KR101087886B1 - Device and method for controlling tcp connection of server - Google Patents

Abstract

The present invention relates to an apparatus and method for controlling a TCP connection of a server capable of quickly responding to an abnormal situation by monitoring an Ethernet link state and a TCP connection state in a state in which a TCP communication server makes a TCP connection from a client. The present invention provides a MAC controller for controlling the transmission and reception of TCP communication data through an Ethernet network connected to a client, and generates a client socket upon connection request of the client, permits the connection, and counts the connection time with the client. And, if there is no request for data from the client while the counted connection time exceeds the preset waiting time, the control unit terminates the connection after destroying the client socket.

Server, client, link, error, connection time

Description

DEVICE AND METHOD FOR CONTROLLING TCP CONNECTION OF SERVER}

The present invention relates to TCP connection control in an Ethernet communication network. In particular, the server of the server that can communicate quickly when an abnormal situation is achieved by monitoring the Ethernet link state and the TCP connection state in a state where the TCP communication server is a TCP connection from the client. A connection control apparatus and a method thereof are provided.

In an Ethernet communication network in a star topology structure, in which a plurality of network devices are connected to a hub, an intelligent electronic device (IED) becomes a server and the monitoring panel (PC) It is a client and sends and receives data through TCP communication through a predetermined communication network.

TCP communication is a reliable communication method, which is widely used in industrial protocols, and has a characteristic of data communication after establishing a connection between a server and a client.

In order to communicate between a client (monitor board) and multiple servers (IED devices), a cable that basically connects a network relay device such as a hub and a device is required.

Even if the cable is disconnected or the hub is broken and the client's connection is abnormally terminated, the server must take countermeasures for reliable communication.

As shown in the network system of FIG. 1, the server 10 is connected to the client 30 with an Ethernet cable through a hub 20.

In the detailed configuration of the server of FIG. 2, the server 10 interprets a reception frame transmitted from the client 30 and generates a transmission frame, and a MAC controller 13 that manages data transmission and reception of the Ethernet layer. And, it is composed of a link connection portion 15 that is responsible for the physical connection of the Ethernet layer.

As shown in FIG. 1, the server is connected to the client through an hub via an Ethernet cable. When the client requests a TCP connection, the server creates a TCP socket and transmits a frame indicating that the connection is established.

After that, the client requests data from the server, and the server constructs response data and sends it to the client at regular intervals, and the client communicates with the server.

The frame received by the server through the Ethernet link of FIG. 2 is input to the control unit 11 via the link connection unit 15 and the MAC controller 13, and the received frame is interpreted by the control unit 11. Subsequently, the controller 11 generates a response frame and transmits the response frame to the client 30 through the MAC controller 13 and the link connection unit 15.

If the data request command from the client is not sent to the server according to the cable disconnection, the client terminates the TCP connection and sends a reconnection request to the new port number. However, because the server is disconnecting the disconnect request from the client, the server does not receive it and remains connected.

In this state, even if the client reconnects to the new port number when the cable is reconnected, the server recognizes that the connection has already been made.

In order to solve this phenomenon, the following two methods are used.

First, the client writes some fixed data at a predetermined period, for example, to the server's promised memory address, and the server's program monitors the fixed data written to the promised memory address, for example, at least 60 seconds. If there is no data change, forcibly terminates the server port and makes the connection wait. This method has a problem in that unnecessary traffic is generated by exchanging separate data frames.

The second method is to use the keep-alive function, which is a TCP socket option. Setting the keep-alive parameter on the server to check if the client is alive will periodically send a check frame to the client.

If there is no response to the keep-alive packet, the problem is considered to be a problem and the TCP socket is closed to wait for the next connection.

However, this keep-alive function is not provided in all TCP / IP products, and increases the price of equipment. Also, setting the keep-alive parameter value is a technically difficult part, and is very dependent on the system, resulting in inefficiency. In addition, since keep-alive packets are periodically transmitted, unnecessary traffic is generated, which causes a problem that traffic efficiency is lowered in a normal state.

The present invention monitors the Ethernet link status in a state where a TCP communication server is connected to a TCP connection from a client, and monitors the TCP connection state using a connection time counter so that a server can quickly respond to an abnormal situation. An apparatus and a method thereof are provided.

Technical means of the present invention for achieving the above object, MAC controller for controlling the transmission and reception of TCP communication data via an Ethernet network connected to the client; And allowing a connection by creating a client socket when the client requests a connection, counting a connection time with the client, and requesting no data from the client while the counted connection time exceeds a preset waiting time. In this case, the control unit for terminating the connection after the client socket is destroyed.

