WO2006011193A1 - Data transmitting apparatus - Google Patents

Abstract

There are a frame division part (11) for dividing data on a frame into a predetermined size of data for each of connections to write the divided data in designated positions of a transmission buffer; a data reading part (13) for reading the divided data from the designated positions of the transmission buffer; a scheduler part (17) for performing a scheduling on a frame-by-frame basis, when a forefront divided data constituting a frame of a particular connection is selected in a round robin manner, such that the output of that frame is given priority until the last divided data constituting the same frame is outputted; and a buffer management part (18) for designating a write position for the frame division part (11) and further designating, in accordance with the decision by the scheduler part (17), a read position for the data reading part (13).

Description

 Specification

 Data transmission device

 Technical field

 TECHNICAL FIELD [0001] The present invention relates to a data transmission apparatus that performs data transfer in consideration of a frame of an upper protocol, and more specifically, while maintaining fairness of data transmission in each connection, The present invention relates to a data transmission device that can shorten the transmission time of the network.

 Background art

 Hereinafter, a conventional communication device on the data transmission side (hereinafter referred to as a data transmission device) will be described. For example, when transmitting and receiving data between communication devices, it has been pointed out that communication quality deteriorates due to an increase in transfer delay time.To reduce this transfer delay time, various solutions have been developed by devising transmission methods. A method has been proposed.

 [0003] Specifically, in a conventional data transmission apparatus, a high-priority buffer for real-time packets and a low-priority buffer for non-real-time packets are used to shorten the packet transfer delay time for which real-time performance is required. And the corresponding packets are stored in each. For example, non-real-time packets are divided and transmitted so that the transmission path is not occupied for a long time, thereby reducing the transfer delay time of real-time packets (high priority data).

 [0004] However, this method has a problem in that non-real-time packets are always divided, so that when there is no real-time packet, the transfer efficiency of non-real-time packets is reduced. Therefore, in Patent Document 1 below, in order to solve this problem, non-real-time packet transfer efficiency is improved by not dividing non-real-time packets when a real-time packet connection is not set. I'm going.

 Patent Document 1: Japanese Patent Laid-Open No. 2002-158702 FIG. 1

 Disclosure of the invention

 Problems to be solved by the invention

However, even if the method described in Patent Document 1 is used, there are many connections. It is difficult to reduce the frame-by-frame transfer delay time of each connection while maintaining fairness of transfer of each connection. In other words, in the conventional data transmission device, if there are many connections, if data transmission is performed while maintaining the fairness of each connection, the time from the start of frame transfer to the end of transfer of the frame becomes longer. There was a problem.

 [0007] Specifically, for example, there are connection power (# 1 1 #n), a transmission queue for each connection is provided in the data transmission device, and several transfer packets are provided for all connections. Consider the case of accumulation. It is also assumed that the frame data of the upper protocol of each connection is divided into three and three packets (A, B, C) are generated.

 [0008] Generally, when a data transmission opportunity is given fairly to each connection, a round robin method is adopted. At this time, in connection # 1, it is necessary to wait for the time required for output for the number of connections from the output of the first packet A1 composing a single frame to the output of B1 of the next packet. That is, if three packets of connection # 1, Al, Bl, and C1, constitute one frame of the upper protocol, 2n pieces of data are required to complete transmission of one frame of the upper protocol of connection # 1. A waiting time equivalent to the time required for packet output is required.

 [0009] On the other hand, the upper protocol frame does not make sense as data unless all the data reaches the receiving side. For example, even if only a part of the upper protocol frame arrives, If the data does not reach the receiving side, all frame data of the higher-level protocol will be discarded. Therefore, in the data transmission device, for example, considering the worst condition such as when the transmission path is at a low rate, when the number of users is large, or when data transfer is postponed by the transfer of best-f auto data, Data transmission must be scheduled so that it is not discarded by the receiving device.

