without the high frequency components of the I, P, and B pictures, the priority of the packet storing the high frequency components of the I, P, and B pictures is low. However, in the hierarchical video program data transmission using the conventional router, when congestion occurs, data packets are randomly discarded as shown in FIG. 25, so even important data packets may be discarded. There have been problems such as a large number of errors in the decoding / playback of video data, or playback stoppage. Therefore, when video data is hierarchized and transmitted to discard data, it is important to selectively discard low priority data packets. However, even in the case of a low-priority packet group, when some packets are transferred to the reception / playback apparatus and some packets are not transferred there, an error occurs for the convenience of decoding video data, When the frame is played back, there is a problem that the image quality is affected, such as flickering on the screen.

Therefore, an object of the present invention is to perform data transmission that effectively uses the bandwidth of a transmission line in a data transmission method in which a meaningful data structure such as a video frame or a slice is divided and IP packets are transmitted and packets are discarded when the transmission network is congested. It is to provide a method and a transmission apparatus.
It is another object of the present invention to provide a data transmission method and a transmission apparatus that effectively utilize the bandwidth of a transmission path when multicasting data.

  Further, the present invention provides a data transmission method for discarding low priority packet data when the transmission network is congested during transmission of video / audio program data composed of a plurality of streams. A system and a transmission apparatus are provided.

  In the present invention, when a meaningful data structure such as a video frame or a slice is divided and transmitted by IP packet, and packet discard due to congestion of the transmission network is performed, the meaningful data (video frame or slice) is discarded from the middle. Instead of starting or ending the discarding, the above object is achieved by starting the packet discarding or ending the discarding in accordance with the start of meaningful data.

  In the present invention, at the time of multicast distribution of data, replication is not started in the middle of meaningful data such as a video frame or a slice, but packet replication is started at the start of meaningful data. This achieves the object.

Further, in the present invention, when video / audio program data composed of a plurality of streams is transmitted, when discarding low priority packet data, as shown in FIG. 24, discarding is performed in units of packets belonging to the low priority class. In other words, the object is achieved by performing selective transmission of data packets belonging to each layer or not.
More specifically, the following means are used.

  First, the present invention provides a sequence for each stream when a server distributes a plurality of streams, for example, program data (hereinafter referred to as “hierarchical program”) composed of hierarchically encoded video data described in the conventional example. A means for assigning a layered header with a number to form layered packet data, a unit for converting data obtained by layered packet data into UDP (User Datagram protocol) packet data, and further converting it into IP (Internet Protocol) packet data; In the DS (Differentiated Services) field value given in the header of the packet data, an identifier (hereinafter referred to as “flow identifier”) data for identifying each stream constituting the hierarchical program, and the discard start or end of the packet data are indicated. Hierarchical program data with identifiers (hereinafter referred to as “control codes”) for execution. The IP packet data, and using the means for delivering.

  Secondly, according to the present invention, in the IP packet data transfer apparatus, the flow identifier data of the hierarchical program data to be processed when congestion occurs and the control code data for executing the discard start or the end of the packet data are retained. When congestion occurs, the start or discard of discarding is performed based on the control code data for executing packet data discarding start / end processing for data having flow identifier data indicating the retained hierarchical program data. The means for executing the termination is used.

  Thirdly, according to the present invention, in a server for shaping hierarchical program data, flow identifier data of hierarchical program data to be processed when the sequence numbers in the hierarchical packet data of the hierarchical program data to be processed are discontinuous, and When control code data for executing start or end of discarding of IP packet data is held, and a predetermined discontinuity occurs in the sequence number for each hierarchical program stream to be processed, the stored hierarchical program stream to be processed For IP packet data having flow identifier data, data is shaped by executing discard start or discard termination based on control code data for executing packet data discard start / end processing. Use means.

  Fourth, the present invention provides a flow identifier data of hierarchical program data to be processed when a sequence number in the hierarchical packet data of the hierarchical program data to be processed is discontinuous in a client that receives and reproduces hierarchical program data. And control code data for starting or terminating the discarding of IP packet data, and when a predetermined discontinuity occurs in the sequence number for each hierarchical program stream to be processed, the stored hierarchical program stream to be processed For the IP packet data having the flow identifier data, the data is shaped by executing the discard start or the discard end based on the control code data for executing the packet data discard start / end processing. Use means to do.

  Fifth, the present invention uses means for creating label data from DS values (flow identifier data and control code data) in a data conversion server that converts IP packet data into MPLS (Multi-Protocol Label Switching) packet data. .

  Sixth, the present invention holds a label to be processed when congestion occurs in an apparatus that performs packet data transfer by MPLS, and when congestion occurs, the MPLS packet data is stored based on the held label. A means for executing the disposal start or the disposal termination is used.

  That is, the packet data transfer method according to the present invention is a packet data transfer method in an IP (Internet Protocol) network or an MPLS (Multi-Protocol Label Switching) network, and data to be processed by a packet data transfer apparatus having a plurality of input / output ports. The identifier (flow identifier) data for identifying the packet and the identifier (control code) data for controlling the process are stored, and the packet data to which the flow identifier data and the control code data are attached are received, and the packet data transfer device When the packet data identified by the flow identifier data is discarded at the time of congestion in the network, the discard start and the discard end are performed based on the control code data.

  When the control code data is a start code, the packet data identified as the packet data to be processed is started to be discarded and discarded from packet data including predetermined control code (start code) data.

  When the control code data is an end code, the discard start and discard end operations of the packet data identified as the packet data to be processed are performed from the packet data next to the packet data including the control code (end code) data.

  When the flow identifier data indicates video data, the control code data is any of a sequence start code, a GOP (Group Of Pictures) start code, a picture (video frame) start code, and a slice start code included in the video data. Control code data created based on the above.

  The packet data transfer method monitors whether or not the amount of buffer data at the output stage of the packet data transfer apparatus is greater than or equal to a predetermined amount (hereinafter referred to as “discard start / end point”), and the buffer data amount increases and is discarded. When packet data having the specified control code data and flow identifier data is received when the start and end points are exceeded, discard of the packet data having the flow identifier data is started, and discarding starts and ends when the buffer data amount decreases. When packet data having predetermined control code data and flow identifier data is received below the point, discarding of the packet data having the flow identifier data can be terminated.

  Also, a plurality of discard start / end points are set in association with different flow identifier data, and whether or not the buffer data amount exceeds the set discard start / end points is monitored, and the buffer data amount increases. When packet data having a predetermined control code data and flow identifier data associated with the discard start / end point is received, the discard of the packet data having the flow identifier data is started. When the packet data amount is decreased and packet data having a predetermined control code data and flow identifier data associated with the discard start / end points is received, the flow is decreased. You may make it complete | finish discard of the packet data with identifier data.

  Whether the buffer data amount at the output stage of the packet data transfer apparatus is greater than or equal to a first predetermined amount (hereinafter referred to as “discard start point”), and a second predetermined amount (less than the first predetermined amount) (Hereinafter referred to as “discard end point”), and if the buffer data amount increases and exceeds the discard start point, packet data having a predetermined control code data and flow identifier data is received. When the packet data having the flow identifier data is started to be discarded, and when the buffer data amount is decreased and the packet data having the predetermined control code data and the flow identifier data is received, the flow identifier data is changed. You may make it complete | finish discard of the packet data which it has.

  Furthermore, a plurality of discard start points and discard end points are set in association with different flow identifier data, and whether or not the buffer data amount is greater than or equal to the set discard start point, and whether or not there are more than the set discard end points. When the packet data amount increases and exceeds each discard start point and packet data having predetermined control code data and flow identifier data associated with the discard start point is received, the flow identifier data is When packet data having flow identifier data associated with the discard end point is received, the discarding of the packet data held is started, and the buffer data amount is reduced to fall below each discard end point. You may make it complete | finish discard of the packet data with the said flow identifier data.

A packet data transfer method according to the present invention includes a plurality of ingress cards each connected to an input line, a plurality of egress cards each having a data discard function and a buffer, and each connected to an output line, and a plurality of ingress cards. Using a router having a card and a switch connected to a plurality of egress cards, a field for setting destination address information, a field for setting flow identifier data for identifying each layer, and discarding A header having a field for setting control code data for starting and ending, and transferring frame data in which frame-coded video frame data is packetized into a plurality of packet data for each layer A packet data input to the ingress card to the switch; When the amount of packet data staying in each buffer exceeds a predetermined threshold value, the packet data input to each buffer is transferred to each flow identifier data. Further, it is characterized in that each layer is discarded based on the control code data.

