JPH0787133A - Buffer control method for data communication node - Google Patents

Abstract

PURPOSE:To secure the required number of buffers by terminal group and to make an effective use of a buffer. CONSTITUTION:A table group 2 consists of the number of using buffers of the entire node and for each communication protocol, table managing the buffer using rate state, and table setting the congestion threshold value of the entire node and for each communication protocol. A communication protocol control part 5 secures a buffer 3 and controls the data communication of terminal groups of each protocol P1 to Pn. A buffer control part 1 refers to the table group 2 and when the buffer using rate of the entire node exceeds the entire congestion threshold value, the buffer using rate of each protocol directs a congestion control part 7 to control the input of the protocol terminal which exceeds the congestion threshold value of the corresponding protocol.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a buffer control method for a data communication node, and in particular, it accommodates a plurality of terminals that support various communication protocols, applications or business contents.
In a packet switching network, etc. that regulates the input of data communication based on the buffer usage rate, secure the minimum required number of buffers for each terminal group that is classified based on the communication format such as communication protocol, application or business content. Also, the present invention relates to a buffer control method for a data communication node suitable for effectively using the buffer.

[0002]

2. Description of the Related Art As a conventional buffer control method in a data communication node or the like, there is a method of managing the total number of buffers used by having a common buffer for all groups according to the congestion state, and a fixed buffer area for each group. There are a method of allocating and managing the buffer usage amount for each group, a method of equally allocating an entire common buffer to connected terminals, and a method of managing the buffer usage number for each terminal.

As a method for managing the total number of buffers used, which has a common buffer for all, Japanese Patent Laid-Open No. 1-32 is known.
No. 0559, JP-A-3-216749 discloses a method for fixedly allocating a buffer area for each group and managing the buffer usage amount for each group. All common buffers are equally allocated to connection terminals. As a method of managing the number of buffers used for each
No. 352 publication is mentioned.

[0004]

In the prior art, according to the method disclosed in Japanese Patent Laid-Open No. 1-320559, when congestion occurs on the whole, communication is restricted from a terminal having a low priority, so that the priority is high. When a terminal uses a large amount of buffers, there is a problem that communication of almost all terminals is restricted, including terminals with low priority that rarely use buffers. In addition, JP-A-3
In the methods described in Japanese Patent Laid-Open No. 216749 and Japanese Patent Laid-Open No. 1-112352, a buffer is allocated in advance for each system or terminal, so that the amount of buffer used or the number of buffers used is limited. Therefore, when congestion occurs when only one system or terminal uses a large amount of buffer and the other system or terminal uses few buffers, the system or terminal uses only the allocated buffer. It will not be possible. That is, there is a problem that the buffer cannot be effectively used.

As described above, in the conventional buffer control method in the data communication node or the like, when congestion occurs, the data communication of the terminal that rarely uses the buffer is regulated, or there is an empty buffer. However, there was a problem that the buffer could not be used effectively.

It is an object of the present invention to, in a data communication node accommodating a plurality of terminals, restrict a data communication input only to a terminal group having a high buffer usage rate when congestion occurs so that the buffer usage rate can be reduced. Effective use of buffers by guaranteeing minimum communication of terminal groups with low bandwidth (guaranteing minimum number of buffers) and allowing uneven usage of buffers for each group until congestion occurs in the entire node It is to provide a buffer control method for realizing the above.

[0007]

The buffer control method of the present invention manages the buffer usage rate of the entire data communication node and the buffer usage rate of each group, and the buffer usage rate of the entire node up to the overall congestion threshold. However, if the overall congestion threshold is exceeded, the buffer usage rate for each group is used to check the buffer usage rate for each group. The data communication input is regulated to the terminals of the group that exceed the threshold value, and the terminals of the group that do not exceed the congestion threshold for each group and the terminals of the group that is less than the buffer usage guarantee value specified for each group. In other words, the input restriction of data communication is not enforced.

