JP4285101B2 - Real-time data communication system, real-time data communication apparatus, and real-time data communication method - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a real-time data communication system for performing real-time data communication, for example.
[0002]
[Prior art]
Generally, when transmitting real-time data, a streaming technique using RTP (Real Time Transport Protocol) on UDP (User Datagram Protocol) has been used. Since UDP is a connectionless protocol and does not have a mechanism such as flow control or retransmission control, it is very suitable for real-time transmission that eliminates the need for data after a certain period of time.
[0003]
On the other hand, TCP (Transmission Control Protocol) is a connection-type protocol and has a flow control and retransmission control mechanism, and has been used for storage-type content download and streaming distribution. HTTP using TCP has been established as a data transmission means that can pass through a firewall or NAT (Network Address Translator), which is difficult with UDP, and is currently very effective.
[0004]
In Patent Document 1, in order to enable connection control in which resource limitation is performed according to a user, the use of a connection request is identified and used according to a resource usage condition set according to the identified user. In a communication processing device that has a connection request from an unspecified number of users, the priority is given to a specific user by determining the resource area to be used and setting many resources that can be used by high-priority users. Discloses a communication processing apparatus capable of establishing a connection.
[0005]
[Patent Document 1]
JP 2003-125022
[0006]
[Problems to be solved by the invention]
However, if the conventional real-time data communication system described above tries to transmit real-time data using HTTP using TCP, the connection by HTTP may be restricted by servers, proxies, routers, etc. The connection was sometimes lost. For example, a proxy server may have a maximum connection time specified by a policy for preventing unauthorized access by a server administrator. In addition, transmission over a long period of time may be cut off by a rule for preventing unauthorized access of an organization operating a network management / connection such as ISP (Internet Service Provider).
[0007]
When the connection is cut off during such transmission, it is necessary to re-establish the HTTP connection again and flow data on the connection. However, this reconnection requires a negotiation with overhead for transmission management, and causes a delay in transmission of real-time data that becomes a data payload for the time spent on it. Therefore, if reconnection takes time, the real-time property of data may be lost.
[0008]
Therefore, the present invention has been made in view of the above points, and does not cause a delay to real-time data due to a sudden disconnection of HTTP connection in real-time data transmission using HTTP effective for communication over a firewall or NAT. It is an object of the present invention to provide a real-time data communication system, a real-time data communication apparatus, and a real-time data communication method that maintain continuous real-time transmission.
[0009]
[Means for Solving the Problems]
The server of the real-time data communication system of the present invention has a receiving unit for receiving input real-time data, and a multiple connection generating unit for generating a plurality of connections so that application software can be identified. A connection management unit that manages the state of a plurality of connections within a reference time, and a transmission unit that transmits real-time data through any of the plurality of connections.
[0010]
In addition, the client has a receiving unit that receives real-time data transmitted from any of a plurality of connections, and a multi-connection generation unit that generates a plurality of connections so that the application software can identify the real-time data. In order to do so, a connection management unit that manages the state of the connection within a reference time and a transmission unit that outputs the received real-time data are provided.
[0011]
Thus, the real-time data communication system of the present invention uses each transmission period within the reference time for each connection of a plurality of connections to transmit one session of real-time data.
[0012]
Therefore, according to the present invention, the following operations are performed.
The server prepares a plurality of connections for transmission by a plurality of connection generators in advance, and when connection connection is detected by the connection / disconnection monitoring unit, the connection switching unit switches to another connection. Then, the data is transmitted to the client via the connection in units of data (for example, in units of packets) by the transmission unit.
[0013]
Here, if the connection is disconnected by the connection disconnection unit based on the reference time set by the reference time setting unit during transmission according to the specifications of the proxy or server (or client), the connection connection / disconnection monitoring unit connects When disconnection is detected, data is transmitted to the client via another connection prepared in advance by the multiple connection generation unit, so that the real-time property of the data can be maintained.