  A packet data transfer apparatus according to the present invention includes a plurality of ingress cards each connected to an input line, a plurality of egress cards each having a data discard function and a buffer and each connected to an output line, and a plurality of ingress A switch connected to the card and a plurality of egress cards, and a field in which destination address information set in the ingress card is set, and a field in which flow identifier data for identifying each layer is set. Packet data obtained by converting a frame-coded video frame data into a plurality of packet data for each layer, having a header provided with a field for setting control code data for starting / ending discarding To the switch and to the egress card corresponding to the value in the address field Packet data transfer device, when the amount of packet data staying in each buffer exceeds a predetermined threshold, packet data input to the buffer, based on the control code data for each flow identifier data, A means for discarding each hierarchical unit is provided.

  The data distribution method according to the present invention is a data distribution method on an IP network, and includes flow identifier data for identifying transmission data and control code data for controlling the start or end of discarding of transmission data during transmission. Are arranged in a DS (Differentiated Services) field in the IP packet header, and the transmission data is distributed.

  The data distribution method according to the present invention is also a data distribution method on the MPLS network, and is a flow identifier data for identifying transmission data and a control for controlling the start or end of discarding of transmission data during transmission. Code data is arranged in a label field in an MPLS packet header, and transmission data is distributed.

  The packet data creation method according to the present invention is a packet data creation method for creating packet data from hierarchical data composed of a plurality of streams. Flow identifier data for identifying each hierarchical data to be transmitted and congestion in the middle of transmission Control code data for starting or ending the discarding operation is generated for each hierarchical data divided for each predetermined size to create hierarchical packet data, and further, a UDP (User Datagram Prtocol) header is added. It is characterized in that it is converted into UDP packet data.

  The data shaping method according to the present invention includes flow identifier data for identifying each hierarchical data of hierarchical data composed of a plurality of streams, and sequence numbers that are continuously given to data divided for each predetermined size. In addition, control packet data for starting or ending the discarding operation of each layer data is added to each layer data divided for each predetermined size to create layer packet data, and further converted into UDP packet data and IP packet data Receiving a data sequence distributed and distributed, reconfiguring UDP packet data and hierarchical packet data from the received IP packet data sequence, discarding data that cannot be reconstructed UDP data, and each flow identifier data A step for confirming the continuity of the sequence number of hierarchical packet data reconstructed every time When the sequence number is discontinuous and the control code data is control code data for starting the discard operation, the layer packet including the next control code data in each layer packet data received as a successor A layer packet including the next control code data in each layer packet data received as a successor when the layer packet data immediately before the data is discarded and the control code data is control code data for terminating the discard operation And a step of discarding the data and converting the subsequent hierarchical packet data into UDP packet data and IP packet data and distributing the data to the same transmission destination as at the time of reception.

  A data shaping device according to the present invention includes flow identifier data for identifying each hierarchical data of hierarchical data composed of a plurality of streams, and sequence numbers that are continuously given to data divided for each predetermined size. In addition, control packet data for starting or ending the discarding operation of each layer data is added to each layer data divided for each predetermined size to create layer packet data, and further converted into UDP packet data and IP packet data Means for receiving converted data, means for reconfiguring UDP packet data and layer packet data from received IP packet data, means for discarding data that cannot be reconfigured if UDP packet data cannot be reconfigured, The continuity of the sequence number of the reconstructed hierarchical packet data is determined for each flow identifier data. Means for confirming, and when the sequence number is discontinuous, if the control code data is control code data for starting the discarding operation, it includes the next control code data in each layer packet data received as a successor Up to the layer packet data immediately before the layer packet data is discarded, and when the control code data is control code data for terminating the discard operation, the next control code data in each layer packet data received as a successor is included. A means for discarding up to layer packet data and converting the subsequent layer packet data into UDP packet data and IP packet data and delivering them to the same destination as when received, and received when the sequence number is continuous When receiving all layer packet data as UDP packet data and IP packet data Characterized by comprising a means for delivering to the same destination.

  The decoding method according to the present invention is continuous to flow identifier data for identifying each hierarchical program data of hierarchical video / audio program data composed of a plurality of video / audio streams and data divided for each predetermined size. A layer packet data is created by assigning a sequence number assigned to each layer data and a control code data for starting or ending the discard operation of each layer data for each layer data, and further creating a UDP packet. Receiving packet data and IP packet data sequence distributed as IP packet data, reconfiguring UDP packet data and hierarchical packet data from received IP packet data sequence, and reconfiguring UDP packet data If not, a step of discarding data that cannot be reconstructed and a reconstructed hierarchy A step of confirming the continuity of the sequence number of the data for each flow identifier data, and when the sequence number is discontinuous, if the control code data is control code data for starting a discarding operation, it is received as a successor The layer packet data immediately before the layer packet data including the next control code data in each layer packet data is discarded, and if the control code data is control code data for ending the discarding operation, it is received as a successor. And discarding the layer packet data including the next control code data in each layer packet data and performing decoding from the layer packet data thereafter.

  The data decoding and reproduction display device according to the present invention is divided into flow identifier data for identifying each hierarchical program data of hierarchical video / audio program data composed of a plurality of video / audio streams, and a predetermined size. The layer packet data is assigned to each layer data obtained by dividing the sequence number continuously given to the data and the control code data for starting or ending the discard operation of each layer data for each predetermined size. Means for generating and receiving IP packet data distributed in the form of UDP packet data and IP packet data; means for reconstructing UDP packet data and hierarchical packet data from the received IP packet data; and UDP packet data If the data cannot be reconstructed, a means to discard the data that cannot be reconstructed and the reconstructed floor Means for confirming the continuity of the sequence number of packet data for each flow identifier data, and when the sequence number is discontinuous, if the control code data is control code data for starting a discard operation Discards the layer packet data immediately before the layer packet data including the next control code data in each received layer packet data, and if the control code data is control code data for ending the discard operation, Means to discard the hierarchical packet data including the next control code data in each received hierarchical packet data and start decoding from the subsequent hierarchical packet data, and received if the sequence number is continuous Means for decoding all layer packet data, and for reproducing and displaying the decoded data Characterized by comprising a stage.

  The packet data copy delivery method (data multicast delivery method) according to the present invention retains flow identifier data for identifying data to be duplicated and control code data for controlling duplication processing. When the packet data identified with the control code data is received and the identified packet data is duplicated, the duplication start and the duplication end are performed based on the control code data for the packet data having the retained flow identifier data. It is characterized by.

  A replica distribution apparatus (data multicast distribution apparatus) according to the present invention includes means for holding flow identifier data for identifying data to be replicated and control code data for controlling replication processing, and flow identifier data Means for receiving packet data to which data and control code data are added, and means for starting replication and ending replication based on control code data for packet data having retained flow identifier data when replicating packet data It is characterized by comprising.

  As described above, according to the present invention, meaningful data such as a video frame or a slice is divided and transmitted by IP packet, and when discarding a packet due to transmission network congestion, meaningful data (video frame (Or, instead of starting or terminating discarding in the middle of a slice), by starting or discarding packets in accordance with the start of meaningful data, the bandwidth of the transmission path can be reduced. It is possible to provide a data transmission method and a transmission system that are effectively utilized.

  In the present invention, at the time of multicast distribution of data, replication is not started from the middle of meaningful data such as a video frame or a slice, but packet replication is started at the start of meaningful data. By doing so, it is possible to provide a data transmission method and a transmission device that effectively utilize the bandwidth of the transmission path.