[0008]

According to the present invention, when the buffer usage rate of the entire data communication node does not exceed the congestion threshold value, even if the buffer is used in each group more than the congestion threshold value specified for each group, The terminals of the group are not judged to be subject to the input restriction of data communication. However, when the buffer usage rate of the entire node exceeds the congestion threshold value, the buffer usage rate of each group is used, so the terminals in the group that are using the congestion threshold value specified for each group or more Is determined to be subject to the input restriction. However, the terminals of the group that are equal to or less than the congestion threshold defined for each group are not judged to be subject to the data communication input restriction. On the other hand, after determining that congestion has occurred, when the buffer usage rate for each group falls below the congestion threshold specified for each group, it is determined that the input restriction for data communication for that group has been released, and the buffer usage rate for the entire node Is below the congestion threshold, it is judged that the input restriction of the data communication of all groups is released.

[0009]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings by taking a packet switching system as an example.

FIG. 1 shows a block diagram of a packet switching system according to an embodiment of the present invention. In FIG. 1, PS1 and P
S2 is a packet switching node (data communication node) having a packet switching function, and connects the packet switching nodes by a trunk line T1. PT11-PT1
n and PT21 to PT2n are terminals connected to the packet switching nodes PS1 and PS2, respectively. Between the packet switching node PS1 and the terminals PT11 to PT1n are subscriber lines L11 to L1n, and the packet switching nodes PS2 and PT are connected.
21 to PT2n are connected by subscriber lines L21 to L2n, and communication is performed using the communication protocols P1 to Pn, respectively. The packet switching node PS1 includes a buffer control unit 1, various table groups 2, a transmission / reception data buffer 3, a congestion control unit 4, and a communication protocol control unit 5 corresponding to each communication protocol. The same applies to the packet switching node PS2. The buffer control unit 1 uses each communication protocol P
1 to Pn, buffer control such as allocation and release of the buffer 3 and a control unit for monitoring the buffer utilization rate of the entire packet switching node and each communication protocol, the congestion control unit 4 receives data according to an instruction from the buffer control unit 1 when congestion occurs. The control unit for instructing the communication protocol control unit 5 to control the input of communication, and the communication protocol control unit 5 includes the communication protocols P1 to P1.
It is a control unit that controls data communication with a Pn terminal.

FIG. 2 shows a configuration example of the buffer control unit 1, the table group 2 and the buffer 3. In FIG. 2, the buffer usage number management table 21 is a table that manages the buffer usage number for the entire packet switching node and for each of the communication protocols P1 to Pn, and is composed of the buffer usage number and the allocated total number of buffers. The buffer management table 22 is a table that manages an unused buffer (empty buffer), and is composed of a leading pointer and an end pointer of the empty buffer. The empty buffers of the transmission / reception data buffer 3 are each chained by a pointer. The buffer usage rate state table 23 is a table for managing the buffer usage rate states for the entire packet switching node and for each of the communication protocols P1 to Pn. 1 "
Is set. The congestion threshold table 24 is a table in which congestion thresholds are set for the entire packet switching node and each of the communication protocols P1 to Pn. Here, the congestion threshold value is set to 80 (%) in all cases, and if it is less than this value, the congestion is released, and if it is more than this value, the congestion is generated.

The buffer control unit 1 is roughly divided into a buffer usage number management unit 10 and a usage rate monitoring unit 15. The buffer usage number management unit 10 further includes an initialization processing unit 11, a buffer allocation processing unit 12, and a buffer release processing unit. It consists of 13. The initial setting processing unit 11 responds to an initial setting request from the system.
This is a part for performing chain creation processing of the buffer 3 and initialization of the buffer usage number management table 22. The buffer allocation processing unit 12 is a unit that, in response to the buffer allocation request issued by each communication protocol control unit 5, extracts the head buffer of the empty buffer from the buffer management table 21 and allocates it to the request source. The buffer release processing unit 13 is a unit that obtains a final empty buffer from the buffer management table 22 in response to the buffer release request issued by each communication protocol control unit 5, and then incorporates the release buffer. The usage rate monitoring unit 15 is a unit that determines whether or not the buffer usage rates of the entire packet switching node and the communication protocols P1 to Pn exceed a congestion threshold. The usage rate monitoring unit 15 is periodically activated.

