JP2000305876A - Method for monitoring connection activity - Google Patents

Abstract

PROBLEM TO BE SOLVED: To decrease frequency in the transmission of monitor signals by sending the monitor signals normally only to some of a plurality of connections with one computer in such a case. SOLUTION: Connections 13, 14, and 15 are established between computers 10 and 11 and the computer 10 stores connection identifiers in the activity monitor management table corresponding to the IP address of the computer 11 (500). Then the monitor signal is sent normally to one connection selected out of the three connections which are already made with the computer 11 (501). If a fault is detected at the connection 13 (501), the monitor signal is sent even to the connections 14 and 15 which are established with the computer 11 (503). Consequently, the frequency of monitor signal transmission is reduced to 1/n.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to a connection activity monitoring method for a communication system.

[0002]

2. Description of the Related Art TC is one of the connection activity monitoring methods.
P / IP (Transmission Control)
Protocol / Internet Protocol
1) There is keepalive. Keep-alive sends a monitoring signal when there is no data to be sent to the communication destination computer for a certain period of time, and the computer receiving the monitoring signal returns a response signal if the connection status is normal, and if the response signal is not returned Determines that a connection failure has occurred.

When the keepalive is used, the network load increases due to the transmission and reception of the monitoring signal and the response signal, and the communication cost in the network for charging the packet increases. Therefore, when the keep-alive is implemented by the RFC 1122, it is defined that the application program can select whether or not to perform the keep-alive for each connection.

[0004]

In a communication system used in a backbone business or the like, it is necessary to monitor the activity of a connection from the viewpoint of reliably detecting a failure. In the conventional method, a monitoring signal is transmitted over all connections. And the network load and communication cost increase.

[0005]

Means for Solving the Problems Since a connection failure mainly occurs due to a network failure or a computer failure, when a plurality of connections are established with one computer, it is necessary to monitor the activity of all the connections. Sex is low. Therefore, the established connection is managed, and when there are a plurality of connections with one computer, a monitoring signal is transmitted only for a part of the connection in a normal state, and the number of times of transmitting the monitoring signal is reduced. When a failure is detected in any of the connections, a monitoring signal is transmitted for all the connections established with the failed computer, and the failure is detected in all the connections.

[0006]

Embodiments of the present invention will be described below.

FIG. 1 is a diagram showing a configuration example of a network according to an embodiment of the present invention. The computer 10 is connected to the computer 11 via a network 12. The connection 13, the connection 14, and the connection 15 are all connections established from the computer 10, and the TCP port 101 and the TCP port 102 of the computer 10 are connected to the TCP port 111 of the computer 11,
Is the TCP port 103 of the computer 11
2 is connected.

FIG. 2 is a configuration diagram of a computer. 2, the computer includes a monitoring control unit 21 and a protocol processing unit 23. The protocol processing unit 23 performs communication processing such as connection establishment, data transfer, and connection release. The signal processing unit 24 in the protocol processing unit 23
It sends and receives monitoring signals and response signals. Monitoring control unit 21
Controls the transmission of the monitoring signal using the signal processing unit 24. The activity monitoring management table 22 is a table for managing an established connection, and exists for each IP address of a communication destination computer.

As shown in FIG. 2, the activity monitoring management table 2
Reference numeral 2 denotes an IP address storage area 221 for storing an IP address of a communication destination computer, and a monitoring candidate list 222 for storing connection identifiers of computers and established connections stored in the IP address storage area 221.

FIGS. 3 and 4 are flowcharts showing the processing of the monitoring control unit 21 and the signal processing unit 24, respectively.

Next, the operation of each unit shown in FIG. 2 will be described with reference to the flowcharts shown in FIGS.

In FIG. 3, when the protocol processing unit 23 establishes a connection, it notifies the monitoring control unit 21 of the connection establishment (step 300).
Refers to the IP address storage area 221, and refers to the activity monitoring management table 22 corresponding to the IP address of the communication destination computer.
(Step 301). If the search is successful (step 302), the connection identifier is registered at the end of the monitoring candidate list 222 (step 304). If the search fails because the corresponding activity monitoring management table does not exist, a new activity monitoring management table is generated, the IP address is set in the IP address storage area 221 (step 303), and the connection identifier is set in the monitoring candidate list 222. Register

The monitoring control unit 21 has a fixed interval (for example, 60
Timeout) in seconds. When the time-out is notified, a connection in a non-communication state is selected one by one from the monitoring candidate list 222 of each activity monitoring management table 22, and a monitoring signal transmission instruction for the selected connection is sent to the signal processing unit 24 (step 305). , Step 306, step 307). FIG. 6 is a flowchart illustrating a method of selecting a connection for which a monitoring signal transmission instruction is issued from the monitoring candidate list 222.

FIG. 6 shows a method for equally selecting a connection for instructing a supervisory signal transmission. First, the connection identifier at the head of the monitoring candidate list 222 is moved to the end of the list (step 600). Next, it is determined whether or not the first connection in the monitoring candidate list 222 is in a non-communication state (step 601).
2). If there is no communication state, the top connection identifier of the monitoring candidate list 222 is moved to the end of the list (step 600).

The above process is repeated until a connection in a non-communication state is found. However, when there is no connection in the non-communication state on the monitoring candidate list 222, the monitoring signal transmission instruction is not performed. By the above processing, the connection identifier of the selected connection is registered at the head of the monitoring candidate list 222, and the connection in the non-communication state can be uniformly selected. In addition to the above selection method, a method of judging the state of all connections on the monitoring candidate list and selecting a connection with which communication has not been performed for the longest time may be considered.

In FIG. 4, when the signal processing unit 24 receives the monitoring signal transmission instruction, the signal processing unit 24 transmits a monitoring signal to the instructed connection and starts a monitoring timer for monitoring a response signal (step 400, step 401, Step 402).

When the signal processing unit 24 receives the response signal,
The monitoring timer is stopped, and a normal completion report is sent to the monitoring control unit 21 (step 403, step 404, step 40).
5). When receiving the normal completion report, the monitoring control unit 21 does not transmit a monitoring signal for another connection (step 308).

When the timeout of the monitoring timer is notified to the signal processing unit 24 (step 406), a connection release process according to the TCP protocol is performed (step 407), and a failure notification of the connection is sent to the monitoring control unit 21. (Step 408). Upon receiving the failure notification (Step 309), the monitoring control unit 21 deletes the connection identifier of the failed connection from the monitoring candidate list 222 of the corresponding activity monitoring management table 22 (Step 310) and stores it in the monitoring candidate list 222. The control unit 24 instructs the signal processing unit 24 to transmit a monitoring signal for all of the connected connections (step 311).

Upon releasing the connection, the protocol processing unit 23 notifies the monitoring control unit 21 and the signal processing unit 24 of the release of the connection. When the connection release is notified to the monitoring control unit 21, the corresponding activity monitoring management table 22
The connection identifier is deleted from the monitoring candidate list 222 (step 312). When the connection release is notified to the signal processing unit 24, the monitoring timer is stopped if the connection is being monitored (steps 409 and 410).

Next, a sequence in the configuration of FIG. 1 will be described with reference to FIG. FIG. 5 shows a case where the computer 10 detects a failure. When the connection 13, the connection 14, and the connection 15 are established between the computer 10 and the computer 11, the computer 10
The connection identifier is stored in the activity monitoring management table corresponding to the P address (step 500).

Normally, a monitor signal is transmitted to one connection selected by the method shown in the flowchart of FIG. 6 among the three connections established with the computer 11 (step 501). When a failure is detected in the connection 13 (step 502), a monitoring signal is also transmitted for the connections 14 and 15 established with the computer 11 (step 503). With the above processing, failure detection can be performed on all connections.

In this embodiment, when a plurality of connections are established between the same computers, the monitoring signal is transmitted only by one connection at the same time. Therefore, when n connections are established between the same computers. Can reduce the number of times of transmission of the monitoring signal to 1 / n as compared with the case where the monitoring signal is transmitted through all connections as before. In addition, since it is possible to reduce the monitoring signal transmission interval to 1 / n of the conventional one without increasing the number of transmissions of the monitoring signal, by shortening the monitoring signal transmission interval, it is possible to prevent failure earlier than before. Can also be detected.

[0023]

According to the present invention, when a plurality of connections are established between the same computers, the number of times of transmission of the monitoring signal is reduced, so that the network load and communication cost required for detecting a failure can be reduced.

[Brief description of the drawings]

FIG. 1 is a diagram illustrating a configuration example of a network according to the present invention.

FIG. 2 is a diagram illustrating a configuration example of a computer according to the present invention.

FIG. 3 is a flowchart of a monitoring control unit in the embodiment.

FIG. 4 is a flowchart of a signal processing unit in the embodiment.

FIG. 5 is a diagram illustrating an example of a sequence diagram of activity monitoring in the embodiment.

FIG. 6 is a flowchart illustrating a method of selecting a connection to be monitored in the embodiment.

[Description of Signs] 10: Computer, 11: Computer, 101, 102, 10
3, 111, 112 ... TCP port, 12 ... network, 13, 14, 15 ... connection, 21 ... monitoring control unit, 22 ... activity monitoring management table, 221 ... IP address storage area, 222 ... monitored connection identifier storage area, 223 ... established connection identifier storage area, 23
... Protocol processing unit, 24 ... Signal processing unit, 300, 30
1,302,303,304,305,306,30
7, 308, 309, 310, 311, 312... Each processing step in the flowchart, 400, 401, 402,
403,404,405,406,407,408,4
09, 410... Each processing step in the flowchart, 50
0, 501, 502, 503... Each processing step in the sequence diagram, 600, 601, 602... Each processing step in the flowchart.

Claims (2)

[Claims] In a connection activity monitoring method of a communication system, when a plurality of connections exist between computers, a fault is detected by transmitting a monitoring signal only for a part of the connections in a normal state. Connection activity monitoring method. 2. A connection activity monitoring method according to claim 1, wherein when a failure is detected in any of the connections, a monitoring signal is transmitted for all the connections established with the failed computer. A connection activity monitoring method characterized by detecting a connection failure.