[0010] The present invention has been made in view of the above, and even when there are many connections, the transfer time for each frame of the upper protocol is shortened while maintaining the fairness of each connection, and the frame unit The purpose of the present invention is to provide a data transmission device capable of improving the reception probability of data. Means for solving the problem

In order to solve the above-described problems and achieve the object, the data transmission apparatus according to the present invention performs data transmission on a frame basis while maintaining fairness of output in a plurality of connections. A data transmission device that divides data on a frame into a predetermined size for each connection, writes the divided data at a specified position in a transmission buffer common to all connections, and simultaneously determines the type of divided data ( Frame dividing means (corresponding to a frame dividing section 11 in the embodiment described later) for generating frame information including the head, continuation, and final) and data reading means (data reading) for reading divided data at a specified position from the transmission buffer And the first divided data constituting the frame of the specific connection is selected by round robin. Thereafter, scheduling means (corresponding to the scheduler unit 17) that performs scheduling in units of frames so that the output of the frame is prioritized up to the final divided data constituting the same frame, and data writing to the transmission buffer A buffer that manages the position, data read position, and free space, designates the write position for the frame dividing means, and further designates the read position for the data read means according to the determination by the scheduling means And a management means (corresponding to the buffer management unit 18).

 The invention's effect

 [0012] In the present invention, for example, when divided data of a frame of a specific connection is selected by round robin, each frame is scheduled so as to give priority to output of divided data constituting the frame. While maintaining the fairness of the data output in the connection, the transmission time of each frame can be shortened.

 Brief Description of Drawings

 FIG. 1 is a diagram showing a configuration of a first embodiment of a data transmission device according to the present invention.

FIG. 2 is a diagram showing an outline of operation of the data transmitting apparatus.

FIG. 3 is a diagram showing how to use a penalty. [FIG. 4] FIG. 4 is a diagram for explaining processing when a frame of a specific connection is longer than a predetermined reference.

 FIG. 5 is a diagram for explaining processing when a plurality of connections are stopped in the middle of frame output.

 [FIG. 6] FIG. 6 is a diagram for explaining processing when the penalty count and the number of transmitted packets of the frame are the same.

7] Fig. 7 is a diagram for explaining how data is handled when a buffer overflow occurs.

8] FIG. 8 is a diagram showing the configuration of the second embodiment of the data transmitting apparatus according to the present invention. 9] FIG. 9 is a diagram showing the configuration of the third embodiment of the data transmitting apparatus according to the present invention. Is

FIG. 10 is a diagram showing a configuration of a data transmitting apparatus according to a fourth embodiment of the present invention.

11] FIG. 11 is a diagram showing a configuration of a fifth embodiment of a data transmitting apparatus useful for the present invention.

 FIG. 12 is a diagram showing a configuration of a data transmitting apparatus according to a sixth embodiment of the present invention.

Explanation of symbols

 1, la, lb, lc, Id, le Data transmitter

 11, 11a, l ib, l id, l ie Frame division

 12 Common transmission buffer for all connections

 13, 13a Data reading part

 14 Packet assembly unit

 15 Packet transmitter

 16, 16d Scheduling information storage unit

 17, 17a, 17c, 17d Scheduler part

18 Buffer manager 21a-l, 21b-l, 21c-l, 21e— 1 connection # 1 transmission buffer

21a— 2, 21b— 2, 21c— 2, 21e-2 Connection # 2 Sending buffer B

 21a-n, 21b-n, 21c_n, 21e_n connection #n transmission buffer

 21d-l-l connection # 1 high priority buffer

 21d-l-2 connection # 1 low priority buffer

 21d-2-l connection # 2 high priority buffer

 21d-2-2 Connection # 2 low priority buffer

 21d-n-l connection # n high priority buffer

 21d-n-2 connection # n low priority buffer

 31b Noffer threshold exceeded detector

 41e Buffer Manager

 BEST MODE FOR CARRYING OUT THE INVENTION

 Hereinafter, embodiments of a data transmission device according to the present invention will be described in detail with reference to the drawings. In addition, this invention is not limited by this embodiment.

 [0016] Embodiment 1.

FIG. 1 is a diagram showing the configuration of the first embodiment of the data transmitting apparatus according to the present invention. The data transmitting apparatus 1 of the first embodiment divides a frame of a plurality of connections and transmits the divided data. Frame division unit 11 to be written to buffer, common transmission buffer unit 12 for all connections, data reading unit 13 for reading transmission buffer data of the designated connection, and header assignment for data read from transmission buffer The packet assembly unit 14 that assembles the packet as a packet, the packet transmission unit 15 that transmits the packet to the transmission path, the scheduling information storage unit 16 that stores information (frame information) necessary for scheduling, and the transmission The scheduler unit 17 that determines the output order of packets (data) to be transmitted and the transmission buffer unit 12 common to all connections are logically connected. The buffer that manages the read address, write address, and empty position address of each connection and controls the writing and reading by the frame dividing unit 11 and the data reading unit 13 so that each connection can be regarded as an independent buffer. And a management unit 18. The frame information identifies, for example, the number of frame divisions and frame priority. Frame type, data length of the divided data and serial number at the time of division (or the remaining data length in the frame), information on whether the data is the final divided data of the frame (or the beginning, continuation, and end information) Contains information. In addition to the frame information, the scheduling information storage unit 16 includes, for example, a transmission buffer address (data write position), penalty information, an upper limit on the number of data that can be continuously output, Information necessary for the scheduler unit 17 and the buffer management unit 18 such as the number of transmission data is also accumulated.