  Further, in the present invention, when video / audio program data composed of a plurality of streams is transmitted, when discarding low priority packet data, the reception / playback side apparatus performs discarding in units of packets belonging to the low priority class. Can provide a data transmission method and a transmission apparatus that prevent disturbance of reproduction.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

  First, the data structure used in this embodiment will be described. FIG. 3 shows an IP packet data conversion method for hierarchical program data used in the present embodiment. Here, as the hierarchical program data, high frequency component data 40 and low frequency component data 41 of an I frame of MPEG video data (MPEG-1 or MPEG-2) will be described as an example. The distributed hierarchical program data 40 is divided into hierarchical data (1) and hierarchical data (2) of 43 and 45 by dividing them into a predetermined size, for example, every 4 kbytes, and given hierarchical headers 42 and 44, respectively. Thus, hierarchical packet data is created. Here, the layer headers 42 and 44 are a 4-bit identifier (hereinafter referred to as “flow identifier”) data for identifying the layer data, and a 4-bit identifier (hereinafter referred to as the head or middle of each layer frame data). , “Control code”) and a 16-bit sequence number assigned continuously to each stream constituting the hierarchical program data. The values of the flow identifier data and control code data will be described in detail with reference to FIG. As sequence numbers, the high-frequency component data 40 and the low-frequency component data 41 of the I frame are cyclically assigned values of 0x0000 to 0xFFFF in hexadecimal notation. However, although the flow identifier data and the control code data are 4 bits and the sequence number is a 16-bit value here, other values may be used.

  Further, the data 42 and 43 and the data 44 and 45 converted into hierarchical packet data are respectively provided with UDP headers 46 and 47 to constitute UDP packet data. Further, the data 46, 42, and 43 and 47, 44, and 45 converted into UDP packet data are each divided into predetermined sizes, for example, 1480 bytes, and IP headers 48, 49, and 50 are added to the top, respectively. Configure packet data. In the example of FIG. 3, since the size of the UDP packet data composed of the data 46, 42 and 43 exceeds 1480 bytes, the 43 hierarchical data (1) is divided into 43-1 and 43-2, and the IP packet It shows the state that is converted into data.

  The DS values shown in 51, 52 and 53 will be described later with reference to FIG. 4, but are values set according to the data transmitted by the IP packet data, and are used to identify each stream constituting the hierarchical program. It is a value composed of flow identifier data and control code data for executing start or end of discarding of packet data. These DS values (data 51, 52, 53) are created based on the flow identifier data (data 55, 56) and control code data (data 60, 61) values in the layer header. The sequence data 65 and 66 are 16-bit sequence numbers that are continuously assigned to each stream as described above.

  FIG. 4 shows details of the DS value set in accordance with the content of data transmitted by the IP packet data. For example, the DS value 51 in FIG. 3 is determined by the content of the data 43-1 to be transmitted, and the DS value 52 is determined by the data content 43-2. Since the data 43-1 described in FIG. 3 is the high frequency component data of the I frame, the flow identifier data is 0x5 and the top data of the high frequency component data of the I frame. 0xD is set. The entire DS value indicated by 51 in FIG. 3 is 0x5D from the flow identifier data (0x5) and the control code data (0xD). Since the data 43-2 in FIG. 3 is high frequency component data of the I frame, the flow identifier data is 0x5. However, since it is intermediate data of the high frequency component data of the I frame, the control code data is 0xC. Is set. As a result, the entire DS value indicated by 52 in FIG. 3 is 0x5C from the flow identifier data (0x5) and the control code data (0xC). Further, since the data 45 in FIG. 3 is the high frequency component data of the I frame, the flow identifier data is 0x5 and the intermediate data of the I frame high frequency component data, so that 0xC is set as the control code data. The As a result, the entire DS value indicated by 53 in FIG. 3 is 0x5C from the flow identifier data (0x5) and the control code data (0xC).

  Except for the high frequency component of the I frame described above, as shown in FIG. 4, when the data transmitted by the IP packet data is a low frequency component of the I frame, 0x6 is assigned as the flow identifier data, and the P frame Is 0x4 as flow identifier data, and 0x3 is assigned as flow identifier data when it is a high frequency component of a P frame. Further, when the data transmitted by the IP packet data is a low frequency component of a B frame, 0x2 is assigned as flow identifier data, and when the data is a high frequency component of a B frame, 0x1 is assigned as flow identifier data. As the control code data, 0xD is assigned when the data to be transmitted is the head data of each video frame, and 0xC is assigned as the control code data when the data to be transmitted is halfway data of the video frame. The above is the relationship between the content of data transmitted by the IP packet data and the DS value, but settings other than the values set in FIG. 4 may be used. In addition, as a code for controlling the start and end of discarding of packet data, it is set depending on whether or not it is the head of each video frame, but other distinctions, for example, whether or not it is the head of GOP (Group Of Pictures), or It is also possible to set by distinguishing whether or not it is the head of a slice of each video frame.

  FIG. 1 is a diagram illustrating a configuration of a packet data transfer apparatus 100 according to an embodiment of the present invention. Specifically, input port 1 (1-1 to 1-N) for connecting to a plurality of input lines, routing unit 2 for routing packet data, and data routed for a plurality of output lines are selected. For selective transmission 3 (3-1 to 3-N) for transmission, buffers 4 (4-1 to 4-N) connected to each selective transmission, and a plurality of output lines It consists of output ports 5 (5-1 to 5-N). Input ports 1 (1-1 to 1-N) can be made of ingress cards. Further, the selective transmission unit 3 (3-1 to 3-N), the buffer 4 (4-1 to 4-N) connected to each selective transmission unit, and the output port 5 (5 for connection to a plurality of output lines) -1 to 5-N) can be made as an egress card 201 (201-1 to 201-N).

  Next, the operation of the packet data transfer apparatus 100 will be described. The packet data transfer apparatus 100 first receives the IP packet data shown in FIG. 3, for example, IP packet data composed of the data 48, 46, 42, and 43-1, etc. -N). The received IP packet data is routed to a predetermined output port (in this case, one of the selective transmission units 3-1 to 3-N) through the routing unit 2 according to the destination address of the IP packet header. The packet data transferred to the selective transmission unit 3 (3-1 to 3-N) is transmitted to the selective transmission unit 3 (3-1 to 3 to 3) by the data amount signal 21 (21-1 to 21-N) of the buffer connected to the subsequent stage. -N) to determine whether or not to transmit to buffer 4 (4-1 to 4-N).

  In the example of the packet data transfer apparatus 100 in FIG. 1, the data amount in the buffer 4 (4-1 to 4-N) is a predetermined data amount (the data amount point in the buffer to start or end the discard of packet data). ) Shows the case of monitoring whether or not there is more than the above, and the signal 21 (21-1 to 21-) indicates whether or not the data amount in the buffer 4 (4-1 to 4-N) is greater than or equal to a predetermined amount. N), the selected transmission unit 3 (3-1 to 3-N) is notified. Thus, it is determined whether or not the received IP packet data is transmitted from the selective transmission unit 3 (3-1 to 3-N) to the buffer 4 (4-1 to 4-N). In the case of transmission, IP packet data is transmitted from the selective transmission unit 3 (3-1 to 3-N) to the buffer 4 (4-1 to 4-N). When not transmitting (discarding), IP packet data is not transmitted from the selective transmission unit 3 (3-1 to 3-N) to the buffer 4 (4-1 to 4-N), and the selective transmission unit 3 (3- 1-3-N). The operation of the selective transmission unit 3 (3-1 to 3-N) will be described later with reference to FIG. The IP packet data transmitted in the buffer 4 (4-1 to 4-N) is distributed from the output port 5 (5-1 to 5-N) to the output line.

  FIG. 2 is a diagram showing the configuration of the selective transmission unit 3 in detail. The selective transmission unit 3 includes a selective transmission determination unit 31, a register 32 that temporarily holds data, and a transfer unit 33 that executes data transfer. The IP packet data routed by the routing unit 2 is temporarily held in the register 32, and the DS value in the IP header is analyzed by the selective transmission determination unit 31. The selective transmission determination unit 31 previously stores and holds the flow identifier data of IP packet data to be discarded when the amount of data in the buffer 4 is equal to or greater than a predetermined amount (during congestion) and control code data for starting / ending the discarding. deep. Here, an example will be described in which a high frequency component (flow identifier data 0x1) of a B frame is set as a flow to be discarded when the buffer 4 is congested.