[0014]
A plurality of connections created in advance by a plurality of connection generation units are transferred to other connections by the connection switching unit for each transmission period within each connection time set so as to have a time difference by the connection shifting connection time setting unit. I will switch. Thereby, simultaneous disconnection due to a certain time-out can be prevented. When the connection connection / disconnection monitoring unit detects disconnection of the connection, the connection disconnected by the plurality of connection generation units is re-established and held.
[0015]
In order to prevent disconnection due to no-communication timeout, a connection that is not used for transmission of real-time data is disconnected by the connection disconnection unit based on the reference time set by the reference time setting unit. To do.
[0016]
As a result, when communicating real-time data over multiple connections, you can selectively use multiple connections to improve transfer efficiency and real-time performance of the data, even if one connection is disconnected, By causing data to flow through another connection, it is possible to prevent a delay caused by re-establishing the connection.
[0017]
DETAILED DESCRIPTION OF THE INVENTION
Embodiments of the present invention will be described below with reference to the drawings as appropriate.
[0018]
FIG. 1 is a diagram showing a configuration of a real-time data communication system applied to the embodiment of the present invention.
In the present invention, a device configuration for transmitting real-time data as shown in FIG. 1 is assumed. In order to make the explanation easy to understand, the HTTP server 1 is set as the transmitting side and the HTTP client 2 is set as the receiving side. An HTTP server 1 and an HTTP client 2 are mounted as an HTTP (Hypertext Transfer Protocol) server or client, and transmit and receive data by HTTP over TCP.
[0019]
In FIG. 1, an HTTP server 1 and an HTTP client 2 are connected by a plurality of HTTP connections 1 (3-1), HTTP connections 2 (3-2), ..., HTTP connections n (3-n). HTTP connection 1 (3-1), HTTP connection 2 (3-2), ... Real-time data DI input to the HTTP server 1 flows on the HTTP connection n (3-n), and real-time from the HTTP client 2 Assume that data DO is output.
[0020]
First, an HTTP request is issued from the HTTP client 2 to the HTTP server 1, and the HTTP server 1 sends real-time data as an HTTP response. Even when the HTTP server 1 and the HTTP client 2 are mounted in reverse to FIG. 1, the HTTP client issues an HTTP request to the HTTP server and transmits the real time data in the HTTP request. The HTTP server returns an HTTP response to the HTTP client when the HTTP client finishes sending all the data or when an error occurs in the HTTP server.
[0021]
FIG. 2 is a diagram showing a configuration of a real-time data communication system connected by HTTP proxy.
When the HTTP proxy 4 is inserted between the HTTP server 1 and the HTTP client 2, the configuration shown in FIG. 2 is taken. In this case, if the HTTP proxy 4 only transfers data carried by HTTP over TCP between the HTTP server 1 and the HTTP client 2, it can be regarded as the same as FIG.
[0022]
Specifically, in FIG. 2, the HTTP proxy 4 sends a plurality of HTTP connections 1 (3-1), HTTP connections 2 (3-2),... HTTP connections n (3 -n) to the HTTP server 1. ), The HTTP proxy 4 is connected to the HTTP client 2 with a plurality of HTTP connections 1 (5-1), HTTP connections 2 (5-2),... It will behave as the server 4-2 connected in 5-n).
[0023]
The plurality of HTTP connections created in FIGS. 1 and 2 described above are created with an arbitrary time difference as will be described later. This is to prevent multiple simultaneous disconnections due to a transfer timeout of the HTTP proxy 4 or the HTTP server 1 (or HTTP client 2).
[0024]
Here, the real-time data DI and DO are data for which delivery is desired to be completed within a certain period. For example, video and audio data handled by live distribution and videophone. The real-time data DI and DO may be video or audio data obtained from a video camera or microphone attached to the HTTP server 1 or RTP packets from other devices (or encoders (compression devices)) connected via a network. It may be streaming data to be transferred.
[0025]
Real-time or streaming data can be divided into packets or smaller units. Even if some of the subdivided data is lost during the communication, if the receiving side can receive subsequent data, it is possible to process the data that arrives according to the time.
[0026]
If the total bandwidth is constant, the total bandwidth that can be obtained even if multiple HTTP connections are used does not change. For example, when real-time data is transmitted using HTTP, the connection is limited due to restrictions on the HTTP server and HTTP proxy. It can be avoided that it is difficult to ensure real-time performance due to being cut off.
[0027]
FIG. 3 is a diagram illustrating a configuration example of a transmission packet and a streaming data packet.
In FIG. 3, the transmission packet 11 has an RTP header 12, and the RTP header 12 has a sequence number SN13 indicating time information and a time stamp TS14.
[0028]
The streaming data packet 15 has streaming data as an RTP payload.
[0029]
Here, the HTTP server 1 on the transmission side can detect a connection in which a delayed packet is generated based on the sequence number SN13 indicating time information and the time stamp TS14, and can switch to another connection.
[0030]
In addition, when the RTP is used as the transmission packet 11 in order to reproduce the streaming data, the receiving-side HTTP client 2 uses the RTP sequence number SN13 and the time stamp TS14 in the transmission packet 11 to order the data. Can be sorted. Note that the reproduction method depends on the streaming data as the RTP payload shown in the streaming data packet 15.
[0031]
FIG. 4 is a diagram showing the configuration of the HTTP server.
In FIG. 4, the HTTP server 1 includes a receiving unit 21 that receives input real-time data DI, and the real-time data DI received by the receiving unit 21 as a plurality of HTTP connections 1 (3-1) and HTTP connections 2 (3 -2),..., HTTP connection n (3-n) to be transmitted as an HTTP connection output CO, a plurality of HTTP connections 1, a plurality of HTTP connections 1 (3-1), HTTP connections 2 (3-2) ... A connection management unit that has a plurality of connection generation units that generate a plurality of HTTP connection n (3-n) in an identifiable manner by the application software program, and manages the state of the plurality of connections in order to transmit real-time data 22 is configured.
[0032]
FIG. 5 is a diagram showing the configuration of the HTTP client.
In FIG. 5, the HTTP server 1 is an HTTP connection input input from any of a plurality of HTTP connections 1 (3-1), HTTP connections 2 (3-2),... HTTP connections n (3 -n). The receiving unit 25 that receives the CI, the transmitting unit 26 that outputs the HTTP connection input CI received by the receiving unit 25 as real-time data DO, a plurality of HTTP connections 1 (3-1), and HTTP connections 2 (3-2) ),... Connection management that has a plurality of connection generation units that generate a plurality of HTTP connection n (3-n) in an identifiable manner by the application software program, and manages the state of the plurality of connections in order to receive real-time data The unit 24 is configured.
[0033]
FIG. 6 is a diagram illustrating the configuration of the connection management unit.
When the connection management unit 22 of the HTTP server 1 transmits real-time data from the HTTP server 1 in response to a request from the HTTP client 2 via the HTTP connection for communication using the HTTP protocol managed by the application software, The connection disconnection unit 33 is configured to disconnect the connection based on the reference time set by the reference time setting unit 31.
[0034]
The connection management unit 22 of the HTTP server 1 enables the application software program to identify a plurality of HTTP connections 1 (3-1), HTTP connections 2 (3-2), ..., HTTP connections n (3-n). A plurality of connection generation units 31 for generating a plurality of connections, and a connection switching unit 36 for switching a plurality of HTTP connections 1 (3-1), HTTP connections 2 (3-2), ..., HTTP connections n (3-n) A plurality of generated HTTP connections 1 (3-1), HTTP connections 2 (3-2)... HTTP connections n (3-n) are sequentially or selectively switched, and other than the real-time data transmission period. Transmits HTTP data from HTTP server 1 while transmitting pseudo data 2 configured to transmit real-time data to. The connection management unit 24 of the HTTP client 2 has a similar configuration.
[0035]
The connection management unit 22 of the HTTP server 1 includes a connection connection / disconnection monitoring unit 34 that monitors a plurality of connections generated by the multiple connection generation unit 31 and a connection disconnected by the connection disconnection unit 33, and monitors connection connection / disconnection. When the connection is detected by the unit 34, the connection switching unit 36 switches to another connection, and when the connection connection / disconnection monitoring unit 34 detects the disconnection, the connection is disconnected by the multiple connection generation unit 31. It is configured to re-establish the connection. The connection management unit 24 of the HTTP client 2 has a similar configuration.
[0036]
The connection management unit 22 of the HTTP server 1 includes a connection shift connection time setting unit 35 that sets a time for shifting each connection time of a plurality of connections generated by the multiple connection generation unit 31, and includes a connection shift connection time setting unit. The connection switching unit 36 is configured to switch to another connection for each transmission period within each connection time set by being shifted by 35. The connection management unit 24 of the HTTP client 2 has a similar configuration.
[0037]
The connection management unit 24 of the HTTP client 2 is configured to include a data rearrangement unit that arranges and outputs the order of real-time data when receiving one session of real-time data from a plurality of connections. The connection management unit 24 of the HTTP server 1 has the same configuration.
[0038]
FIG. 7 is a diagram showing identification of a plurality of connections.
In the HTTP server 1, when transmitting real-time data over a plurality of connections using the protocol HTTP managed by the upper application software 42 of the operating system OS 41, the plurality of connections generated so that the application software 42 can be identified are sequentially provided. Alternatively, it can be selectively used to separate real-time data.
[0039]
At this time, the application software 42 switches the URL1, URL2, URL3,... Corresponding to the reproduction software 43-1, the reproduction software 43-2, the reproduction software 43-3,. Indicates the type of real-time data, and identifies a session using URL, HTTP authentication (HTTP header), or other unique authentication methods. A session identifier is provided for one session, and a plurality of HTTP connections are permitted for one URL using the same session identifier. In addition, data separation by switching is performed in units of streaming data packets.
[0040]
Further, in the HTTP client 2, when receiving real-time data on a plurality of connections using the protocol HTTP managed by the application software 42, the plurality of connections generated so that the application software 42 can be identified are sequentially or selectively selected. It is possible to rearrange the order of real-time data.
[0041]
At this time, the application software 42 switches the URL1, URL2, URL3,... Corresponding to the reproduction software 43-1, the reproduction software 43-2, the reproduction software 43-3,. Identifies the connection. In addition, the rearrangement of the order of the data separated by switching is performed in units of streaming data packets.
[0042]
In the HTTP client 2, the application software 42 identifies that one connection has been disconnected, and is detected as an error in both the HTTP server 1 and the HTTP client 2. When establishing a disconnected connection, the application software 42 identifies URL1, URL2, URL3,... And determines that the connection is for real-time data transmission.
[0043]
In addition, the determination as to whether or not the connection is for real-time data transmission is specifically that a real-time data URL is provided and the HTTP connection established by accessing it is a connection for real-time data transmission. Application software 42 identifies.
[0044]
For example, when a URL “http: //.../...real_Time-streaming? Session = 123456” is accessed, a connection for transmitting real-time data can be established. However, since simultaneous real-time data is scattered to multiple terminals when accessed simultaneously by multiple terminals, a session identifier is provided in the URL as “http: //.../...real_Time-streaming? Session = 123456”. Alternatively, information may be inserted into an HTTP request or response header.
[0045]
Next, details of the operation of the real-time data communication system applied to the embodiment of the present invention configured as described above will be described below.
[0046]
FIG. 8 is a flowchart showing the operation of the HTTP server.
FIG. 8 shows steps until the HTTP server 1 that transmits real-time data creates the HTTP connection 3 and transmits real-time data. The details are described below.
[0047]
In step S1, a plurality of HTTP connections are created. Specifically, when the HTTP server 1 starts processing on the transmission side, an HTTP connection 3 for transmitting real-time data to the HTTP client 2 on the reception side is created. A plurality of HTTP connections 3 are created by the multiple connection generator 31.
[0048]
When one connection is created, the process moves to step S2. In step S2, it is determined whether real-time data has been input, whether the real-time data input has timed out, or not. Specifically, the HTTP server 1 obtains real-time data from a device such as a camera directly attached to the transmission side as described above, or obtains a packet such as RTP from another device on the network. is there. In the embodiment of the present invention, both cases are similarly treated as inputs. When real-time data is input, the process may be immediately performed to the next processing step S3, or a plurality of data may be accumulated in a buffer or a queue and then the process may proceed to step S3. Further, in step S2, when an error such as an input time-out occurs, a transition is made to S3 so that the block does not block until real-time data is input after waiting for input of real-time data.
[0049]
In step S3, it is determined whether or not there is another connection. If there is another connection, in step S4, the other connection becomes the connection time set by shifting by the connection shifting connection time setting unit 35. Thus, after being scheduled to be created later with a time difference, all connections are automatically created. After the connection, the connection state shifted in time is maintained.
[0050]
In step S5, an error check is made to determine whether the input of real-time data has timed out in step S2 or not. Specifically, determination of error check such as input timeout of real-time data is performed by the connection connection / disconnection monitoring unit 34 of the connection management unit 22 of the HTTP server 1. At this time, the reference set by the reference time setting unit 31 is used. When no communication timeout occurs based on the time and disconnection of the connection is detected, the multiple connection generation unit 31 re-establishes the disconnected connection.
[0051]
Thereafter, when an HTTP connection once connected is disconnected for some reason, the connection is immediately reconnected by the plurality of connection generation units 31 so that a predetermined number or more of connections can always be used. In addition, in order to prevent a no-communication timeout based on the reference time set by the reference time setting unit 31 and disconnection of the connection by the connection disconnection unit 33, a fixed pseudo-time other than the real-time data transmission period is used. Data may be sent out.
[0052]
If there is no error such as timeout, the process proceeds to step S6. If there is an error such as a timeout, the process proceeds to step S8.
[0053]
In step S6, an HTTP connection for transmitting the data input in step S2 is selected from the HTTP connection created in step S1. Specifically, the selection of the HTTP connection is performed by the connection switching unit 36, and the selection method may specify the created HTTP connections in order, or if the data has been transmitted before, it is used so far. The existing HTTP connection may be used continuously, or when data is newly transmitted or when the HTTP connection is changed by some judgment, one of the HTTP connections created in step S1 may be selected and transmitted. good.
[0054]
In step S7, the input real-time data is transmitted through the selected HTTP connection. If the transmission fails, the real-time data may be discarded as missing data and preparations may be made for transmitting the next data in step S2. Even when the transmission is successful, if redundant data transmission is required, the process may be repeated so as to move to step S6 again and transmit the same data through another HTTP connection.
In step S8, there is no input for a certain period of time, and the process ends.
[0055]
FIG. 9 is a flowchart showing the operation of the HTTP client.
FIG. 9 shows steps until the HTTP client 2 on the real-time data receiving side creates the HTTP connection 3, receives the real-time data, and outputs it. The details are described below.
[0056]
In step S11, a plurality of HTTP connections are created. Specifically, when the reception side process starts, the HTTP client 2 creates an HTTP connection 3 for receiving the real time data from the HTTP server 1 on the transmission side. The HTTP client 2 may actively establish a connection to the HTTP server 1 or may wait for a connection from the HTTP server 1. A plurality of HTTP connections 3 are created by the multiple connection generator 31. When one connection is created, the process moves to step S12. In step S14, other connections are scheduled and maintained to be created later with a time difference. Thereafter, when an HTTP connection once connected is disconnected for some reason, the connection is immediately reconnected so that a predetermined number or more of connections can always be used.
[0057]
In step S12, the process waits until data can be received from the HTTP connection, and receives data when it becomes possible. If data is received, the process proceeds to step S13. When a reception timeout occurs in all HTTP connections, if all HTTP connections are switched from the transmission side, the process proceeds to step S16. The timeout is detected by the connection connection / disconnection monitoring unit 34 of the connection management unit 24 of the HTTP client 2, and at this time, a no-communication timeout occurs based on the reference time set by the reference time setting unit 31. When disconnection is detected, the connection disconnected by the multiple connection generator 31 is re-established.
[0058]
In step S13, it is determined whether or not there is another connection. If there is another connection, in step S14, the other connection has the connection time set by shifting by the connection shifting connection time setting unit 35. Thus, after being scheduled to be created later with a time difference, all connections are automatically created. After the connection, the connection state shifted in time is maintained.
[0059]
When all the connections are created, the process proceeds to step S15, and the received data is output in step S15. Pseudo data for preventing no-communication timeout is discarded without being output. The output destination may be a playback device such as a monitor or a speaker directly connected to the HTTP client 2, or may be transferred to another device (or a decoder (decompression device)) on the network. When the output ends, the process returns to step S12.
[0060]
Here, since the order is changed between the real-time data since it is received through a plurality of HTTP connections, the order is corrected by the transmission unit 26, and the data is rearranged when output. The output is arranged in order by part. In this case, a buffer or queue is provided immediately before output, real-time data of a certain amount or period is accumulated, and the oldest data is output in order.
[0061]
In step S16, the process ends because there is no data to be received or the HTTP connection is disconnected.
[0062]
FIG. 10 is a diagram illustrating an example of a transfer state using a plurality of HTTP connections.
FIG. 10 is a diagram showing a transfer state using a plurality of HTTP connections for each HTTP connection. Respective HTTP connections 1 (51), 2 (52), 3 (53), 4 (54),... N (5n), n + 1 (5n + 1) are time differences T2-T1, T3-T2, ... Effective period 56 of all HTTP connections created from Tn + 1-Tn, from HTTP connection creation (connection) 54 to HTTP connection discard (disconnection) 55, or a part of the real time indicated by diagonal lines Real-time data transmission is performed during the data transmission period 58-1 to 58-n + 1. The disconnected connection indicated at 55 is immediately reconnected and maintained for transmission.
[0063]
As a result, the real-time data is in the HTTP connection 1 (51), 2 (52), 3 (53), 4 (54),... N (5n), n + 1 (5n + 1) in the connected state. Since it is guaranteed that the data flows through the continuous real-time data transmission periods 58-1 to 58-n + 1, transmission without interruption is possible.
[0064]
Here, the HTTP connections 1 (51), 2 (52), and 3 (53) are set so that no communication timeout occurs based on the reference time set by the reference time setting unit 31 and the connection is not disconnected. 4 (54),... N (5n), n + 1 (5n + 1) HTTP connection valid period 56 or real-time data transmission period 58-1 to 58-n + 1 is set within the reference time. The
[0065]
Also, in order to prevent the connection disconnection unit 33 from disconnecting due to a no-communication timeout based on the reference time set by the reference time setting unit 31, the HTTP connections 1 (51), 2 (52) ), 3 (53), 4 (54),... N (5n), n + 1 (5n + 1) HTTP connection real-time data transmission periods 58-1 to 58-n + 1 other than the valid period 56 May send certain pseudo data.
[0066]
The real-time data transmission periods 58-1 to 58-n + 1 indicated by hatching indicate the real-time data transmission periods and do not represent individual packets. As a method for transmitting real-time data, data is written on the HTTP connection in units of real-time data within the real-time data transmission period 58-1 to 58-n + 1. In this case, normally, send () of TCP is executed. A certain amount of data is transmitted as a packet.
[0067]
FIG. 11 is a diagram for explaining the connection switching operation.
In FIG. 11, the connection management unit 22 of the HTTP server 1 responds to a request from the HTTP client 2 through the HTTP connection 1 (3-1) using the HTTP protocol managed by the application software for the HTTP server 1 to real-time data. The HTTP connection 1 (3-1) is suddenly disconnected by the connection disconnection unit 33 based on the reference time set by the reference time setting unit 31 based on the operation policy of the HTTP server 1 or the like. is there. At this time, transmission of the packets P1, P2, P3, P4, and P5 is stopped in the HTTP connection 1 (3-1). Further, the transmission of the packet P6 is stopped in the HTTP connection 2 (3-2) that has already been disconnected.
[0068]
When the connection connection / disconnection monitoring unit 34 detects disconnection of the HTTP connection 1 (3-1), the HTTP connection 1 (3-1) and HTTP connection 2 (3-2) disconnected by the multiple connection generation unit 31. Re-stretch. At this time, real-time data cannot be transmitted until the connection is reestablished and completely restored.
[0069]
Therefore, a plurality of HTTP connection 1 (3-1), HTTP connection 2 (3-2),... HTTP connection n (3-n) are generated in advance by the plurality of connection generation units 31 so that the application software can identify them. In addition, for each transmission period within each connection time set by shifting by the connection shifting connection time setting unit 35, the connection switching unit 36 selects another HTTP connection n (3-n), HTTP connection n + 1. Switch to (3-n + 1).
[0070]
As a result, packets P7, P8, and P9 are transmitted through HTTP connection n (3-n), packets P10, P11, and P12 are transmitted through HTTP connection n + 1 (3-n + 1), and subsequent packets P13, P14, and P15 are also transmitted. The data is transmitted through the HTTP connection 1 (3-1), the HTTP connection 2 (3-2),..., The HTTP connection n (3-n), the HTTP connection n + 1 (3-n + 1), which has been restored.
[0071]
Conventionally, due to abrupt disconnection by an HTTP proxy or HTTP server (or HTTP client), real-time data transmission by HTTP must accept the delay caused by re-establishing the HTTP connection, and this delay also affects subsequent packets. As a result, the real-time property of a large amount of data has been lost. However, if the embodiment of the present invention is used, it is possible to avoid the influence of delay caused by the re-establishment of the HTTP connection and to maintain the real-time property.
[0072]
In general, UDP has been mainly used for real-time data transmission, but it is difficult for UDP transmission to cross a firewall, but on the other hand, the means of transmission over a firewall using HTTP proxy is currently widely used. On the other hand, if the embodiment of the present invention is used, real-time data transmission across a firewall can be performed with higher quality.
[0073]
Embodiments of the present invention are not limited to those described above, but are not limited to HTTP, but also include HTTPS (Hypertext Transfer Protocol, Secure), FTP (File Transfer Protocol), SSL (Secure Sockets Layer), and TLS The present invention can be applied to various protocols used on TCP.
[0074]
【The invention's effect】
According to the present invention, by providing a mechanism capable of dynamically switching an HTTP connection and transmitting real-time data, even if the HTTP connection is disconnected due to an HTTP proxy or an HTTP server (or client) connection timeout, a delay due to connection reconnection It is possible to transmit real-time data without being influenced by the above.
[Brief description of the drawings]
FIG. 1 is a diagram showing a configuration of a real-time data communication system applied to an embodiment of the present invention.
FIG. 2 is a diagram showing a configuration of a real-time data communication system connected by HTTP proxy.
FIG. 3 is a diagram illustrating a configuration example of a transmission packet and a streaming data packet.
FIG. 4 is a diagram showing a configuration of an HTTP server.
FIG. 5 is a diagram illustrating a configuration of an HTTP client.
FIG. 6 is a diagram illustrating a configuration of a connection management unit.
FIG. 7 is a diagram showing identification of a plurality of connections.
FIG. 8 is a flowchart showing the operation of the HTTP server.