 [0017] Next, an outline of the operation of the data transmitting apparatus configured as described above will be described. FIG. 2 is a diagram showing an outline of the operation of the data transmission apparatus.

 [0018] In the present embodiment, for example, in the lower protocol, the upper protocol frame of connection # 1 is divided into three (data A1, data B1, data C1) (processing of frame divider 11). Furthermore, it is assumed that a packet (packet Al, packet Bl, packet C1) corresponding to each data is generated by attaching a packet header to each (processing of the packet assembly unit 14). Also, the top data of the frame is (H), the continuous data of the frame is (C), the final data of the frame is (T), and information on these divided data is held in the scheduling information storage unit 16 as frame information. .

 In the present embodiment, the data is held in a buffer having a logical queue configuration of the divided data power connection # 1. Note that the actual physical configuration of the common transmission buffer unit 12 for all connections is not connection-specific, but is configured to hold each connection in a single cell buffer. Therefore, the buffer management unit 18 designates a write address for the frame division unit 11, and further designates a read address for the data reading unit 13 based on the determination of the scheduler unit 17. This is logically equivalent to configuring a queue for each connection. However, if the physical configuration is a single cell buffer, a threshold value is not set for each connection, and if a large amount of data is written to a specific connection, the amount of data that can be written to other connections will decrease. Become.

Thus, in the data transmitting apparatus of the present embodiment, for example, as shown in FIG. 2, data Al 1, Bl 1, and C 1 are held in logically configured buffer 12_1 of connection # 1. Will be. Other connections # 2—Connection # n buffer 12-2 In 1-n, the data divided in the same way is retained.

In such a situation, the scheduler unit 17 of the present embodiment starts a data transmission scheduling operation. Note that the scheduler unit 17 of this embodiment determines a connection to be selected by the round robin method when all the connections have the same condition.

 [0022] For example, when round robin is started from connection # 1, the scheduler unit 17 first selects data A1 (H) and performs control for outputting the data (control to the buffer management unit 18). ). Next, the scheduler unit 17 selects the data B1 (C) of the connection # 1 and gives control to output the data in order to prioritize the output in units of frames. At this time, if it is a normal round robin, the data A2 of the connection # 2 should be selected. Therefore, the scheduler unit 17 maintains the fairness between the connections. A penalty is assigned and the count of the penalty is held in the scheduling information storage unit 16.

 [0023] Next, the scheduler unit 17 selects the data C1 (T) of the connection # 1 as data to be output next. At this time, if it is a normal round robin, the data A3 of the connection # 3 should be selected. Therefore, in order to maintain the fairness between the connections, the scheduler unit 17 changes the connection # 1 to the connection # 3. A penalty is assigned and the count of the penalty is held in the scheduling information storage unit 16.

 [0024] Next, a method for using the penalty will be described. Figure 3 is a diagram showing Penanoleti's method for 禾 IJ. FIG. 3 shows a case where the divided data is accumulated in the buffer (12-1-12-n) of each connection so that the top data of the frame will be the next output. Also, it is assumed that connection # 1 has a penalty for connection # 2 (1 count).

[0025] For example, in the case of normal round robin output, the scheduler unit 17 must first select the data of the connection # 1. Here, the connection # 1 is Since there is a penalty for connection # 2, first check whether there is data available for connection # 2 before selecting data for connection # 1. Connection # 2 has data A2 (H) that can be output. Therefore, the scheduler unit 17 selects data A2 (H) and performs control for outputting the data. At the same time, the penalty count for connection # 2 that existed in connection # 1 is decremented by one to eliminate the penalty.

 Next, scheduler unit 17 selects data B2 (C) of connection # 2, and performs control for outputting the data. Connection # 2 is then penalized (1 count) for connection # 3, which would have been selected in the normal round robin. Thereafter, the scheduler unit 17 performs control for outputting the data in the buffer while repeating the addition and subtraction of the penalty in the same manner as described above.