  The packet discarding starts from packet data having control code data of 0xD, and for a discarding flow, a flag indicating that the discarding is in progress is accumulated and held in the selective transmission determining unit 31 and discarded. When the flag is set and the congestion of the buffer 4 is eliminated, transmission from the packet data having the control code data 0xD to the buffer 4 is started. That is, when the buffer 4 is congested, discarding is not started while packet data having a DS value of 0x1C is received, and discarding is started after receiving IP packet data having 0x1D. If the discard flag is set and the congestion of the buffer 4 is eliminated, the discarding is not completed while the packet data having the DS value of 0x1C is received, and the IP packet data having the 0x1D is received. The selective transmission determination unit 31 controls the transfer unit 33 so as to end the discarding. The transfer unit 33 performs an operation of transferring or discarding the data read from the register to the buffer under the control of the selective transmission determination unit 31.

  The above is the configuration and operation of the selective transmission unit 3. For the selective transmission or discard described with reference to FIGS. 1 and 2, only one discard start / end point is set according to the amount of data in the buffer, and only one layer is selectively transmitted or discarded. However, two or more discard start / end points may be set, and two or more layers may be selectively transmitted or discarded. An example of the case where two discard start / end points are set to selectively transmit and discard two-layer packet data will be described. The flow identifier data of the IP packet data to be discarded when the point 1 is exceeded, the control code data for starting / ending the discard, and the discard start / end point 2 (discard start / end point 2> discard start / end point 1) The flow identifier data of the IP packet data to be discarded when exceeding and the control code data for starting / ending the discarding are accumulated and held.

  For example, when the amount of data in the buffer 4 exceeds the discard start / end point 1, the high frequency component (flow identifier data 0x1) of the B frame is set as the flow to be discarded, and the discard start / end point 2 is exceeded. An example in which the low frequency component (flow identifier data 0x2) of the B frame is set as a flow to be discarded in addition to the flow will be described. The packet discarding starts from packet data having control code data of 0xD, and for a discarding flow, a flag indicating that the discarding is in progress is accumulated and held in the selective transmission determining unit 31 and discarded. When the flag is set and the congestion of the buffer 4 is eliminated, transmission from the packet data having the control code data 0xD to the buffer 4 is started.

  That is, when the buffer 4 is congested and the amount of data in the buffer 4 exceeds the discard start / end point 1, discard is not started while packet data having a DS value of 0x1C is received, and an IP having 0x1D is received. After receiving the packet data, the high-frequency component of the B frame is discarded. When the amount of data in the buffer 4 exceeds the discard start / end point 2, the flow of the flow identifier data 0x1 continues to be discarded, and for the low frequency component of the B frame, packet data having a DS value of 0x2C is received. Discarding is not started during reception, but discarding is started after IP packet data having 0x2D is received. When the discard flag is set for the high frequency component of the B frame and the low frequency component of the B frame and the amount of data in the buffer 4 falls below the discard start / end point 2, the low frequency component of the B frame ( As for the flow identifier data 0x2), transmission from the packet data having the control code data 0xD to the buffer 4 is started. The high frequency component (flow identifier data 0x1) of the B frame is subsequently discarded. Similarly, when the discard flag is set for the high-frequency component of the B frame and the amount of data in the buffer 4 falls below the discard start / end point 1, packet data having a DS value of 0x1C is received. The selective transmission determination unit 31 controls the transfer unit 33 so that the discarding process is not terminated and the discarding process is terminated after receiving the IP packet data having 0x1D.

  FIG. 5 is a modification 110 of the packet data transfer apparatus 100 according to the embodiment of the present invention. Input ports 1 (1-1 to 1-N) can be made of ingress cards. Further, the selective transmission unit 13 (13-1 to 13-N), the buffer 14 (14-1 to 14-N) connected to each selective transmission unit, and the output port 5 (5 for connection to a plurality of output lines) -1 to 5-N) can also be made as an egress card 202 (202-1 to 202-N).

  The packet data transfer apparatus 100 in FIG. 1 monitors whether or not the amount of data in the buffer 4 exceeds a predetermined amount of data (data amount point in the buffer to start or end discarding of packet data) (signal 21). In the packet data transfer apparatus 110 in FIG. 5, the data amount in the buffer 14 is changed to the first data amount (packet data of the packet data). Whether or not there is a data amount point in the buffer to be discarded: hereinafter referred to as “discard start point”, or there is a second data amount (data amount point in the buffer in which packet data is to be discarded): The point at which the packet is transmitted or discarded by the selective transmission unit 13 is monitored at two points (signals 22 and 23) whether or not there are more or less (referred to as “discard end point”). Todeta is different from the transfer apparatus 100 and 110. However, the first data amount is larger than the second data amount (discard start point> discard end point). Hereinafter, the operation of the packet data transfer apparatus 110 will be described with reference to FIG.

The selective transmission unit 13 illustrated in FIG. 6 includes a selective transmission determination unit 131, a register 132 that temporarily stores data, and a transfer unit 133 that executes data transfer. The IP packet data routed by the routing unit 2 is temporarily held in the register 132, and the DS value in the IP header is analyzed by the selective transmission determination unit 131. In the selective transmission determination unit 131, the flow identifier data of IP packet data to be discarded when the amount of data in the buffer 14 is equal to or larger than the first predetermined amount (discard start point), the control code data for starting discard, and the buffer When the data amount in 14 is less than the second predetermined amount (discard end point), control code data for terminating discard is stored and held. Here, an example will be described in which the high frequency component (0x1) of the B frame is set as a flow to be discarded when the amount of data in the buffer 14 is equal to or greater than the discard start point.

The packet discarding starts from packet data having control code data of 0xD, and for a discarding flow, a flag indicating discarding is accumulated and held in the selective transmission determination unit 131 and discarded. When the flag is set and the amount of data in the buffer 14 is less than the discard end point, transmission from the packet data having the control code data 0xD to the buffer 14 is started. That is, even when the amount of data in the buffer 14 is equal to or greater than the discard start point, discard is not started while packet data having a DS value of 0x1C is received, and after IP packet data having 0x1D is received. Start disposal. Even when the discard flag is set and the amount of data in the buffer 14 is less than the discard end point, discarding is not completed while packet data having a DS value of 0x1C is received, and 0x1D is set. The selective transmission determination unit 131 controls the transfer unit 133 so as to end the discarding after receiving the received IP packet data. The transfer unit 133 performs an operation of transferring or discarding the data read from the register to the buffer under the control of the selective transmission determination unit 131.

  The above is the operation of the selective transmission unit 13. For the selective transmission or discard described with reference to FIGS. 5 and 6, only one discard start point and discard end point are set, and only one layer is selectively transmitted or discarded. Two or more discard start points and discard end points may be set, and two or more layers may be selectively transmitted or discarded. An example in which two sets of discard start points and discard end points are set to selectively transmit and discard two-layer packet data will be described. In the selective transmission determination unit selective transmission determination unit 131, the buffer 14 The flow identifier data of the IP packet data to be discarded and the control code data for starting the discarding, and the data amount in the buffer 14 are the second place when the data amount of the packet is equal to or larger than the first predetermined amount (discarding start point 1). Flow of control code data that terminates discard when it is less than a fixed amount (discard end point 1), and IP packet data that is discarded when the amount of data in the buffer 14 is equal to or greater than a third predetermined amount (discard start point 2) The identifier data and the control code data for starting the discard, and the control code for terminating the discard when the amount of data in the buffer 14 is less than the fourth predetermined amount (discard end point 2). Keep accumulating hold over data. However, the disposal start point 2> the disposal start point 1> the disposal end point 2> the disposal end point 1. Here, when the amount of data in the buffer 14 is greater than or equal to the discard start point 1, the high frequency component (0x1) of the B frame is set as the flow to be discarded. It is assumed that the frequency component (0x2) is set.

The packet discarding starts from packet data having control code data of 0xD, and for a discarding flow, a flag indicating discarding is accumulated and held in the selective transmission determination unit 131 and discarded. When the flag is set and the amount of data in the buffer 14 is less than the discard end point, transmission from the packet data having the control code data 0xD to the buffer 14 is started. That is, even when the amount of data in the buffer 14 is greater than or equal to the discard start point 1, discarding is not started while packet data having a DS value of 0x1C is received, and IP packet data having 0x1D is received. Starts to discard the high-frequency component (0x1) of the B frame. Similarly, when the amount of data in the buffer 14 is equal to or greater than the discard start point 2, while the packet data having the DS value of 0x2C is being received, the low frequency component (0x
The discard of the low frequency component (0x2) of the B frame is started after receiving the IP packet data having 0x2D without starting the discard of 2). Even when the discard flag is set for the low frequency component (0x2) of the B frame and the data amount in the buffer 14 is less than the discard end point 2, packet data having a DS value of 0x2C is received. The selective transmission determining unit 131 controls the transfer unit 133 so that the discard of the low-frequency component (0x2) of the B frame is not terminated while the IP packet data having 0x2D is received and the discard is terminated. I do. During this time, the high-frequency component (0x1) of the B frame continues to be discarded. When the discard flag is set for the high-frequency component of the B frame, packet data having a DS value of 0x1C is received even when the amount of data in the buffer 14 further decreases and becomes less than the discard end point 1. The transfer unit 133 is controlled so that the discarding is not finished while the IP packet data having 0x1D is received and the discarding is finished.