Conventionally, when the buffer usage rate of the entire packet switching node is monitored to regulate the input of data communication, for example, the terminals P11 and P21 of the communication protocol P1 use a large amount of buffers and the buffer usage number management table 21 If the total buffer usage rate obtained by dividing the total number of buffer usages of the packet switching node by the total number of buffers exceeds the congestion threshold, terminals P12 to P of other communication protocols are used.
1n and P22 to P2n are also subject to data communication input regulation. In addition, when the number of buffers used is assigned in advance for each communication protocol and the usage rate is monitored, only the assigned buffers can be used, so that the buffers cannot be effectively utilized. The present invention manages the buffer usage rate of the entire packet switching node and the buffer usage rate of the communication protocols P1 to Pn (for each group), and uses the buffer usage rate of the entire packet switching node up to the overall congestion threshold. When the overall congestion threshold is exceeded, the buffer usage rate of each communication protocol is used, and the input restriction of data communication is applied to the terminals of the communication protocol whose buffer usage rate of a certain communication protocol exceeds the congestion threshold. By controlling the other terminals that do not exceed the congestion threshold for each communication protocol and the terminals that are below the buffer usage guarantee value specified for each communication protocol and do not restrict the input of data communication. , It guarantees the minimum number of buffers required for each communication protocol and enables effective use of buffers.

The operation of the buffer controller 1 according to the present invention will be described below with reference to FIGS. 1 and 2 according to the flow charts of FIGS.

FIG. 3 is a flow chart of the buffer usage number management unit 10. Steps 111 to 113 are the initialization processing unit 11, steps 121 to 123 are the buffer allocation processing unit 12, and steps 131 to 133 are the buffer release processing unit 13. It corresponds to each processing of.

When the packet switching nodes PS1 and PS2 are started up, an initial setting request is made to the usage number management unit 10 of the buffer control unit 1 on each node.

The usage number management unit 10 determines the request type (step 100). In the case of the initial setting request, the chain of the buffer 3 is created, and the start pointer and the end pointer of the buffer management table 22 are set (step 11).
1). Next, the number of buffer chains created in step 111 is set as the total number of buffers to the total number of buffers in the buffer usage number management table 21 (step 11).
2). Next, the total number of buffer allocations for each of the communication protocols P1 to Pn is calculated, and the buffer usage number management table 21
Is set to the total number of buffer allocations for each communication protocol (step 112).

The buffer allocation total number Bi for each communication protocol is calculated, for example, by the following equation 1. Bi = Si / S · Bn (Formula 1) Si: Total line speed or protocol speed class of terminals supporting the communication protocol Bn: Total number of buffers obtained in step 111 S: Lines of all terminals Total of speed or protocol speed class Here, the line speed and protocol speed class are data that can be given from outside as configuration data.

Next, the terminals PT11 to PT1n and the terminal PT
When communicating with 21 to PT2n, a buffer is always required on the packet switching nodes PS1 and PS2. At that time, the communication protocol control unit 5 makes a buffer allocation request to the used number management unit 10 of the buffer control unit 1.

The usage number management unit 10 determines the request type (step 100). In the case of an allocation request, the head buffer of the empty buffer is obtained from the buffer management table 22,
It is removed from the buffer chain and the pointer of the next empty buffer is set as the head pointer of the buffer management table 22 (step 121). Next, the total buffer usage number in the buffer usage number management table 21 is updated (+1).
(Step 122). Finally, the number of buffers used for each communication protocol of the communication protocol of the buffer number management table 21 is updated (+1) (step 12).
3).

When the buffer used for communication between terminals becomes unnecessary, the communication protocol control unit 5 requests the used number management unit 10 of the buffer control unit 1 to release the buffer.

The usage number management unit 10 determines the request type (step 100). In the case of a release request, the entire buffer usage number in the buffer usage number management table 21 is updated (-
1) Do (step 131). Next, the buffer usage number of the communication protocol of the communication protocol of the buffer usage number management table 21 is updated (-1) (step 1).
32). Finally, the final empty buffer is obtained from the buffer management table 22, the release request buffer pointer is set to the next buffer pointer, and the release request buffer pointer is set to the final pointer of the buffer management table 22 (step 133). .

FIG. 4 is a flow chart of the buffer utilization rate monitoring unit 15. The usage rate monitoring unit 15 is periodically activated on each node while the packet switching nodes PS1 and PS2 are operating. The buffer utilization state table 23 shows the states (congestion state, release state) of the entire packet switching node and each communication protocol at the time of the previous monitoring.