 [0027] Next, a process when the frame of a specific connection is longer than a predetermined reference will be described. Fig. 4 is a diagram for explaining the processing when the frame of a specific connection is longer than a predetermined reference, and this processing avoids a situation where a specific connection occupies output for a long time. To do.

 [0028] For example, when the output frame is longer than a specific reference (predetermined length), the data constituting the frame of one connection is preferentially output from the buffer. As a result, a specific connection occupies the output for a long time, and there is a concern that the frame output of other connections will be suppressed for a long time. Therefore, in the present embodiment, for example, an upper limit is set for the number of packets that can be continuously output for one frame, that is, the number of data that can be continuously selected, for the scheduling information storage unit 16. .

[0029] Figure 4 shows the case where the output of connection # 2 is stopped while the last one packet (data A2 (T)) is left due to the above limit while outputting the data of the frame of connection # 2. is doing. In this case, the frame output of connection # 2 has not been completed, but the scheduler unit 17 stops scheduling in units of frames based on the upper limit set in the scheduling information storage unit 16, and the normal round In Robin, select data A3 (H) of connection # 3 and perform control to output data A3 (H). When the scheduling unit 17 stops scheduling in units of frames as described above, the scheduling information storage unit 16 holds the number of transmission data in the frame of the connection. In addition, by setting this upper limit, for example, real-time performance is required and flexibility is set. For a frame whose frame length is shorter than a predetermined reference, it is possible to perform operations such as “not dividing”.

 [0030] Next, processing when a plurality of connections are stopped in the middle of frame output due to the upper limit of the number of packets that can be continuously output for one frame will be described. FIG. 5 is a diagram for explaining processing when a plurality of connections are stopped during frame output. Note that FIG. 5 shows processing when connection # 1 and connection # 2 are stopped in the middle of frame output and a series of round-robin output scheduling operations are completed due to the above upper limit.

 [0031] Normally, since the scheduler 17 starts a new round robin, the connection #

 In this embodiment, there are multiple connections whose data output is stopped in the middle of the frame. Select a connection with a large amount of data sent to. For example, the direction of connection # 2 over connection # 1 When the number of data output in one frame is larger, scheduler unit 17 selects connection # 2, data A2 (C), and then data B2. Control for outputting (T). Thereafter, the scheduler unit 17 selects connection # 1, which is also in the middle of frame output, and performs control to output data A1 (T). Even in such processing, penalty addition / subtraction is performed in the same manner as described above.

 Next, a process when a new round robin is performed, a plurality of connections are stopped during frame output, and the penalty count and the number of transmitted data of the frame are the same will be described. FIG. 6 is a diagram for explaining processing when the penalty count and the number of transmitted packets of the frame are the same.

[0033] In such a case, the scheduler unit 17 considers the characteristics of the frame, and selects a connection so that priority is given to the output of a frame that requires more renotime characteristics. For example, if the frame in the middle of output of connection # 1 is a frame that transfers non-realtime data, and the frame in the middle of output of connection # 2 is a frame that transfers real-time data, the scheduler unit 17 preferentially selects data A2 (C) of connection # 2 and performs control to output the data. The schedule Yuri The scheduling information storage unit 16 retains the frame characteristics necessary for encoding.

[0034] Next, data handling when a buffer overflow occurs will be described. Fig. 7 is a diagram for explaining how to handle data when a buffer overflow occurs.

 [0035] As described above, the transmission buffer unit 12 common to all connections is logically managed for each connection. Physically, the transmission buffer unit 12 is written in a single buffer. 18 is managed logically by connection. In FIG. 7, for simplification, the number of data that can be recorded in the transmission buffer 12 common to all connections is six. For example, data A1 (C) and data Bl (C) of the frame in the middle of output of connection # 1 The data A2 (C), data B2 (T), and the data A3 (H), data B3 (T) of the untransmitted frame of connection # 3 are saved and all connections are saved. Assume that there is no more free space in the common transmission buffer unit 12 where other data can be written.

 [0036] In this state, if further continuation data (C) of connection # 1 arrives, normally all newly connected data cannot be written in all connection common transmission buffer section 12. Therefore, in the frame division unit 11 of the present embodiment, in order to prioritize transfer in units of frames, the data already written in the buffer is searched for data with a low priority, for example, best-fault data. If such data exists, the data is discarded in units of frames, and newly arrived output data is written in the discarded empty area.