  The packet data transfer apparatus described with reference to FIG. 1 or FIG. 5 is an apparatus that selectively transmits or discards packet data when the transmission network is congested. Is an apparatus that performs data shaping when a part of the hierarchical program data described with reference to FIG. 3 is randomly discarded and transmitted. Data shaping will be described by taking, for example, a case where the data 40 of FIG.

Originally, data 40 distributed from the program distribution server includes IP packet data composed of data 48, 46, 42, and 43-1, IP packet data composed of data 49 and 43-2, and data The IP packet data composed of 50, 47, 44, and 45 should be transmitted through the transmission network, but a part of the data, for example, data 48, 46, 42, and 43-1 due to congestion in the transmission network. Is discarded in the middle of transmission, the IP packet data composed of the remaining data 49 and 43-2 and 50, 47, 44, and 45 are composed to the receiving device. IP packet data is transmitted. The receiving device can reproduce the high-frequency component of the I frame when all the data constituting the data 40 is received, but cannot reproduce it when only a part of the data can be received. Therefore, since the data that is wasted even if it is transmitted to the receiving device is transmitted, the transmission network is effectively used by discarding the data that is wasted even if it is transmitted. In the data shaping server, discarding the IP packet data composed of the data 49 and 43-2 and the IP packet data composed of the data 50, 47, 44 and 45 means data shaping. . Next, the configuration and operation of the data shaping server 120 will be described with reference to FIG.

  The data shaping server 120 shown in FIG. 7 includes a CPU unit 122 connected by a bus 121, a program accumulation unit 123 that accumulates a data shaping program that operates on the CPU unit 122, and accumulation of data necessary during data shaping processing. It comprises a main memory 124 for holding data, an input unit 125 for receiving data, and an output unit 126 for outputting the shaped data. The data shaping server 120 first loads the processing program stored and held in the program storage unit 123 into the CPU unit 122 and starts data shaping processing. The detailed processing operation of the CPU unit 122 will be described later with reference to FIG.

  When the data shaping processing operation is started, the input of data is accepted from the input unit 125, and the continuity of sequence numbers assigned to each layer, which is the layer header of the input data shown in FIG. When discontinuity is detected, invalid data discarding (data shaping) operation processing is performed as described above. Specifically, the UDP data is reconstructed for each layer with respect to the received data to monitor whether or not the UDP packet data is completed, and the incomplete data is discarded. For data that has been completed as UDP data, the sequence number of each layer (flow identifier data 0x6 to 0x1) is monitored to check whether the data is continuous. The UDP data having the control code data value 0xC is discarded until the UDP packet data having the control code data value 0xD indicating the head of the frame is created. After the UDP packet data having the control code data value 0xD is created, the received data of each layer is redistributed as IP packet data from the output unit 126. The above is the schematic processing operation of the data shaping server. Next, a detailed processing operation performed by the CPU unit 122 of the data shaping server will be described with reference to FIG.

FIG. 8 shows a flowchart of processing executed by the CPU unit 122 of the data shaping server 120.
(1) When the data shaping server 120 is powered on, the data shaping processing program stored and held in the program storage unit 123 is loaded into the CPU unit 122, and the data shaping processing operation is started (step 200).
(2) When the data shaping processing operation starts, UDP data is created from the IP packet data received from the input unit 125, and it is determined whether or not the data is completed as UDP data (steps 201 and 202).
(3) If it is determined in step 202 that a part of the UDP data is discarded and the UDP data cannot be reconstructed, the incomplete UDP data is discarded (steps 202 and 203).
(4) If it is determined in step 202 that the UDP data can be reconstructed, the UDP data is reconstructed and the flow to be subsequently processed (for example, flow identifier data 0x6 to 0x1 in the hierarchy header) It is determined whether or not there is (steps 202 and 204).
(5) If it is determined in step 204 that the flow is not to be processed, IP packet data is redistributed from the output unit 126 (steps 204 and 205).
(6) If it is determined in step 204 that the flow should be processed, the sequence number value in the layer header is subsequently acquired, and it is determined whether each layer is continuous (steps 204 and 206). 207).
(7) If it is determined in step 207 that the sequence number is discontinuous, a discard setting is made for the flow of the discontinuous layer, and then the control code data value in the layer header is acquired (steps 207 and 208). 209). If the discard setting has already been performed, the discard setting is continued.
(8) If it is determined in step 207 that the sequence numbers are consecutive, the control code data value in the layer header is acquired (steps 207 and 209).
(9) After acquiring the control code data value, it is determined whether or not the control code data value is in the middle (0xC) of the layer frame data (step 210).
(10) If it is determined in step 210 that the control code data value is in the middle of the layer frame data (0xC), it is subsequently determined whether the data is to be discarded (steps 210 and 211).
(11) If it is determined in step 211 that the data is not to be discarded, it is redistributed as IP packet data from the output unit 126 (steps 211 and 212).
(12) If it is determined in step 211 that the data is to be discarded, the data is discarded and the creation of the next UDP data is started (steps 211, 213, 201).
(13) If it is determined in step 210 that the control code data value is not in the middle of the layer frame data (0xC), it is subsequently determined whether or not the control code data value is the head (0xD) of the layer frame data. (Steps 210 and 214).
(14) If it is determined in step 214 that the control code data value is not the head (0xD) of the layered frame data, it is redistributed as IP packet data from the output unit 126 (steps 214 and 212).
(15) If it is determined in step 214 that the control code data value is the top (0xD) of the layer frame data, the discard process for the corresponding layer frame data is terminated, and the output unit 126 re-transmits it as IP packet data. Distribute (steps 214, 215, 212). If the discard setting is not performed, the redistribution operation is continued. The above is the flow of processing executed by the CPU unit 122 of the data shaping server 120.

  The data shaping server shown in FIG. 7 is a device that performs data shaping and redistribution when a part of the hierarchical program data shown in FIG. 3 is randomly discarded and transmitted. The reception / playback apparatus shown is a device that shapes received data, performs decoding by a decoding unit (decoder), and performs playback by a display / playback unit. Regarding data shaping, basically the same processing as the data shaping server is performed.

  9 includes a CPU unit 142 connected by a bus 141, a program storage unit 143 that stores a data shaping program that operates on the CPU unit 142, and storage of data necessary during data shaping processing. A main memory 144 for holding data, an input unit 145 for receiving data, a decoding unit 146 for decoding the shaped data, and a display reproduction unit 147 for reproducing and displaying the decoded signal Consists of. First, the reception / playback apparatus 140 loads the processing program stored and held in the program storage unit 143 to the CPU unit 142 and starts a data reception / playback operation. The detailed processing operation of the CPU unit 142 will be described later with reference to FIG.

  When the data reception / reproduction operation is started, data input is accepted from the input unit 145, and the continuity of the sequence number assigned to each layer, which is the layer header of the input data shown in FIG. When discontinuity is detected, invalid data discarding (data shaping) operation processing is performed as described above. Specifically, the UDP data is reconstructed for each layer with respect to the received data to monitor whether or not the UDP packet data is completed, and the incomplete data is discarded. For data that has been completed as UDP data, the sequence number of each layer (flow identifier data 0x6 to 0x1) is monitored to check whether the data is continuous. The UDP data having the control code data value 0xC is discarded until the UDP packet data having the control code data value 0xD indicating the head of the frame is created. After the UDP packet data having the control code data value 0xD is created, the received hierarchical program data is decoded by the decoding unit 146, and the decoded data is displayed and reproduced by the display reproduction unit 147. The above is the schematic processing operation of the reception / playback apparatus.

