JPS632439A - Burst communication system - Google Patents

Abstract

PURPOSE:To reproduce unperiodically produced burst information, while maintaining its produced time interval, by transmitting burst produced time, etc., together with the burst information by means of an originating node and reading them by means of an incoming node. CONSTITUTION:At an originating node the produced and terminated times of each burst 8, 9, and 10 are known by means of an absolute time determining means and the known times are added to header information 8-1, 9-1, and 10-1 and footer information 8-2, 9-2, and 10-2 parts. If received burst information is reproduced as it is at a terminating node, the information produced interval at the originating node cannot be maintained, since the burst information respectively have different delays (d1, d2, and d3). At the receiving node, therefore, the produced time information t1, t2, and t3 at the header parts are read and the readout of data are adjusted so as to absorb the fluctuation of the delays. Finally, the received data are reproduced in such a way that the reproduced data can have a fixed permissible delay K containing unperiodical bursts.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to a burst communication method (conventional (Problems to be Solved by the Technology and Invention) Conventional communication systems had the following drawbacks from the perspective of preserving the temporal order of information when handling information that occurs in bursts. .

In time division multiplexing (TDM), communication channels are allocated periodically, so information output regularly from a terminal is Almost preserved. but,
For information that is generated in bursts, that is, with irregular changes in the information generation interval, sampling interval, and amount of information per sample, in the TDM system, the information is constantly sent periodically to the corresponding channel in the communication network. is transmitted or receives information periodically through the channel, which is extremely inefficient. This will be a big problem, especially in future networks that handle high-speed burst information.

On the other hand, packet communication systems essentially allow for burst information transmission. However, in conventional packet communication systems, although the relative order of the information sequences included in each packet is preserved, the generation time of each piece of information cannot be determined even if fluctuations in delay time within the communication network are absorbed by a buffer. .

Therefore, if the sampling rate of the information source is fixed or regular, the information generation time can be reconstructed relatively from the packet order on the receiving side, but if the information source is completely bursty, For example, the sampling rate is variable, making it difficult to reproduce information accurately in time due to fluctuations in delays and data loss in transmission and processing.

Furthermore, with conventional packet communication methods, real-time communication may not be satisfied due to large fluctuations in packet delay time caused by processing waiting time, etc. To avoid this, information that requires strong real-time performance may not be satisfied. Methods of prioritizing packets are being considered so that the loaded packets can be transmitted with priority. However, when transmitting information that occurs in bursts as variable-length bursts containing a large amount of data at high speed, fixed priorities such as those used in packets do not improve communication quality from the perspective of real-time communication. It is difficult to set the optimal priority that satisfies the following.

The present invention provides means for enabling communication that preserves the information generation interval, which is difficult to achieve in conventional packet communication, without causing a decrease in efficiency as in TDM, for burst information that occurs irregularly. Furthermore, by using the means of storing information generation intervals, it is possible to achieve fine-grained information according to the requirements for real-time handling of information and the traffic conditions within the network, which was difficult to achieve with conventional packet communications. , and realizes highly efficient priority control.

(Means for Solving the Problems) The present invention has a means for determining the absolute time when information is generated at the information source or the relative time from a reference point, and By adding time information, even if the information is generated non-periodically, or is sampled non-periodically, or if there are large fluctuations in delay time within the network, the receiving side can The most important feature is that it enables burst communication in which information is reproduced correctly. Conventional packet communication technology does not consider transmitting the irregularly occurring burst information as described above in a burst manner, so the received signal is The difference was that it was playing the .

Furthermore, as a type of time information, the present invention provides information representing the time margin required to reach the target, such as the target absolute time to reach the target node, based on the delay time conditions (real-time nature) requested by the information source to the network. By adding this to a burst, each node has the feature of being able to control fine-grained priority according to the processing status of the burst within the network. This differs from conventional packet communications in that even if priorities are used, they are not changed according to the processing status within the network.