 [0037] For example, if the frame data of connection # 3 is best-f-auto data, data A3 (H) and data B3 (T) of connection # 3 are discarded from the transmission buffer unit 12 common to all connections. The frame division unit 11 writes the data C1 (C) of connection # 1 in this empty area.

[0038] As described above, in the present embodiment, data transfer is performed in the lower protocol while being conscious of the frame of the upper protocol. As a result, it is possible to shorten the transmission time in the frame unit of the upper protocol while maintaining the fairness of the packet (data) output in each connection. In the present embodiment, the frame Even though some of the data is transferred, it is difficult for the receiving side to discard the frame that has already been transferred due to a transfer delay and the transferred frame is discarded on the receiving side. The frame-by-frame reception probability can be improved. Note that these effects are particularly noticeable when the transfer speed of the transmission path between the data transmission device and the data reception device is low.

 [0039] Also, in this embodiment, since the frame assembly is completed in a short time in the data receiving apparatus that is the communication counterpart, reception for waiting for data until all the packets constituting the frame arrive. The amount of data remaining in the buffer can be reduced. Therefore, when the reception buffer is configured as a common buffer for each connection, the amount of reception buffer can be significantly reduced.

 [0040] Embodiment 2.

 In the first embodiment described above, the common transmission buffer unit 12 for all connections employs a method in which a single buffer is shared by each connection. On the other hand, in the second embodiment, a case where the transmission buffer power is physically separated in connection units will be described.

 [0041] FIG. 8 is a diagram showing the configuration of the second embodiment of the data transmitting apparatus according to the present invention. The data transmitting apparatus la of the second embodiment divides a frame of multiple connections, and A frame division unit 11a that writes data to a connection-specific transmission buffer and a connection

# 1 Transmission buffer unit 21a_l, connection # 2 Transmission buffer unit 21a_2, ..., connection # n Transmission buffer unit 21 a—n, data reading unit 13a that reads data from the specified transmission buffer for each connection, and packet to be transmitted And a scheduler unit 17a that determines an output order of the data and gives an instruction to the data reading unit 13a. Note that the same components as those in the first embodiment described above are denoted by the same reference numerals and description thereof is omitted.

 [0042] Next, the operation of the data transmitting device la configured as described above will be described. Here, only processing different from the first embodiment will be described.

[0043] The frame dividing unit 11a that has received the frame data of each connection divides the frame. At the same time as writing the frame information at the time of allocation to the scheduling information storage unit 16, the divided data is written to the connection # 1 transmission buffer unit 21a_l—connection #n transmission buffer unit 21a_n provided for each connection.

Note that the operation of the scheduler unit 17a is basically the same as that of the scheduler unit 17 of the first embodiment described above. The scheduler unit 17a of the present embodiment is different from the scheduler unit 17 described above in the process of instructing the data reading unit 13a of the data reading order of the selected connection. Also, in the first embodiment, when a buffer overflow occurs during the output of data of a specific frame and the continuous data of that frame is input, the connection with the frame data of the best buffer of the common buffer is used. However, in this embodiment, since the buffer is managed individually, this processing is not performed.

[0045] As described above, in the present embodiment, the data transmission device includes an individual transmission buffer for each connection. As a result, it is possible to obtain the same effects as those of the first embodiment described above without being affected by the frame data of other connections.

 [0046] Embodiment 3.

 In the above-described second embodiment, since the connection-specific transmission buffer is provided, for example, when a buffer overflow occurs during frame transmission and continuous data is input during frame transmission, As you can see, no data is being discarded on a frame-by-frame basis. Therefore, in this embodiment, the discard processing of data in units of frames is applied to a configuration including a connection-specific transmission buffer.

FIG. 9 is a diagram showing the configuration of the third embodiment of the data transmitting apparatus according to the present invention. The data transmitting apparatus lb of the third embodiment includes a frame dividing unit l ib and a connection # 1 transmission. Buffer section 21b_l, connection # 2 transmission buffer section 21b-2,..., Connection #n transmission buffer section 21b_n, and buffer threshold value excess detection section 31b. Note that the same components as those in the first or second embodiment described above are denoted by the same reference numerals and description thereof is omitted.

[0048] Next, the operation of the data transmission device lb configured as described above will be described. This Here, only processing different from that of the second embodiment will be described.

[0049] The frame division unit l ib writes the divided data to the connection # 1 transmission buffer unit 21b_l—connection #n transmission buffer unit 21b_n provided for each connection in the same process as the frame division unit 11a described above. .