Next, a detailed processing operation performed by the CPU unit 142 of the reception / playback apparatus will be described with reference to FIG. FIG. 10 shows a flowchart of processing executed by the CPU unit 142 of the reception / playback apparatus 140.
(1) When the power of the reception / playback device 140 is turned on, the reception / playback processing program stored and held in the program storage unit 143 is loaded into the CPU unit 142 and the reception / playback processing operation is started (step 250).
(2) When the reception / reproduction processing operation is started, UDP data is created from the IP packet data received from the input unit 125, and it is determined whether or not the UDP data is completed (steps 251 and 252).
(3) If it is determined in step 252 that a part of the UDP data is discarded and the UDP data cannot be reconstructed, the UDP data that is not completed is discarded (steps 252 and 253).
(4) If it is determined in step 252 that the UDP data can be reconstructed, the UDP data is reconstructed, and the flow to be subsequently processed (for example, the flow identifier data 0x6 to 0x1 in the hierarchy header) It is determined whether or not there is (steps 252 and 254).
(5) If it is determined in step 254 that the flow is not to be processed, the data is discarded (steps 254 and 255).
(6) If it is determined in step 254 that the flow should be processed, the sequence number value in the layer header is subsequently acquired, and it is determined whether each layer is continuous (steps 254, 256). 257).
(7) If it is determined in step 257 that the sequence number is discontinuous, a discard setting is made for the flow in the discontinuous layer, and then the control code data value in the layer header is acquired (steps 257 and 258). 259). If the discard setting has already been performed, the discard setting is continued.
(8) If the sequence number is consecutive in the determination in step 257, the control code data value in the layer header is subsequently acquired (steps 257 and 259).
(9) After acquiring the control code data value, it is determined whether or not the control code data value is in the middle (0xC) of the hierarchical frame data (step 260).
(10) If it is determined in step 260 that the control code data value is in the middle of the layer frame data (0xC), it is subsequently determined whether the data is to be discarded (steps 260 and 261).
(11) If it is determined in step 261 that the data is not to be discarded, hierarchical program data is created (steps 261 and 266).
(12) If it is determined in step 261 that the data should be discarded, the data is discarded and the creation of the next UDP data is started (steps 261, 263, 251).
(13) If it is determined in step 260 that the control code data value is not in the middle of the layer frame data (0xC), it is subsequently determined whether or not the control code data value is the head (0xD) of the layer frame data. (Steps 260 and 264).
(14) If it is determined in step 264 that the control code data value is not the head (0xD) of the layer frame data, a data discarding operation is performed, and creation of the next UDP data is started (steps 264, 266, 251).
(15) If it is determined in step 264 that the control code data value is the head (0xD) of the layer frame data, the discard process of the corresponding layer frame data is terminated and layer program data is created (step) 264, 265, 267). If the discard setting is not performed, the hierarchical program data creation operation is continued.
(16) After creating the hierarchical program data, the decoding unit 146 decodes the hierarchical program data, and the display reproduction unit 147 displays and reproduces the decoded data.
The above is the flow of processing executed by the CPU unit 142 of the reception / playback apparatus.

  FIG. 11 shows a modification 150 of the packet data transfer apparatus 100 shown in FIG. The packet data transfer apparatus 100 in FIG. 1 monitors whether or not the amount of data in the buffer 4 exceeds a predetermined amount of data (data amount point in the buffer to start or end discarding of packet data) (signal 21). The operation of selecting whether to transmit or discard the packet data in the selective transmission unit 3 has been described. However, in the packet data transfer apparatus 150 according to the present embodiment, the packet data is selectively transmitted and discarded. In addition, a data duplication method for multicast distribution of data will be described.

  The packet data transfer apparatus 150 shown in FIG. 11 has basically the same configuration as that of the packet data transfer apparatus 100 shown in FIG. 1 except that the input port 1 (1−1) for connecting to a plurality of input lines is different. 1 to 1-N) and a routing unit 2 for routing packet data, a duplicate transmission unit 7 (7-1 to 7-N) for creating data for multicast is provided. Hereinafter, the operation of the packet data transfer apparatus 150 will be described focusing on the operation different from that of the packet data transfer apparatus 100 of FIG. The transfer device 150 first transmits the IP packet data shown in FIG. 3, for example, IP packet data composed of the data 48, 46, 42, and 43-1, to the plurality of input ports 1 (1-1 to 1-1. N). The IP packet data received from a plurality of input ports is transferred to the copy transfer unit 7 (7-1 to 7-N), and the packet data is copied if necessary, and the IP packet header is passed through the routing unit 2. Route to a predetermined output port (in this case, one of the selective transmission units 3-1 to 3-N). Since the processing operation after the routing is the same as the operation of the packet data transfer apparatus 100 of FIG. 1, the detailed configuration and operation of the duplicate transmission unit 7 will be described below with reference to FIG.

FIG. 12 is a diagram showing the configuration of the duplicate transmission unit 7 in detail. The duplicate transmission unit 7 includes a duplicate transmission determination unit 71, a register 72 that temporarily holds data, and a duplicate creation unit 73 that creates a duplicate of data. The IP packet data received from the input port 1 is temporarily held in the register 72, and the DS value in the IP header is analyzed by the duplicate transmission determination unit 71. The duplicate transmission determination unit 71 stores and holds in advance the flow identifier data of IP packet data to be duplicated at the time of multicast distribution, the control code data for starting the duplication, and the control code data for completing the duplication operation. Here, a description will be given taking a high-frequency component (flow identifier data 0x5) of an I frame as an example of a flow to be copied. In addition, packet data replication creation starts from packet data having control code data of 0xD, and a flag indicating that replication is in progress is accumulated in the duplicate transmission determination unit 71 for a flow that is being duplicated and distributed. When it is no longer necessary to create a replica, the specification is such that the replica creation ends from the packet data having the control code data 0xD. That is, when creating a replica, the replica creation is not started while the packet data having 0x5C as the DS value is received, and the replica creation is started after receiving the IP packet data having 0x5D. In addition, when the copy creation flag is set and it is no longer necessary to create a copy, the copy creation is not completed while packet data having a DS value of 0x5C is received, and IP packet data having 0x5D is received. The copy transmission determination unit 71 controls the copy creation unit 73 so that the copy creation is terminated after the reception of the copy. The above is the configuration and operation of the duplicate transmission unit 7.

  FIG. 13 is a diagram showing a configuration of the hierarchical program data distribution apparatus shown in FIG. The program distribution apparatus 150 shown in FIG. 13 includes a CPU unit 152 connected by a bus 151, a program storage unit 153 that stores a program distribution program that operates on the CPU unit 152, and storage of data necessary during the program distribution process. It comprises a main memory 154 that holds data, a data storage unit 155 that stores and holds hierarchical program data, and an output unit 156 that distributes packetized data.

  The program distribution device 150 first loads the processing program stored and held in the program storage unit 153 into the CPU unit 152, and starts a packet data distribution operation. The detailed distribution operation of the CPU unit 152 will be described later with reference to FIG. When the distribution operation of the hierarchical program data is started, the hierarchical program data stored and held in the data storage unit 155 is read into the main memory unit 154 and divided into a predetermined size for each layer, and each divided data is described with reference to FIG. Hierarchical packet data is created with a hierarchical header attached. Further, UDP packet data is created by adding a UDP header to the data converted into hierarchical packet data, further converted into IP packet data, and distributed from the output unit 156.

FIG. 14 shows a flowchart of the distribution operation executed by the CPU unit 152 of the program distribution device 150.
(1) When the power of the program distribution device 150 is turned on, the program distribution program stored and held in the program storage unit 153 is loaded into the CPU unit 153, and the hierarchical data distribution operation is started (step 300).
(2) When the distribution operation of hierarchical data is started, the hierarchical program data is read from the data storage unit 155 that stores the hierarchical data, and it is determined whether the hierarchical video data has been read up to the frame break (steps 301 and 302).
(3) If it is determined in step 302 that the frame break has not been read, the process returns to step 301 and data is read again (steps 302 and 301).
(4) If it is determined in step 302 that the data has been read up to the frame break, the read data is divided by a predetermined size, for example, every 4 kbytes (steps 302 and 303).
(5) After division into a predetermined size, the hierarchy header described in FIG. 3 is attached to each hierarchy data (step 304).
(6) UDP packet data is configured by adding a UDP header to the layer packet data to which the layer header is added (step 305).
(7) The UDP packet data is divided, and the IP packet data is configured by adding the IP header described in FIG. 3 to each divided data (step 306).
(8) The data converted into IP packet data is distributed from the output unit 156 (step 307).