Furthermore, by using the priority control described above, bursts with a clear delay can be rejected at intermediate nodes in the network. This differs from conventional packet communications in that they are discarded after reaching the last node, and enables highly efficient control by preventing invalid data from flowing within the network in real time.

(Example) The present invention will be explained in detail below.

First, as a first embodiment of the present invention, a burst communication method using determination of absolute time will be described.

First, the concept of time determination according to this embodiment will be explained with reference to FIG. In FIG. 1(a), 1 is a master clock in the central network, and 2 to 4 are clocks that are synchronized with the master clock 1 and receive absolute time information. The positional relationship between master clock 1 and clocks 2 to 4 on the time axis is as shown in FIG. 1(b). 5～
7 is a burst communication node synchronized with clocks 2 to 4;
In this example, 5 is the originating node, 6 is the relay node, and 7 is the destination node.

By distributing absolute time in consideration of transmission delay in a network synchronization system, it is technically possible to reduce the value of Te shown in FIG. 1(b) to 1 m5ec or less.

Next, a first embodiment will be described using an example of burst information generated at the transmitting node 5 having such absolute time determination means.

FIG. 2(a) is an example of burst information generated at the transmitting node 5. Burst 8 is from time t1 to t'1, burst 9 is from t2 to t"2, burst 10 is from t"
Continue from 3 to t°3.

The transmitting node 5 knows the occurrence time and end time of each burst by means of absolute time determination, and stores them in header information 8-1.9-1゜lo-1 and footer information 8-2.9-2.10. - Add to 2 parts. Figure 2 (b
) are occurrence time information 8-1a, 9-1a and 1 added to header information 8-1゜9-1 and 1O-1, respectively.
0-1a is illustrated.

Since the burst information received by the receiving node 7 generally has different delays (dl, d2, d3) depending on traffic conditions and routing, it is not possible to maintain the information generation interval at the transmitting node 5 if the information is reproduced as is. Therefore, at the receiving node 7, the occurrence time information 1. ．． Read 12°t3 and adjust the final data reading time to absorb delay fluctuations within the allowable delay range at 8-3.9-3°l in Figure 2 (C).
(o-3 part), the received data is finally regenerated so that the aperiodic burst has a certain allowable fixed delay (FIG. 2(d)).

Although absolute time is used as time information in this embodiment, more generally time information indicating a difference from the time at a reference point within a network may be used instead. The burst end time in FIG. 2 is valid in the second and third embodiments.

As a second embodiment of the present invention, in addition to the means in the first embodiment, the margin of delay time for each burst to reach the receiving node is evaluated using absolute time, and priority is given. This shows a method that has means for adding degrees.

FIG. 3 is an explanatory diagram of the second embodiment of the present invention, and FIG.
In a), 11 is the originating node, 12 is the relay node, 1
3 is a receiving node, 14 is a transmitting terminal, and 15 is a receiving terminal. FIG. 3(b) shows burst data 1 generated at time t4 in the originating node 11 and ended at time tI4.
6 is shown. The originating node 11 includes occurrence time information (ta) 16 in the header information 16-1 of the burst data 16.
-1a is added, but information 16-1b regarding the permissible delay time until the arrival/peak time or the permissible arrival time (1,) is added based on the real-time requirement roughly determined by the class of the information source ( Figure 3 (C)).

The relay node 12 evaluates the real time margin for the burst data 16 based on the time t5 when the burst data 16 arrives at the relay node 12, the allowable arrival time III j o , and the destination node address. (Figure 3(d)). For example, y=jo
t5 te=to −T, (1) where, t,,: Expected transmission delay from relay node 12 to destination node 13, Tp=t5+t,: Estimated time of arrival at destination node 13, y: Real time margin, shall be.

The priority for burst data 16 is changed as follows depending on the value of y.

■If y≧Th, the priority is set to normal.

However, Th is a value of y that satisfies real-time performance with normal priority.

■If Th2≧y, reject the burst data.

However, Th2 is a value of y such that it is determined that the data reproduced at the destination node 13 clearly does not satisfy real-time characteristics.