Specifically, the controller is characterized in that the count after initializing the access time when the client requests the connection.

According to another aspect of the present invention, there is provided a MAC controller that detects an abnormality of an Ethernet network connected to a client and generates a link error interrupt; And creating a client socket when the client requests a connection and allowing the connection, and setting a link error flag when a link error interrupt is transmitted from the MAC controller, and destroying the client socket when the link error flag is set. And a controller for terminating the connection with the client afterwards.

Another technical means of the present invention for achieving the above object is a MAC controller for controlling the transmission and reception of TCP communication data through an Ethernet network connected to the client, and detects it when the Ethernet network is abnormal and generates a link error interrupt. ; And a communication control unit which allows a connection by creating a client socket when the client requests a connection, and connects when there is no request for data while the connection time is counted when the connection with the client exceeds a preset waiting time. And a control unit including an access time counter for terminating the connection and an interrupt processing unit for terminating a connection with a client after setting a link error flag when a link error interrupt is transmitted from the MAC controller. .

The controller may be configured to destroy the generated client socket when the connection with the client is terminated. The server may be an intelligent electronic device, and the client may be a monitoring panel.

Technical method of the present invention for achieving the above object, the server comprising the steps of confirming the connection request of the client through the created server socket; Allowing a connection by creating a client socket when a connection request is made from the client; Initializing and counting an access time after allowing the access; And terminating the connection with the client and destroying the client socket when there is no data request from the client while the counted connection time exceeds the preset waiting time.

Another technical method of the present invention for achieving the above object, the server comprises the steps of confirming the connection request of the client through the created server socket; Allowing a connection by creating a client socket when a connection request is made from the client; After allowing the connection, the server checks whether the Ethernet network is abnormal and generates a link error interrupt in abnormality; And if the link error interrupt occurs, terminating the connection with the client after setting the link error flag and destroying the client socket.

As described above, the present invention provides a link error detection and a connection timer function of a server, such as disconnection of an Ethernet cable between a hub and a server, an abnormality of a hub itself, an disconnection of an Ethernet cable between a hub monitor, and an abnormal state of a client (monitor panel) itself. There is an advantage that can quickly deal with abnormal phenomena such as.

In addition, by detecting a link error, it is possible to quickly determine and display the abnormal state of the connection cable between the hub and the server (IED device), and because it uses an access timer, unnecessary traffic is not exchanged between the server and the client. This does not happen, and there is no need for additional setup or programming on the client, which is easy to install and maintain.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

3 is a block diagram illustrating a network system including a server and a client according to the present invention, and includes a server (100), a hub (200), and a client (300).

The server 100 is connected to the client 300 through the Ethernet network and the hub 200. When the client 300 makes a TCP connection request, the server 100 creates a TCP socket and generates a frame indicating that the connection is established. It is composed of an intelligent electronic device (IED) for transmitting to the client (300).

The server 100 basically interprets the reception frame transmitted from the client 300 and creates a transmission frame, the MAC controller 120 in charge of data transmission and reception of the Ethernet layer, and the physical layer of the Ethernet layer. It is configured to include a link connector 130 that is in charge of the connection.

In the case of the controller 110, the client 300 creates a client socket to allow a connection and counts a connection time with the client 300, and the counted connection time exceeds a preset waiting time. When there is no request for data from the client 300, the client socket is configured to terminate the connection after destroying the client socket.

In addition, the controller 110 allows a connection by creating a client socket when the client 300 requests a connection, and sets a link error flag when a link error interrupt is transmitted from the MAC controller 120. If the flag is set, it is configured to terminate the connection with the client 300 after destroying the client socket.

In addition, the MAC controller 120 is configured to control transmission and reception of TCP communication data through an Ethernet network connected to the client 300 and to detect a link error when the Ethernet network is abnormal and generate a link error interrupt.

The control unit 110 is configured to include a communication control unit 111, the connection time counter 113 and the interrupt processing unit 115, as shown in FIG. 3, the communication control unit 111 of the client 300 It is configured to allow a connection by creating a client socket upon a connection request, and to generate a transmission frame by analyzing a reception frame transmitted from the client 300. The connection time counter 113 is connected to the client 300. If there is no request for data while the access time is counted and the counted access time exceeds the preset waiting time, the connection is terminated, and the interrupt processor 115 transmits a link error interrupt from the MAC controller 120. After setting the link error flag, the connection with the client 300 is terminated.