  The above is the program distribution operation executed by the CPU unit 152 of the program distribution device 150. In the present embodiment, distribution of hierarchical data using MPEG-1 / 2 encoding as shown in FIG. 3 has been described as an example. However, MPEG-4 encoded data is used as an encoding method. It may be used. In MPEG-4 encoding, since encoding is performed in units of objects, selective data transmission performed in MPEG-1 / 2 is performed by assigning flow identifier data and control code data to each object and distributing IP packets. / Disposal and duplicate transmission are possible.

  FIG. 15 shows a configuration in which IP packet data is converted into MPLS (Multi-Protocol Label Switching) packet data, the converted MPLS packet data is transmitted through the MPLS network, and is further converted into IP packet data for transmission. Yes. In FIG. 15, areas 75 and 77 indicate the IP network, and area 76 indicates the MPLS network area. In the network in which the IP network and the MPLS network shown in FIG. 15 are interconnected, the IP packet data 71 is converted into the MPLS packet data 72 by the data type conversion device A 160 and is transmitted by the MPLS network 76. The MPLS packet data 73 is converted into IP packet data 74 by the data type conversion apparatus B 170 and transmitted through the IP network 77.

  FIG. 16 shows correspondence between IP packet data and MPLS packet data for transmitting transmission data in a network in which the IP network and the MPLS network described in FIG. 15 are interconnected. A method for converting IP packet data and MPLS packet data will be described later with reference to FIGS. As the IP packet data 82 shown in FIG. 16, the MPEG-1 / 2 data described in FIG.

  When the IP packet data transmits the MPEG-1 / 2 data, the value shown in FIG. 4 can be used as the DS value 80 in the IP packet header 81. When the IP packet data (configured from data 81 and 82) configured in this way is converted into MPLS packet data, it can be configured by adding an MPLS header 90 (32 bits) to each IP packet data. is there. The MPLS header 90 includes a label 91 (20 bits) for identifying data, an experimental field 92 (3 bits), 93 (1 bit) indicating the end of the stack, TTL ( Time to Live) 94 (8 bits). Here, for example, a fixed value 7 (binary 111) is used as the experimental field 92 value, the value of the stack field 93 is 0 (binary 0) in the first case of transmission data, and the last. In this case, 1 (1 in binary system) is assigned, and as the value of TTL94, for example, a value such as 15 (000011111 in binary system) can be used. As the value of the label value 91, as shown in FIG. 17, by adding a 12-bit conversion pad (binary 111111111111) to the upper bits of the DS value in the IP packet header, the DS value can be easily set. Can create a label value of MPLS packet data.

  FIG. 18 shows a configuration of a data type conversion apparatus A that converts 71 IP packet data into 72 MPLS packet data. FIG. 19 is a flowchart of a data type conversion program executed by the CPU unit 162 of the data type conversion apparatus A in FIG.

  A data type conversion apparatus 160 shown in FIG. 18 includes a CPU unit 162 connected by a bus 161, a program storage unit 163 that stores a data type conversion program that operates in the CPU unit 162, and a data type conversion process. It comprises a main memory 164 that stores and holds necessary data, an input unit 165 that receives IP packet data, and an output unit 166 that outputs MPLS packet data. The data type conversion apparatus A 160 first loads the processing program stored and held in the program storage unit 163 into the CPU unit 162, and starts the data type conversion processing operation. The detailed distribution operation of the CPU unit 162 will be described later with reference to FIG. When the data type conversion processing operation is started, IP packet data 71 is received from the input unit 165 and temporarily stored in the main memory 164. An MPLS header is created with reference to the DS value in the header of the received IP packet data. The created MPLS header is attached to the head of the IP packet data to create MPLS packet data, and 72 MPLS packet data are output from the output unit 166.

FIG. 19 shows a flowchart of the data type conversion processing operation executed by the CPU unit 162 of the data type conversion apparatus A 160.
(1) When the data type conversion apparatus A is powered on, the data type conversion processing program stored and held in the program storage unit 163 is loaded into the CPU unit 162, and the data type conversion processing operation is started (steps 350 and 351). ).
(2) When the data type conversion processing operation is started, reception of IP packet data is started from the input unit 165, and it is determined whether or not IP packet data has been received (step 351).
(3) If it is determined in step 351 that IP packet data has not been received, the IP packet data reception waiting operation is continued (step 351).
(4) If it is determined in step 351 that IP packet data is received, the MPLS packet header described with reference to FIGS. 16 and 17 is created from the DS value in the IP packet header (step 352).
(5) After creating the MPLS packet header, the MPLS packet header is added to the head of the IP packet data, and the MPLS packet data is output from the output unit 166 (step 353). The data type conversion processing operation executed by the CPU unit 162 of the 160 data type conversion apparatus A has been described above.

  FIG. 20 shows the configuration of a data type conversion apparatus B that converts 73 MPLS packet data into 74 IP packet data. FIG. 21 is a flowchart of a data type conversion program executed by the CPU unit 172 of the data type conversion apparatus B in FIG.

A data type conversion device 170 shown in FIG. 20 includes a CPU unit 172 connected by a bus 171, a program storage unit 173 that stores a data type conversion program that operates on the CPU unit 172, and a data type conversion process. The main memory 174 stores and holds necessary data, an input unit 175 that receives MPLS packet data, and an output unit 176 that outputs IP packet data. First, the data type conversion apparatus 170 loads the processing program stored and held in the program storage unit 173 to the CPU unit 172, and starts the data type conversion processing operation. The detailed distribution operation of the CPU unit 172 will be described later with reference to FIG. When the data type conversion processing operation is started, the MPLS packet data 73 is received from the input unit 175 and temporarily stored in the main memory 174. Received MPL
The header data of the S packet data is deleted, IP packet data is created, and the output unit 176 outputs 74 IP packet data.

FIG. 21 shows a flowchart of the data type conversion processing operation executed by the CPU unit 172 of the data type conversion apparatus B 170.
(1) When the data type conversion device B is powered on, the data type conversion processing program stored and held in the program storage unit 173 is loaded into the CPU unit 172, and the data type conversion processing operation is started (steps 360 and 361). ).
(2) When the data type conversion processing operation is started, reception of MPLS packet data is started from the input unit 175, and it is determined whether or not MPLS packet data is received (step 361).
(3) If it is determined in step 361 that MPLS packet data has not been received, the MPLS packet data reception waiting operation is continued (step 361).
(4) If it is determined in step 361 that MPLS packet data is received, the header data of the MPLS packet data is deleted, and IP packet data is created (step 362).
(5) After creating the IP packet data, the output unit 176 outputs the IP packet data (step 363). The above is the data type conversion processing operation executed by the CPU unit 172 of the 170 data type conversion apparatus B.

FIG. 22 is a diagram illustrating a configuration of an MPLS packet data transfer device 500 according to an embodiment of the present invention. Specifically, input ports 501 (501-1 to 501 -N) for connecting to a plurality of input lines, a routing unit 502 for routing MPLS packet data, and data routed for a plurality of output lines To connect a selective transmission unit 503 (503-1 to 503-N) for selective transmission, a buffer 504 (504-1 to 504-N) connected to each selective transmission unit, and a plurality of output lines Output port 505 (505-1 to 505-N). The input ports 501 (501-1 to 501 -N) can be created by ingress cards. Further, the selective transmission unit 503 (503-1 to 503-N), the buffer 504 (504-1 to 504-N) connected to each selective transmission unit, and the output port 505 (505) for connecting to a plurality of output lines. -1 to 505-N) can be made as the egress card 205 (205-1 to 205-N).

  The MPLS packet data transfer apparatus 500 first receives MPLS packet data shown in FIG. 16, for example, data having a value shown in FIG. 17 as a label value from a plurality of input ports 501 (501-1 to 501-N). To do. The received MPLS packet data is routed to a predetermined output port (here, any one of the selective transmission units 503-1 to 503-N) through the routing unit 502 according to the label value. The MPLS packet data transferred to the selection data unit 503 (503-1 to 503-N) is transmitted to the selection transmission unit 503 (503-1 to 503-1) by the data amount signal 521 (521-1 to 521-N) of the buffer connected to the subsequent stage. 503-N), it is determined whether or not to transmit to the buffer 504 (504-1 to 504-N).

  In the MPLS packet data transfer apparatus 500 of FIG. 22, the data amount in the buffer 504 (504-1 to 504-N) is a predetermined data amount (data amount point in the buffer to start or end discarding of MPLS packet data). ) Shows the case of monitoring whether or not there is more than the above, and a signal 521 (521-1 to 521-) indicates whether or not the amount of data in the buffer 504 (504-1 to 504 -N) is a predetermined amount or more. N), the selected transmission unit 503 (503-1 to 503-N) is notified. Thereby, it is determined whether or not the received MPLS packet data is transmitted from the selective transmission unit 503 (503-1 to 503-N) to the buffer 504 (504-1 to 504-N). In the case of transmission, MPLS packet data is transmitted from the selective transmission unit 503 (503-1 to 503-N) to the buffer 504 (504-1 to 504-N). When not transmitting (discarding), MPLS packet data is not transmitted from the selective transmission unit 503 (503-1 to 503-N) to the buffer 504 (504-1 to 504-N), and the selective transmission unit 503 (503-503) is transmitted. 1-503-N). The operation of the selective transmission unit 503 (503-1 to 503-N) will be described later with reference to FIG. The MPLS packet data transmitted into the buffer 504 (504-1 to 504-N) is distributed from the output port 505 (505-1 to 505-N) to the output line.

  FIG. 23 is a diagram illustrating the configuration of the selective transmission unit 503 in detail. The selective transmission unit 503 includes a selective transmission determination unit 531, a register 532 that temporarily stores data, and a transfer unit 533 that executes data transfer. The MPLS packet data routed from the routing unit 502 is temporarily held in the register 532, and the selection transmission determination unit 531 analyzes the label value in the MPLS header. The selective transmission determination unit 531 previously stores and holds the label value of MPLS packet data to be discarded when the data amount in the buffer 504 is equal to or larger than a predetermined amount (during congestion).

  Here, an example will be described in which the high frequency component of the B frame (the flow identifier data in the label is 0x1) is set as a flow to be discarded when the buffer 504 is congested. In addition, MPLS packet discarding starts from MPLS packet data whose control code data in the label has 0xD. For a discarding flow, a flag indicating that discarding is in progress is made in the selective transmission determination unit 531. If the discard flag is set and the congestion of the buffer 504 is eliminated, transmission from the packet data having the control code data 0xD in the label to the buffer 504 is started. That is, when the buffer 504 is congested, discarding is not started while MPLS packet data having 0xFFF1C as a label value is received, and discarding is started after receiving MPLS packet data having 0xFFF1D as a label value. Also, when the discard flag is set and the congestion of the buffer 504 is eliminated, the discard is not finished while the packet data having 0xFFF1C as the label value is received, and the MPLS packet data having 0xFFF1D is received. Then, the selective transmission determination unit 531 controls the transfer unit 533 so that the discarding is completed. The transfer unit 533 performs the operation of transferring the MPLS packet data read from the register 532 to the buffer 504 or discarding it under the control of the selective transmission determination unit 531.

  The above is the configuration and operation of the selective transmission unit 503. For the selective transmission or discard described with reference to FIGS. 22 and 23, only one discard start / end point is set according to the amount of data in the buffer, and only one layer is selectively transmitted or discarded. However, two or more discard start / end points may be set, and two or more layers may be selectively transmitted or discarded. Further, depending on the amount of data in the buffer, the discard start point and the discard end point may be set individually and selectively transmitted or discarded. The above is the configuration and processing operation of the MPLS packet data transfer apparatus 500 shown in FIGS.

  As described above, for a data flow having a predetermined identifier when the transmission network is congested, discarding is started from packet data having predetermined control code data, and discarding is started from packet data having predetermined control code data. The selective packet transmission / discard method to be completed can be realized in an ATM (Asynchronous Transfer Mode) network or an IPv6 (Internet Protocol Version 6) network.

  In the ATM network, data transmission is performed using 53-byte ATM cells. The 53-byte ATM cell is composed of a 5-byte ATM cell header and 48-byte data. The 5-byte cell header includes VPI (Virtual Path Identifier) and VCI (Virtual Channel Identifier) as fields for identifying a flow, and PT (Payload Type) indicating a data type. Thus, selective transmission / discarding can be realized by mapping a predetermined identifier to VPI and VCI and using PT as the predetermined identifier. Specifically, each layer of data composed of a plurality of layers is mapped to VPI and VCI. For example, an ATM PT that transmits the top data of a frame of MPEG data and a PT that transmits data in the middle of the frame are different values. When congestion occurs in the ATM switch, discarding is started from the ATM cell having the PT that transmits the head data of the video frame, and after the discarding starts, the head data of the video frame is transmitted. By terminating the discarding from the ATM cell having the PT to be performed, the selective packet transmission / discarding function realized in the IP network or the MPLS network can be realized.

  The IPv6 network is the next generation network of the IP (exactly IPv4) network, and the basic IPv6 header has a 3-byte flow label area. Therefore, the 1-byte DS field value used in the IPv4 network is used. By adding a 2-byte conversion pad (for example, 0xFFFF) to the head of DS, the DS field can be easily associated with the flow label, and the selective packet transmission / discard function can be realized.

The block diagram which shows an example of the packet data transfer apparatus by this invention. The block diagram of the selection transmission part of the packet data transfer apparatus shown in FIG. The figure which shows the structural example of the packet data by this invention. The figure which shows the relationship between the transmission data by this invention, and DS (Differentiated Services) field value (flow identifier data, control code data). The block diagram which shows the other example of the packet data transfer apparatus by this invention. FIG. 6 is a block diagram of a selective transmission unit of the packet data transfer apparatus shown in FIG. 5. The figure which shows the structural example of the data shaping server by embodiment of this invention. The flowchart which shows the operation example of the data shaping server by this invention. The figure which shows the structural example of the reception reproducing | regenerating apparatus by this invention. The flowchart which shows the operation example of the reception reproducing | regenerating apparatus by this invention. The block diagram which shows the other example of the packet data transfer apparatus by this invention. The block diagram of the duplication transmission part of the packet data transfer apparatus shown in FIG. The figure which shows the structural example of the program data delivery apparatus by this invention. The flowchart which shows the operation example of the program data delivery apparatus by this invention. The figure explaining the transmission network which interconnected the IP network and the MPLS network. The figure which shows the example of a response | compatibility with IP packet data and MPLS packet data by this invention. The figure which shows the label data structural example of MPLS (Multi-Protocol Label Switching) packet data by this invention. The figure which shows the structural example of the data type conversion apparatus by this invention. The flowchart which shows the operation example of the data type conversion apparatus shown in FIG. The figure which shows the other structural example of the data type conversion apparatus by this invention. The flowchart which shows the operation example of the other example of the data type conversion apparatus shown in FIG. The block diagram of the other example of the packet data transfer apparatus by this invention. The block diagram of the selective transmission part of the other example of the packet data transfer apparatus shown in FIG. Explanatory drawing of the video data transmission in the router by this invention. Explanatory drawing of the video data transmission in the conventional router.

Explanation of symbols

  1: input port, 2: routing unit, 3: selective transmission unit, 4: buffer, 5: output port, 21: data amount signal, 31: selective transmission determination unit, 32: register, 33: transfer unit, 100, 110 : Packet data transfer device, 120: data shaping server, 140: reception / playback device, 150: packet data transfer device, 160: data type conversion device, 170: data type conversion device, 500: MPLS packet data transfer device

Claims (2)

  Holds flow identifier data for identifying data to be duplicated and control code data for controlling duplication processing, receives packet data to which the flow identifier data and control code data are added, and identifies the data A packet data duplication distribution method, wherein when duplicating packet data, duplication start and duplication are performed based on the control code data for packet data having the retained flow identifier data.   Means for holding flow identifier data for identifying data to be duplicated and control code data for controlling duplication processing; and means for receiving packet data to which the flow identifier data and control code data are attached And packet data having means for starting and ending replication based on the control code data for the packet data having the retained flow identifier data when the packet data is replicated Replication distribution device.

Images