[0050] Connection # 1 transmission buffer unit 21b_l—connection #n transmission buffer unit 21b_n

When the divided data is written, the data storage amount is notified to the buffer threshold excess detection unit 31b.

 [0051] The buffer threshold value excess detection unit 31b holds a buffer threshold value for each connection in advance, and monitors whether each buffer unit exceeds the threshold value. For example, if there is a connection that exceeds the threshold value, the buffer threshold value excess detection unit 31b notifies the frame division unit l ib to that effect. Then, the frame dividing unit l ib discards data in units of frames when a frame arrives without performing frame division of the corresponding connection during the period in which the threshold value is continuously exceeded.

 [0052] Thus, in the present embodiment, the amount of data is monitored in units of buffer units so that the buffer unit for each connection does not hold data exceeding the threshold, and for example, exceeds the threshold value. In this case, all data of the corresponding frame is discarded. As a result, the same effects as those of the second embodiment can be obtained, and the reception probability of data in units of frames can be further improved.

 In the present embodiment, the case of providing connection-specific transmission buffers has been described. However, the present invention is not limited to this, and a common buffer for all connections as in Embodiment 1 is provided, which is logically This can also be applied to the case where a separate transmission buffer is configured for each connection.

 [0054] Embodiment 4.

 In the fourth embodiment, the reception probability of data in units of frames is improved by processing different from that in the third embodiment.

[0055] FIG. 10 is a diagram showing a configuration of the data transmitting apparatus according to the fourth embodiment of the present invention. The data transmitting apparatus lc of the fourth embodiment includes a connection # 1 transmission buffer unit 21c-1 and a connection. # 2 Transmission buffer section 21c_2, ..., connection # n Transmission buffer section 21c—n and previous In addition to the processing of the scheduler unit 17a described above, a scheduler unit 17c that prioritizes the output of the frame when the buffer unit exceeds a predetermined data storage amount is provided. Note that the same reference numerals are given to the same configurations as those in the first, second, or third embodiment, and the description thereof is omitted.

 [0056] Next, the operation of the data transmission device lc configured as described above will be described. Here, only processing different from that of Embodiment 3 will be described.

 [0057] In the present embodiment, connection # 1 transmission buffer unit 21c—one connection #n transmission buffer unit 21c-n determines whether or not data exceeds the buffer threshold for each connection. Monitor autonomously. When there is a connection that exceeds the threshold, each buffer unit notifies the scheduler unit 17c to that effect.

 [0058] When the scheduler unit 17c receives information indicating that the threshold value has been exceeded, scheduling is performed with priority given to output of frame data of the corresponding connection.

 [0059] As described above, in the present embodiment, when the buffer unit for each connection exceeds a predetermined data accumulation amount, scheduling is performed so that the data of the corresponding frame is preferentially transmitted. I decided to do it. As a result, the same effects as in the second embodiment can be obtained, and the reception probability of data in units of frames can be further improved.

 In the present embodiment, the case where the transmission buffer for each connection is provided has been described. However, the present invention is not limited to this, and a common buffer for all connections as in the first embodiment is provided to logically. This can also be applied to the case where a separate transmission buffer is configured for each connection.

 [0061] Embodiment 5.

 In the above Embodiment 1 and Embodiment 4, a separate transmission buffer (physically or logically) is provided for each connection. However, in Embodiment 5, a single transmission buffer is provided for each connection. For example, a case will be described in which the data of each connection is stored in a plurality of transmission buffers according to type.

FIG. 11 is a diagram showing the configuration of the data transmitting apparatus according to the fifth embodiment of the present invention. The data transmitting apparatus Id according to the fifth embodiment includes a frame dividing unit id and a connection # 1 high priority. Destination buffer 21d-ll, connection # 1 Low priority buffer 21d-l-2, connection # 2 High priority buffer unit 21d_2_l, Connection # 2 Low priority buffer unit 21d_2_2, ..., Connection #n High priority buffer unit 21d_n_l, Connection #n Low priority buffer unit 2 In addition to the processing of the scheduler unit 17c, a scheduler unit 17d that prioritizes the output of frames with high priority is provided.

 [0063] Next, the operation of the data transmitting device Id configured as described above will be described. Here, only processing different from Embodiments 1 to 4 will be described.

 [0064] For example, the priority for each frame type (corresponding to the data type) is registered in advance in the frame division unit id for each connection, and the frame division unit id receives the frame. When dividing, the divided data is written in the corresponding buffer according to the registered priority. At the same time, information for each buffer (information on priority) is written to the scheduling information storage unit 16d.