Upon activation, first, the buffer usage rate of the entire packet switching node is obtained by dividing the buffer usage number of the buffer usage number management table 21 by the total number of buffers (step 200). Next, if the obtained buffer utilization rate of the entire packet switching node exceeds the overall congestion threshold value of the congestion threshold value table 24, the congestion occurrence state is reached. If not, the congestion release state and the current buffer utilization rate are used. Judging the status, buffer usage status table 2
Check the change from the overall buffer utilization state of 3 (step 210). Then, if there is no change in the buffer usage rate state (congestion release state ⇒ congestion release state, congestion occurrence state ⇒ congestion occurrence state), go to step 211, and if the buffer usage rate state is congestion release state ⇒ congestion occurrence state, step 221. If the buffer usage rate status is congestion occurrence status ⇒ congestion occurrence status, the process proceeds to step 231.

If there is no change in the overall buffer usage rate state, it is judged whether the overall buffer usage rate state is the congestion release state or the congestion occurrence state (step 211). And
If the entire buffer usage rate is in the congestion release status, the processing is terminated, and if the buffer usage status is in the congestion occurrence status, the processing proceeds to the processing according to the buffer usage rate for each communication protocol after step 212.

First, it is judged whether or not the processing for all communication protocols is completed (step 212), and if completed, the processing is ended. If not completed, the buffer usage rate of the communication protocol is obtained by dividing the buffer usage number of the communication protocol in the buffer usage number management table 21 by the total number of buffer allocations (step 21).
3). Next, it is checked whether the buffer usage rate of the communication protocol exceeds the congestion threshold value of the communication protocol in the congestion threshold value table 24, and the current buffer usage rate state (congestion state, release state) is determined. A determination is made and a change from the buffer utilization rate state of the communication protocol in the buffer utilization rate state table 23 is checked (step 214). Then, if there is no change, the process directly returns to step 212. When congestion is changed to a congestion release state, "0" is set to the buffer use rate state of the communication protocol of the buffer use rate state table 23 (step 215), and the congestion control unit performs the congestion release process of the communication protocol. 4 is notified (step 216). Further, when the state is changed from the congestion release state to the congestion occurrence state, "1" is set to the buffer usage rate state of the communication protocol of the buffer usage rate state table 23 (step 217), and the congestion regulation process of the communication protocol is performed. Notify the congestion control unit 4 (step 218). Then, the process returns to step 212.

When the overall buffer usage rate state is the congestion release status ⇒ congestion occurrence status, "1" is set to the overall buffer usage rate status in the buffer usage status table 5 (step 221). Then, the process proceeds from step 222 onward according to the buffer usage rate for each communication protocol.

First, it is judged whether or not the processing for all communication protocols is completed (step 222), and if completed, the processing is ended. If not completed, the buffer usage rate of the communication protocol is calculated by dividing the buffer usage number of the communication protocol in the buffer usage number management table 21 by the total number of buffer allocations (step 22).
3). Next, it is checked whether the buffer usage rate of the communication protocol exceeds the congestion threshold of the communication protocol in the congestion threshold table 24 (step 224). Returns to step 222. On the other hand, if the congestion threshold is exceeded, "1" is set to the buffer utilization state of the communication protocol of the buffer utilization state table 23 (step 225), and the congestion control process of the communication protocol is changed to the congestion control unit. 4 is notified (step 226). Then, the process returns to step 222.

If the overall buffer utilization state is the congestion occurrence state → congestion release state, "0" is set to the overall buffer utilization state of the buffer utilization state table 23 (step 231). Next, the cancellation process for each notification protocol after step 232 is performed.

First, it is judged whether or not the processing for all communication protocols is completed (step 232), and if completed, the processing is ended. If not completed, it is checked whether the buffer usage rate of the communication protocol in the buffer usage rate table 23 is in the congestion occurrence state (step 2).
33). Then, in the case of the congestion release state, step 232
Return to. On the other hand, in the case of the congestion occurrence state, "0" is set to the buffer usage rate state of the communication protocol of the buffer usage rate state table 23 (step 234), and the congestion control unit 4 is notified of the congestion release processing of the communication protocol. (Step 235). Then, the process returns to step 232.

As can be seen from the above description, in the present embodiment, the buffer utilization rate state is monitored for each communication protocol, and when the buffer utilization rate state of the entire packet switching node becomes a congestion state, each communication protocol state is monitored. It becomes possible to regulate the input of data communication only for the terminal of the communication protocol in which the buffer usage rate state of each is in the congestion state. Therefore, it is possible to guarantee communication (a minimum number of buffers) of a terminal of a communication protocol having a low buffer usage rate. Further, it is possible to realize the effective use of the buffer by allowing the number of buffers used in each communication protocol to be offset until the buffer utilization rate of the entire packet switching node becomes congested.

When determining whether the buffer usage rate state of each communication protocol is the congestion occurrence state, it is not always necessary to make a determination based on the congestion threshold value, and a communication protocol that can be externally given as configuration data is provided. It is obvious that the determination may be made based on the buffer use guarantee value (number of buffers) for each.

Although the case where the present invention is applied to the packet switching system has been described above, it is obvious that the present invention is not limited to the packet switching system but can be widely applied to all data communication systems. Further, the grouping of the terminals is not limited to the communication protocol, but may be classified based on the purpose or the business content.

[0034]

As described above, according to the present invention,
In a data communication system such as a packet switching network, in addition to the buffer usage rate of the entire data communication node, to monitor the buffer usage rate for each terminal group classified based on the communication mode such as communication protocol, application or business content. ,
Even if congestion occurs in the entire node when there is a bias in the buffer usage rate between each terminal group, terminals in the terminal group with a low buffer usage rate are not subject to data communication regulation, so that The minimum communication of can be guaranteed. Further, until the congestion occurs in the entire data communication node, the bias of the buffer usage rate between the terminal groups is allowed, so that the buffer can be effectively used.

[Brief description of drawings]

FIG. 1 is a system configuration diagram of an embodiment of the present invention.

2 is a diagram illustrating a buffer control unit, various tables in FIG.
It is a figure which shows the structural example of a buffer.

FIG. 3 is a processing flowchart of a buffer usage number management unit in FIG.

FIG. 4 is a processing flowchart of a buffer usage rate monitoring unit in FIG.

[Explanation of symbols]

PS1, PS2 Packet switching node (data communication node) PT11 to PT1n, PT21 to PT2n Terminal 1 Buffer control unit 2 Table group 3 Transmission / reception data buffer 4 Congestion control unit 5 Communication protocol control unit 10 Buffer usage number management unit 15 Buffer usage rate monitoring Part 21 Buffer usage number management table 22 Buffer management table 23 Buffer usage rate state table 24 Congestion threshold table

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Toru Nishiyama, Tohru Nishiyama 810 Shimoimaizumi, Ebina, Kanagawa Prefecture Office Systems Division, Hitachi, Ltd. (72) Kenichi Aoki 6-81, Onoue-cho, Naka-ku, Yokohama, Kanagawa Hitachi Software Engineering Stock Association In-house

Claims (3)

[Claims] 1. A buffer control method in a data communication node accommodating a plurality of terminals, wherein the plurality of terminals are divided into a plurality of groups, and a data transmission / reception buffer is variably assigned to each group, and the entire data communication node is provided. Set the buffer usage threshold (hereinafter referred to as the overall congestion threshold) and the buffer usage threshold for each group (hereinafter referred to as the congestion threshold for each group) in advance, The buffer usage rate of the entire data communication node and the buffer usage rate of each group are periodically monitored, and if the buffer usage rate of the entire data communication node does not exceed the overall congestion threshold, whichever group Do not restrict the data communication even for the terminal of the above, and the buffer usage rate of the entire data communication node is congested as a whole. If the value exceeds a certain value, the buffer control of the data communication node is characterized in that the data communication is regulated to the terminals of the group whose buffer usage rate for each group exceeds the congestion threshold of the group. Method. 2. The buffer control method for a data communication node according to claim 1, wherein a plurality of terminals accommodated in the data communication node are grouped based on a communication form such as a communication protocol, a purpose or a work content. A method for controlling a buffer of a data communication node, characterized by: 3. The buffer control method for a data communication node according to claim 1, wherein the buffer utilization rate for each group is a line speed defined for each line in the group or a protocol speed class defined for each communication channel. Is calculated for each group, and is calculated from the total number of buffers assigned to each group based on the ratio and the number of used buffers counted for each group.