That is, the frame transmitted to the server 100 through the Ethernet network is input to the control unit 110 via the link connection unit 130 and the MAC controller 120 of the server 100 and then the communication control unit ( 111) the received frame is interpreted. Subsequently, the controller 110 generates a response frame and transmits the response frame to the client 300 through the MAC controller 120 and the link connection unit 130.

The client 300 is configured as a monitoring panel such as a PC that communicates with the server 100 while repeating an operation of requesting data from the server 100 and receiving response data from the server 100 at regular intervals.

On the other hand, the server 100 is connected to the control unit 110 and the fixed period timer 140 connected to the bus interface, the memory 150 connected to the control unit 110 and the bus interface, and the control unit 110 Key input unit 160 for changing the input idle time displayed on the LCD display unit 171 and an LCD display unit 171 connected to the control unit 110 to display a connection termination timer 140 value. ) And a display unit 170 each having an LED display unit 175 that is turned on according to a control signal of the controller 110 at the time of a link error Fail.

In addition, the MAC controller 120 is connected to the control unit 110 and the link connecting unit 130 and DMA and the Ethernet interface, respectively, the link connecting unit 130 is connected to the connector 180, the connector 180 is It is connected to an external Ethernet network.

4 is a diagram illustrating a data flow between the controller 110 and the MAC controller 120 according to the present invention. The controller 110 and the MAC controller 120 exchange TCP communication data when performing normal communication. When the link state is not in the normal state, the MAC controller 120 generates a link error (Fail) interrupt to the controller 110 so that the controller 110 can recognize the link state.

In the Ethernet communication network of the star topology structure connected to the hub 200 as shown in FIG. 3, the IED device becomes the server 100 and the monitoring panel becomes the client 300.

In order for the server 100 (IED device) to communicate with the client 300 (monitoring panel), an IP address, a TCP port number, and a timeout (wait time) setting, which are necessary parameters for TCP communication, are required, and the control unit of the server 100 is provided. 110 receives such a parameter through the key input unit 160 and stores it in the memory 150.

When a TCP connection request is received from the client 300 at the server 100, TCP communication with the client 300 is performed.

On the other hand, the control unit 110 of the server 100 operates the connection time counter 113 in a state of TCP connection in accordance with the connection monitoring routine, the link state monitoring is also performed.

The data request command of the client 300 is input to the controller 110 via the link connection unit 130 and the MAC controller 120 through the Ethernet network.

When the controller 110 interprets the request command, prepares and outputs a response frame, the controller 110 is eventually transmitted to the client 300 through the MAC controller 120, the link connection unit 130, and the Ethernet network.

If a link error (Fail) occurs due to a disconnection between the hub 200 and the server 100, such as an Ethernet network, the MAC controller 120 immediately detects this and delivers a link error interrupt to the controller 110. The controller 110 terminates the TCP connection according to the link error interrupt, and controls the LED display unit 175 to light up the LED to indicate that a link error has occurred.

When the Ethernet network is replaced and returned to the normal state, the controller 110 turns off the LED of the LED display unit 175 again to inform the user that the state has been returned to the normal state.

The determination of the link error between the hub 200 and the client 300 counts the connection time when the client 300 is connected, and when the counted connection time exceeds the set waiting time, the controller 110 connects to the TCP connection. Terminate and wait for connection. Here, the set waiting time may be set to a value between 10 seconds and 60 seconds. Of course, even if the client 300 is not reset because the normal connection is not terminated or the client 300 does not communicate for a long time due to a software error after maintaining the connection, the connection time exceeds the preset waiting time. The server 100 terminates the connection and waits for the connection.

If the client 300 requests a new connection after returning to the normal state, the server 100 reconnects in the standby mode and TCP communication is resumed.

Therefore, in the present invention, the connection time count and link error interrupt information are used to monitor the TCP connection. When the increased value through the connection time counter 113 after the connection of the client 300 is greater than a preset waiting time, It is possible to return to the connection standby state more quickly than the existing method by combining the method of terminating the TCP connection with the network and the method of terminating the connection by generating the link error interrupt when an abnormality occurs by monitoring the status of the Ethernet network. .

5 is a flowchart illustrating a connection process of a server according to the present invention, which will be described below with reference to FIGS. 3 and 4.

First, the server 100 executes the TCP server 100 program at boot time, generates one socket, and registers itself as a server socket (S1).

Subsequently, the server 100 checks whether a connection is requested from the client 300 through the server socket (S2).

If the client 300 is requested to connect, the client 300 permits the connection request, and creates a client socket for communication with the client 300 (S3 and S4).

Subsequently, the server 100 increases the access time count after initializing the count value to '0' before increasing the access time count of the client 300 (S5).

Subsequently, the server 100 checks whether there is a request for data from the client 300 through the client socket (S6), and if there is a request for data, transmits the corresponding data to the client 300 (S7). The server 100 waits for re-request of data after initializing the access time count (S6).

However, when the request for data is not received, the server 100 determines whether the access time counted through the access time counter 113 exceeds a preset waiting time (S8), and the access time exceeds the waiting time. If the connection is terminated, the client socket is terminated after the connection is terminated to return to the connection waiting state (S11).

In addition, if a link error interrupt signal is transmitted from the MAC controller 120 while the server 100 is in operation, the interrupt processor 115 sets a link fail flag, and the server 100 sets the link error flag. If the error flag is checked and set (S9), after the connection is terminated, the client socket is destroyed to return to the connection waiting state (S11). For example, the MAC controller 120 transmits a link error interrupt signal to the controller 110 when a link state abnormality is detected due to disconnection of the Ethernet network, and the interrupt processor 115 of the controller 110 sends a link error flag. Set to the set value ('1' or '0').

When the TCP connection termination request is received from the client 300 during operation (S10), the server 100 terminates the TCP connection and then terminates the client socket and returns to the connection waiting state (S11).

Therefore, in the present invention, an IED device, which is a server 100 for TCP communication, monitors a line state in a state in which a TCP connection is made from a client 300, or monitors a TCP connection state using an access time counter 113, and thus is abnormal. If it is determined that the situation is terminated quickly, the connection can be terminated and the standby state can be waited for another TCP connection.

Accordingly, the present invention provides a link error detection and a connection timer 140 function of the server 100, an Ethernet network disconnection between the hub 200 and the server 100, an abnormality of the hub 200 itself, and a hub 200 between the monitoring panel. It is possible to cope better with the conventional technology against abnormal phenomena such as disconnection of Ethernet network and abnormal state of the client 300 (monitor panel) itself.

That is, by detecting a link error, it is possible to quickly determine and display an abnormal state of the connection network between the hub 200 and the server 100 (IED device), and use the watchdog timer 140 so that the server 100 and the client Since no separate check frames are exchanged between the 300, unnecessary traffic does not occur, and no additional setting or programming is required for the client 300, thereby facilitating installation and maintenance.

Preferred embodiments of the present invention are disclosed for purposes of illustration, and those skilled in the art will be able to make various modifications, changes, and additions within the spirit of the present invention. Therefore, such modifications, changes and additions should be determined not only by the claims below, but also by equivalents to those claims.

1 is a network system including a server and a client according to the prior art.

2 is a view showing a detailed configuration of a server according to the prior art.

3 is a block diagram showing a network system including a server and a client according to the present invention.

4 is a diagram illustrating a data flow between a controller and a MAC controller according to the present invention.

5 is a flowchart showing a server connection process according to the present invention.

* Explanation of symbols for the main parts of the drawings

100: server 110: control unit

111: communication control unit 113: connection time counter

115: interrupt processor 120: MAC controller

130: link connection 140: timer

150: memory 160: key input unit

170: display unit 180: connector

200: hub 300: client

Claims (8)

delete delete A MAC controller which detects an abnormality of the Ethernet network connected to the client and generates a link error interrupt; And When the client requests a connection, the client socket is created to allow the connection, and when the link error interrupt is transmitted from the MAC controller, the link error flag is set; when the link error flag is set, the client socket is destroyed. And a controller for terminating the connection with the client. A MAC controller that controls transmission and reception of TCP communication data through an Ethernet network connected to a client, and detects when an abnormality of the Ethernet network occurs, thereby generating a link error interrupt; And A communication control unit which allows a connection by creating a client socket when the client requests a connection; and when there is no request for data while the connection time is counted and the counted connection time exceeds a preset waiting time. A control unit including a connection time counter for terminating and an interrupt processing unit for terminating a connection with a client after setting a link error flag when a link error interrupt is transmitted from the MAC controller; . The method of claim 4, wherein The control unit TCP connection control device of the server, characterized in that for destroying the generated client socket when the connection is terminated with the client. The method according to any one of claims 3 to 5, And the server is an intelligent electronic device and the client is a monitoring panel. delete The server confirms the connection request of the client through the created server socket; Allowing a connection by creating a client socket when a connection request is made from the client; After allowing the connection, the server checks whether the Ethernet network is abnormal and generates a link error interrupt in abnormality; And When the link error interrupt occurs, terminating a connection with a client after setting a link error flag, and destroying a client socket.

Images