 [0065] When scheduling the output of data for each connection, the scheduler unit 17d schedules the output data in consideration of the priority of each connection held in the scheduling information storage unit 16d. Except for the processing that prioritizes the output of a frame with a high priority, the processing is the same as in the first, second, third, or fourth embodiments.

 [0066] As described above, in this embodiment, the frame data of each connection is classified as a type.

 It was decided to store it in multiple transmission buffers according to (frame priority). As a result, the same effects as those of the first to fourth embodiments described above can be obtained, and frame data with higher priority can be transmitted with priority.

 [0067] Embodiment 6.

 In the above Embodiment 1 and Embodiment 5, the data after dividing the frame is written in each transmission buffer. However, in Embodiment 6, the frame itself before dividing the frame is stored in each frame. A case will be described in which a packet is assembled by dividing it after writing to the transmission buffer and reading a frame from each transmission buffer.

FIG. 12 is a diagram showing the configuration of the sixth embodiment of the data transmitting apparatus according to the present invention. The data transmitting apparatus le of the sixth embodiment includes a connection # 1 transmission buffer unit 21e-1 and a connection. # 2 Transmission buffer section 21e_2, ..., connection # n Transmission buffer section 21e_n A frame division unit l ie and a buffer management unit 41e.

[0069] Next, the operation of the data transmission device le configured as described above will be described. Here, only processing different from Embodiments 1 to 15 will be described.

[0070] In the present embodiment, assuming that the frame is divided based on the information that the frame obtained from each transmission buffer unit is written, the buffer management unit 41e assumes the frame information. Is written to the scheduling information storage unit 16.

[0071] Then, the frame dividing unit ie reads out the frame from each transmission buffer based on the scheduling by the scheduler unit 17a, divides and outputs the frame.

[0072] As described above, even when the frame itself before dividing the frame is written in each transmission buffer, and the frame is divided and assembled after reading the frame from each transmission buffer, it is the same as the above-described embodiment. Similar effects can be obtained.

 Industrial applicability

As described above, the data transmission device according to the present invention is useful as a communication device that constitutes a communication system that performs data transfer in consideration of higher-order protocol frames, and particularly handles frames of a plurality of connections. Suitable for communication devices.