■If Th,>y>Th2, raise the priority higher than normal.

However, in this case, multiple priorities can be further set depending on the level of y.

The relay node 12 can perform the above processing simply by checking the header information +6-1.

(Effects of the Invention) As explained above, according to the present invention explained based on the first embodiment, if the originating and terminating nodes recognize the absolute time, the originating node can determine the burst occurrence time. etc., along with the burst information, and read it at the destination node, this can be used to transmit the burst information that occurs non-periodically, despite fluctuations in delay within the network.
Reproduction is possible while maintaining the generation time interval.

Further, according to the present invention described based on the second embodiment, each node has means for determining the time, and the originating node has means for adding burst occurrence time information and permissible arrival time information. By checking the above information included in the header, the relay node can perform processing with different priorities without accumulating burst data, and each relay node can perform adaptive processing from the viewpoint of maintaining real-time performance. Can perform priority processing. Furthermore, due to real-time nature, invalid data is immediately discarded at the node where it becomes clear. By combining these, extremely efficient communication becomes possible.

[Brief explanation of the drawing]

FIG. 1 is a conceptual diagram of time determination according to an embodiment of the present invention, and FIG.
The figure is an explanatory diagram of the first embodiment of the invention, and FIG. 3 is an explanatory diagram of the second embodiment of the invention. 1: Master clock in the network, 2 to 4: Clock in the network, 5 to 7: Nodes in the network, 8: Burst data generated at time t at the originating node, 9: Burst data generated at time t2 at the originating node Burst data generated at time t3 at all originating nodes, s-t, 9-1.10-1 + header information, 8-1a,
9-1a, 10-1a: Occurrence time information, 8-2.
9-2.10-2: Footer information, 11: Originating node,
12: Relay node, 13: Receiving node, 14: Originating terminal, 15: Receiving terminal, 16: Burst data generated at time t4 at terminal 14, 16-1+header information, 16-2: Footer information, 16-
1a: Occurrence time information.

Claims (3)

[Claims] (1) Examples of codes indicating the beginning and end of data and for transmitting these data in order to transmit variable speed, variable length data blocks that are generated aperiodically or partially periodically from an information source In a burst communication method in which a node within the network processes a burst to which information that the network should identify has been added, the originating node is provided with a first and a second means for adding the time information as a relative value or an absolute value to the burst data, and the destination node reads the time information added to the burst data to determine the arrival time of the received burst data. A burst communication system characterized by providing a third means for reproducing burst information while absorbing fluctuations and maintaining correct temporal relationships. (2) Examples of codes indicating the beginning and end of data and the transmission of these data in order to transmit variable-speed, variable-length data chunks that are generated aperiodically or partially periodically from an information source In a burst communication method in which a node within the network processes a burst with information that the network should identify, etc., the originating node is provided with a a second means for adding the above-mentioned time information as a relative value or an absolute value to the burst data; and a goal to be reached to the communication partner by identifying the conditions regarding the delay required by the information source that generated the information to the network. A fourth means is provided for adding information representing a time margin regarding delay from the viewpoint of time or real-time to the burst data, and the receiving node reads the time information added to the burst data to transmit the received burst data. A third means for reproducing burst information while absorbing fluctuations in the arrival time of burst data and maintaining a correct temporal relationship, and the estimated transmission time of the burst occurrence, arrival time at a node, and from the node to the target receiving node. A burst communication method characterized by providing a fifth means for raising the priority of bursts with less time margin from the normal priority by evaluating the real time margin in consideration of time. (3) Among bursts with little time margin at each node, the estimated time Tp of arrival at the destination node is expected to be clearly delayed by t compared to the required target arrival time To, and t is expected to arrive at the destination node. A patent characterized in that the burst is rejected if the delay exceeds a tolerance value determined to be necessary for reproducing data while maintaining temporal continuity at a node, that is, while maintaining real-time property. A burst communication method according to claim 2.