FIG. 9 is a flowchart showing the operation of an HTTP client.
FIG. 10 is a diagram illustrating an example of a transfer state using a plurality of HTTP connections.
FIG. 11 is a diagram illustrating a connection switching operation.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 ... HTTP server, 2 ... HTTP client, 3 ... HTTP connection, 4 ... HTTP proxy, 11 ... Transmission packet, 12 ... RTP header, 13 ... Sequence number SN, 14 ... Time stamp TS, 15... Streaming data packet, 21... Reception unit, 22... Connection management unit, 23... Reception unit, 24... Connection management unit, 26. Reference time setting unit, 33 ... Connection disconnection unit, 34 ... Connection connection / disconnection monitoring unit, 35 ... Connection shift connection time setting unit, 36 ... Connection switching unit, 41 ... OS, 42 ... Application software , 43 …… Playback software

Claims (12)

A server comprising: an input data receiving unit that receives input of real-time data; and an output unit that outputs the real-time data received by the input data receiving unit in units of divided data that is data having a predetermined data length; In a data communication system configured with a client having a divided data receiving unit that receives data input,
Either or both of the server and the client are
A plurality of connections are generated with different generation times, and one of the plurality of connections is selected as a transmission path for the divided data, and the selection is performed at predetermined time intervals or the selected A connection management unit is provided that switches when connection disconnection is detected and regenerates a connection that has been disconnected among the plurality of connections.
Data communication system. The connection management unit disconnects each of the plurality of connections when a predetermined reference time determined in advance from the start of generation of the connection.
The data communication system according to claim 1. Switching of the connection selection by the connection management unit is performed in units of divided data transmitted through the transmission path.
The data communication system according to claim 1. Selection of the connection by the connection management unit is sequentially performed based on the order in which the connections are generated.
The data communication system according to claim 1. The connection management unit newly selects another connection different from the connection when a delay packet is generated in the connection selected as the transmission path.
The data communication system according to claim 1. The connection management unit newly selects another connection different from the connection when the transmission failure of the divided data is detected from the server.
The data communication system according to claim 1. The output unit of the server transmits pseudo data in a section other than the transmission section of the divided data for all of the plurality of connections.
The data communication system according to claim 1. The connection management unit rearranges the divided data input to the divided data receiving unit in the order in the real-time data.
The data communication system according to claim 1. An input data receiving unit for receiving input of real-time data;
A plurality of connections are generated with different generation times, and one of the plurality of connections is selected as a transmission path for the divided data, and the selection is switched at predetermined time intervals, and A connection management unit that regenerates a connection for a disconnected connection among a plurality of connections,
An output unit that outputs the real-time data received by the input data receiving unit in units of divided data that is data having a predetermined data length, via the connection selected by the connection management unit;
server. A divided data receiving unit that receives input of divided data obtained by dividing real-time data for each predetermined data length output from the server;
A plurality of connections are generated with different generation times, and one of the plurality of connections is selected as a transmission path for the divided data, and the selection is performed at predetermined time intervals or the selected A connection management unit that performs switching when connection disconnection is detected and regenerates a connection that has been disconnected among the plurality of connections is provided.
client. Receiving real-time data input;
A plurality of connections are generated with different generation times, and one of the plurality of connections is selected as a transmission path for the divided data, and the selection is performed at predetermined time intervals or the selected Switching at the time of detecting the disconnection of the connection, and regenerating the connection for the disconnected connection among the plurality of connections;
Outputting the received real-time data in units of divided data, which is data having a predetermined data length, via the selected connection.
Data communication method. A step of receiving input of divided data obtained by dividing real-time data output by a server for each predetermined data length;
A plurality of connections are generated with different generation times, and one of the plurality of connections is selected as a transmission path for the divided data, and the selection is performed at predetermined time intervals or the selected Switching at the time of detecting disconnection of the connection, and regenerating the connection for the disconnected connection among the plurality of connections.
Data communication method.

Images