Claims

The scope of the claims  [1] In a data transmission apparatus that performs data transmission on a frame basis while maintaining output fairness in multiple connections,  Frame information that divides the data on the frame into a predetermined size for each connection and writes the divided data to the specified position in the transmission buffer common to all connections, and at the same time includes the type of divided data (first, continuous, last) A frame dividing means for generating the frame, a data reading means for reading out the divided data at the specified position from the transmission buffer, and the first divided data constituting the frame of the specific connection by round robin. Scheduling means for performing scheduling in units of frames so that priority is given to output of the frames up to the final divided data to be  It manages the data write position, data read position, and free space for the transmission buffer, designates the write position for the frame dividing means, and further reads the data read means according to the determination by the scheduling means. A buffer management means for specifying  A data transmission device comprising: [2] The scheduling means includes  As a result of performing the scheduling in units of frames, when the divided data (single or plural) of other connections to be selected by round robin is not selected, the other connections are performed in the connection in which the scheduling in units of frames is performed. Penalize the connection for  The data transmission device according to claim 1, wherein an opportunity for selecting divided data is not given to the connection on which the scheduling in units of frames is performed until the penalty is eliminated. [3] Set the maximum number of data that can be selected continuously for one frame,  The scheduling means includes If the data on the frame of the connection for which the scheduling is performed in units of frames exceeds the upper limit, the divided data exceeding the upper limit is not selected. 3. The data transmission apparatus according to claim 2, wherein the scheduling in units of frames is stopped, and then the divided data of the next connection is selected by round robin. [4] The scheduling means includes  When the frame unit scheduling is stopped halfway, the number of transmission data in that frame is held,  4. The remaining divided data of a connection having a large number of transmission data is selected when a new scheduling is performed in a state where there are a plurality of connections in which scheduling in frame units is stopped in the middle. Data transmission device. [5] Preserve frame characteristics,  The scheduling means includes  When there are a plurality of connections in which the scheduling for each frame is stopped in the middle, and when the penalty and the number of transmission data are the same, a new scheduling is performed, and real-time characteristics are required based on the characteristics of the frame. 5. The data transmission device according to claim 4, wherein scheduling is performed so that priority is given to frame output. [6] The frame dividing means includes:  If an overflow occurs in the transmission buffer while writing the frame division data, it is detected by searching the frame division data with lower priority from the division data already written in the transmission buffer. 2. The data transmitting apparatus according to claim 1, wherein the divided data is discarded in units of frames, and the subsequent divided data in the frame being written is written in the empty area after the discarding. [7] The transmission buffer is provided with a threshold for the amount of data stored for each connection, and for connections whose data storage exceeds the threshold, the entire corresponding frame is discarded. 1. The data transmission device according to 1. [8] The transmission buffer is provided with a threshold for the amount of data stored for each connection, 2. The data transmission device according to claim 1, wherein the scheduling means performs scheduling so that a connection in which a data accumulation amount exceeds the threshold value gives priority to output of the corresponding frame. [9] The region for each connection of the transmission buffer is divided according to the priority of the frame, and the scheduling means performs scheduling in units of frames so as to give priority to the output of the divided data of the frame having a high priority. The data transmission device according to claim 1, wherein  [10] In a data transmission apparatus that performs data transmission based on a frame unit while maintaining output fairness in multiple connections,  The data on the frame is divided into a predetermined size for each connection, and the divided data is written to a separate transmission buffer for each connection, and at the same time, frame information including the type of divided data (first, continuous, last) is generated. Frame dividing means;  Data reading means for reading out the divided data from the transmission buffer based on a scheduling result;  When the first divided data constituting a frame of a specific connection is selected by round robin, scheduling is performed in units of frames so that priority is given to the output of the frame up to the final divided data constituting the same frame. Scheduling means,  A data transmission device comprising:  [11] The scheduling means includes:  As a result of performing the scheduling in units of frames, when the divided data (single or plural) of other connections to be selected by round robin is not selected, the other connections are performed in the connection in which the scheduling in units of frames is performed. Penalize the connection for  The data transmission device according to claim 10, wherein an opportunity for selecting divided data is not given to the connection on which the scheduling in units of frames is performed until the penalty is eliminated. [12] Set the maximum number of data that can be selected continuously for one frame,  The scheduling means includes If the data on the frame of the connection for which the scheduling is performed in units of frames exceeds the upper limit, the divided data exceeding the upper limit is not selected. 12. The data transmission apparatus according to claim 11, wherein the scheduling in units of frames is stopped, and then the divided data of the next connection is selected by round robin. [13] The scheduling means includes:  When the frame unit scheduling is stopped halfway, the number of transmission data in that frame is held,  13. The remaining divided data of a connection with a large number of transmission data is selected when a new scheduling is performed in a state where there are a plurality of connections in which the scheduling for each frame is stopped halfway. Data transmission device. [14] Hold the special frame,  The scheduling means includes  When there are a plurality of connections in which the scheduling for each frame is stopped in the middle, and when the penalty and the number of transmission data are the same, a new scheduling is performed, and real-time characteristics are required based on the characteristics of the frame. 14. The data transmission device according to claim 13, wherein scheduling is performed so that frame output is prioritized.  [15] A threshold is set for the amount of data stored in each individual transmission buffer for each connection, and the connection for which the data storage amount of the transmission buffer exceeds the threshold is applicable. 11. The data transmission apparatus according to claim 10, wherein the entire frame is discarded.  [16] Set a threshold for the amount of data stored in each transmission buffer for each connection,  11. The data according to claim 10, wherein the scheduling means performs scheduling so that priority is given to output of a corresponding frame for a connection in which a data accumulation amount of the transmission buffer exceeds the threshold value. Transmitter device. [17] The individual transmission buffer is further divided according to the priority of the frame for each connection, and the scheduling means performs scheduling in units of frames so as to give priority to the output of the divided data of the frame having a high priority. 11. The characteristic of claim 10 Data transmission device.  In a data transmission apparatus that performs data transmission on a frame basis while maintaining fairness of output in multiple connections,  Assuming a situation in which a frame written in a separate transmission buffer is divided for each connection, a buffer management means for generating frame information including the type of divided data (first, continuous, last),  When the first divided data that constitutes a frame of a specific connection is selected by round robin, the scheduling in units of frames is performed so that priority is given to the output of the frame up to the final divided data that constitutes the same frame. Scheduling means for performing  A frame dividing means for reading out an individual transmission buffer power frame for each connection based on a result of scheduling, and dividing data on the read frame into a predetermined size;  A data transmission device comprising: