CN1315288C - Uninterrupted Switching Method of Dual Network Communication System - Google Patents

Abstract

The invention provides a communication method of a dual communication network that realizes network switching without interrupting network communication when the network fails. This network communication method classifies network information according to real-time information and non-real-time information. The network information has a time stamp. The polling information is carried out from the current operating network, and the actively uploaded information is simultaneously sent from the two operating networks. Its advantage is that when the network is normal, it only sends polling commands for the running network segment, reducing network load; when the network fails, it only switches for the faulty node, saving time; No data is lost during network switching.

Description

Uninterrupted Switching Method of Dual Network Communication System

technical field

The invention relates to the network communication technology of electric power system automation, provides the realization method of the double network communication system based on Ethernet, especially realizes uninterrupted real-time communication.

Background technique

The dual network communication system is based on the Ethernet communication technology and is applied to the automation communication of the power system based on the Ethernet. Uninterrupted real-time communication is realized by utilizing the characteristics of high reliability, fast speed, and large data volume of Ethernet communication and the mechanism of actively sending information. The implementation method is used for Ethernet-based dual network redundancy backup, which can be used for both substation layer equipment and bay layer embedded equipment, so as to realize network hot backup and smooth switching.

At present, the common implementation methods of network switching mainly include:

1. Divide the communication network into two network segments, the main network and the backup network, one of which is the main network and the other is the backup network. When the network is normal, network information is only transmitted through the current running network segment. When the main network fails, the network management program will switch the entire network to another network segment, and the application program will also start to receive the network information of this network segment. The advantage of this kind of network switching is that the principle is simple and the implementation is simple; the disadvantage is that the overall network switching takes a lot of time, and network information may be lost during the period from network failure to switching.

2. Divide the communication network into two network segments, the main network and the backup network. One of the network segments is the main network and the other is the backup network. The management program determines which network segment the node uses according to the network status of each network node. . When the network is normal, the network information of each node is only transmitted through the current running network segment of the node. When the network connection of a certain network node fails, the network management program switches the network node to another network segment, and the application program also Start receiving the network information of the network segment of the network node. The advantage of this kind of network switching is that the switching time is saved only for the faulty node, and the implementation is relatively simple; the disadvantage is that the network information may still be lost during the time period from the network failure to the switching.

3. Divide the communication network into two network segments, the main network and the backup network. One of the network segments is the main network and the other is the backup network. The management program determines which network segment the node uses according to the network status of each network node. . When the network is normal, the network information is transmitted through the two network segments of each node at the same time. After the network information reaches the receiving end, the receiving end judges that the network information is valid when both network segments receive the same network information; when a network When the network connection of a node fails, the network management program switches the network node to another network segment, and at the same time notifies the application program that it only needs to receive the network information of one network segment. The advantage of this kind of network switching is that it only saves switching time for the faulty node and has high information reliability; however, it is difficult to ensure that network information will not be lost during the period from network failure to switching.

Contents of the invention

In order to overcome the above-mentioned problems in the existing network switching technology, the present invention provides an implementation method of an Ethernet-based dual-network communication system that realizes network switching without interrupting network communication when a network fails.

The principle of this implementation method is: divide the communication network into two network segments, the main network and the backup network, one of which is the main network A, and the other is the backup network B, and the management program is based on the network status of each network node Decide which network segment this node uses. According to real-time and non-real-time requirements, the information is divided into polling information and active uploading information. Polling information is mainly for timing query information with little change or no change, and active uploading information is mainly for changes that need to be updated immediately. Displacement information or over-limit information; when the network is normal, the polling information is only transmitted through the current running network segment A of the network node, and the actively uploaded information is transmitted through the two network segments A and B of each node at the same time. The network information received by the network segment is judged as valid information. After the network information arrives at the receiving end, the polling information is processed directly, and the information sent actively is determined according to the time stamp in the information whether the information is the latest information. If it is the latest information, it will be updated immediately, and if it is not the latest information, the information will be discarded; When the network connection of a network node fails, the network management program switches the network node to another network segment B, and the polling information is transmitted through the switched network segment B, and the information is actively sent because there are two network segments A Both send it to B, so it will not be lost. After the switchover, disconnect the connection with the faulty network segment of the network node, and send the active information only from network segment B; after a delay, reconnect the faulty network segment of the network node, and if the fault is eliminated, connect success. After the reconnection is successful, the proactively sent information is resent from the two network segments A and B at the same time.

The advantage of this kind of network switching is that switching time is saved only for faulty nodes, and information reliability is high; at the same time, data will not be lost during network switching.

The basic features of this network switching method are:

1. The information is classified according to the polling information and the active sending information, and the active sending information is transmitted on two network segments at the same time.

2. The network information has a time stamp, and it is judged whether the network information has been updated according to the time stamp. If updated, discard the message; otherwise, update the message.

3. The network switching takes the network node as the unit, and only performs network switching for the faulty node.

4. In the process of network switching, since the active uploading information of the transmission change is sent from two network segments at the same time, the information will not be lost.

Description of drawings

Figure 1a and Figure 1b show a schematic diagram of the switching process of the network from normal to faulty.

Fig. 2 shows a flowchart of the sending of the polling command.

Figure 3 shows a flowchart of the reply to the polling command.

Fig. 4 shows a flowchart of the transfer of unsolicited information.

Fig. 5 shows a flowchart of the reception of polling information.

FIG. 6 shows a flow chart of receiving unsolicited information.

Fig. 7 shows the flowchart of network switching.

Figure 8 shows the flowchart of network recovery.

Detailed ways

The realization method of the dual network communication system based on Ethernet includes the network information transmission mode, the diagnosis of the network failure state, the network switching mode, the network restoration and so on.

1) Network information transmission method

There are two ways to transmit network information: polling and active uploading. Polling information is generally carried out at regular intervals and does not have real-time characteristics; actively sending information generally only sends changed data. In the present invention, both information transmission methods are used, and the polling method is used to periodically call for data. Since the amount of data is relatively large, it is only carried out through the current operating network segment. Actively uploading data is used to transmit changed data, and this part of data is carried out simultaneously on two network segments, so that when one of the network segments fails, the information can be transmitted continuously through the other network segment.

2) Diagnosis of network fault status

Network fault detection is a prerequisite for network switching, so it is very important to quickly and accurately judge network faults. In order to ensure that the network switching method is universal, the present invention uses a combination of network protocol and network management to judge the fault state of the network. Method 1, the standard network protocol will judge the running status of the network, and actively notify the application program, and the application program uses the returned information to judge the network failure; method 2, the network management program judges whether the current running network segment is available by sending polling commands successfully , if sending the polling command is unsuccessful, the cumulative number of sending failures, when the number of sending failures reaches the limit condition, it is determined that the network segment is faulty. The two ways of judging network faults have their own disadvantages. Method 1 is related to the network protocol. When a physical failure occurs, it takes longer to judge the network failure than method 2. Method 2 has nothing to do with the network protocol. When a non-physical failure occurs, the network The failure time is longer than the method 1. Therefore, in the present invention, the two methods are combined, and if any method determines a network failure, it is determined that a network failure has occurred.

3) Network switching method

With the guarantee of the network information transmission mode, the network switching mode can be very flexible. The method adopted by the present invention is to immediately switch the operating network segment of the network node to the standby network segment after detecting a certain network segment failure of a certain network node, disconnect the faulty network segment of the network node, and release network connection resources and Maintenance information for this connection.

4) Network recovery

An important function of the network management program is to automatically restore the network. According to the recovery time definition of the configuration file, the network management program will regularly try to connect to the faulty network segment. Since this part of the program uses an independent thread, it will not block the operation of other programs . When the network management program detects that the connection is successful, it automatically maintains the information of the connection and re-performs related initialization work. In addition, even if the faulty network returns to normal, it will not switch back to this network segment immediately, because the two network segments are in an equal relationship, and which network segment is currently running has no impact on the application.

Contains content description and corresponding functions

Figure 1a and Figure 1b show a schematic diagram of the switching process of the network from normal to fault, Figure 1a is when the network is running normally, and Figure 1b is when the network is faulty. As shown in Figure 1a and Figure 1b, the upward arrow indicates the active sending of information, and the downward arrow indicates the query command; when the network is running normally, the A network segment indicates the current operating network segment, and the polling information and active sending information are transmitted, and the B The network segment only transmits the active upload information; when the network fails, the B network segment indicates the current running network segment, and transmits the polling information and the active upload information, and the A network segment is disconnected.

According to the present invention, the communication network is divided into two network segments, the main network and the backup network, wherein one network segment is the main network A, and the other network segment is the backup network B, and the management program determines the node according to the network status of each network node Which network segment to use. According to real-time and non-real-time requirements, the information is divided into polling information and active uploading information. Polling information is mainly for timing query information with little change or no change, and active uploading information is mainly for changes that need to be updated immediately. Displacement information or over-limit information; when the network is normal, the polling information is only transmitted through the current running network segment A of the network node, and the actively uploaded information is transmitted through the two network segments A and B of each node at the same time. The network information received by the network segment is judged as valid information. After the network information arrives at the receiving end, the polling information is processed directly, and the information sent actively is determined according to the time stamp in the information whether the information is the latest information. If it is the latest information, it will be updated immediately, and if it is not the latest information, the information will be discarded; When the network connection of a network node fails, the network management program switches the network node to another network segment B, and the polling information is transmitted through the switched network segment B, and the information is actively sent because there are two network segments A Both send it to B, so it will not be lost. After the switchover, disconnect the connection with the faulty network segment of the network node, and send the active information only from network segment B; after a delay, reconnect the faulty network segment of the network node, and if the fault is eliminated, connect success. After the reconnection is successful, the proactively sent information is resent from the two network segments A and B at the same time.

1. Send polling command

As shown in Figure 2, the polling command is sent regularly, and the polling command is only sent from the current running network segment, and the return status of the sending is used to judge whether the network is normal.

The functions of each step in Figure 2 are explained as follows:

S01-Send a polling command from the current running network segment. Query the current running network segment of the network node, and send a polling command to the network node from the network segment.

S02-judging whether the sending is successful. In this step, it is necessary to judge whether the sending is successful according to the actual situation of the network protocol.

S03-Accumulated times of sending failures. Taking the network node as a unit, the number of sending failures of each network node is respectively accumulated according to the situation of each network node.

S04-clear the number of sending failures. After the polling command is sent successfully, it means that the network is in good condition, so the number of network failures can be cleared.

S05-judging whether the number of failures exceeds the limit. Judging whether the number of failures exceeds the limit is generally determined according to the network situation.

S06—determining that the network is faulty. Once the number of sending failures exceeds the limit, it is determined that the network has failed, but it will not be processed.

2. Response to polling command

As shown in Figure 3, the data is organized according to the polling command, and a reply is made from the network segment that receives the polling command.

The functions of each step in Figure 3 are explained as follows:

S01-judging whether a polling command is received.

S02-judging whether to actively upload. The polling response needs to determine whether there is an active upload at that time. If there is an active upload, it needs to send the active upload information immediately.

S03—sending a polling response to the network that has received the polling command. If no information is actively sent at that time, a polling response can be sent immediately according to the content of the polling command.

3. Actively upload information transmission

As shown in Figure 4, the actively uploaded information only includes data that has shifted or exceeded the limit. Since these information need to be delivered in a timely and reliable manner, they are sent from two network segments at the same time.

The functions of each step in Figure 4 are explained as follows:

S01-judging whether to actively upload.

S02-Send active upload information to two network segments at the same time. In order to ensure that the unsolicited information is not lost, it is necessary to send unsolicited information to two network segments at the same time.

4. Reception of polling information

As shown in FIG. 5 , since polling information has a lower priority than actively sending information, it can be updated directly.

The functions of each step in Figure 5 are explained as follows:

S01-judging whether a polling response is received.

S02—Update polling response information. If a poll reply message is received, the received message can be updated as needed.

5. Actively send information to receive

As shown in Figure 6, there are two copies of the same information actively sent, so it needs to be judged according to the time stamp of the information. If it is not repeated, it will be updated immediately; otherwise, the repeated information will be discarded.

The functions of each step in Figure 6 are explained as follows:

S01-judging whether to receive the actively sent information.

S02-judging whether the information already exists. Since the actively uploaded information is sent from two network segments at the same time, it is necessary to determine whether the information has been updated to avoid repeated data updates.

S03 - update the information sent actively. If the proactively submitted information has not been updated, the proactively submitted information needs to be updated.

6. Network switching

As shown in Figure 7, according to the network failure situation, switch the network and be responsible for clearing network resources.

The functions of each step in Figure 7 are explained as follows:

S01-judging whether the network is faulty. For the method of judging network faults, refer to the description of FIG. 1 .

S02-Switch the current running network to the standby network. Once a network node has a network failure, it is necessary to switch the operating network of the network node to the backup network.

S03-Disconnect the network connection, release network resources and maintain information. After the network is switched, it is necessary to disconnect the failed network and release related network resources and maintenance information.

7. Network recovery

As shown in Figure 8, network recovery is to realize self-repair of temporary network failures without affecting the normal operation of the system.

The functions of each step in Figure 8 are explained as follows:

S01-judging whether the timing time is up. It generally takes a while to reconnect to the network, and the waiting time is determined by the application.

S02-Try to connect to the faulty network segment. If the timer expires, you can try to connect to the faulty network segment.

S03-judging whether the connection is successful.

S04-Maintain network information and perform relevant initialization. If the connection is successful, the information returned by the connection needs to be maintained and used by other related programs.

Claims (7)

1, a kind ofly realizes when network failure that network switches and the communication means based on the dual network communication system of Ethernet of uninterrupted network service, communication network is divided into major network, is equipped with two network segments of net, one of them network segment is a major network, another network segment is to be equipped with net, hypervisor determines this node to use which network segment as major network according to the network condition of each network node, when the network connection of certain network node is broken down, network supervisor switches to another network segment with this network node, it is characterized in that:

Be polling message with the network information according to real time information and non real-time information classification and initiatively go up carry information, and time stamp is set in the network information;

Network just often, polling message only transmits by the current operation network segment of network node;

Initiatively go up two network segments of carry information by each node and transmit simultaneously, the network information that any network segment receives all is judged to be effective information;

After the network information arrived receiving terminal, polling message was directly handled;

The network information initiatively goes up carry information and then decides according to the time stamp in the information whether this information is up-to-date information after arriving receiving terminal, thereby whether decision is upgraded;

Behind the network failure, the fault network node is switched to another network segment;

After the fault of networking, polling message transmits by the network segment after switching.

2, communication means as claimed in claim 1 wherein further comprises the diagnosis of network failure state, the step of network recovery.

3, communication means as claimed in claim 1 is characterized in that:

Polling mode is used for regularly calling data, is only undertaken by the current operation network segment;

The data of sending data to be used to transmit variation in the active, these data are carried out simultaneously at two network segments.

4, communication means as claimed in claim 2 is characterized in that the diagnosis algorithm of network failure state comprises the mode that adopts procotol, wherein:

Procotol is judged network failure, uses operating system mechanism notification application, and utilizes return information to judge network failure.

5, communication means as claimed in claim 2 is characterized in that the diagnosis algorithm of network failure state comprises the mode that adopts network management, wherein:

The network supervisor utilization sends polling order and whether judges successfully whether the current operation network segment is available, if it is unsuccessful to send polling order, accumulative total sends the frequency of failure, reaches restrictive condition when sending the frequency of failure, judges that then fault has appearred in this network segment.

6, communication means as claimed in claim 3 is characterized in that:

Only switch at the fault network node;

After detecting certain network segment fault of certain network node, the operation network segment with this network node switches to the standby network segment immediately;

Disconnect the fault network segment of this network node, and the maintenance information that is connected with this of releasing network connection resource.

7, communication means as claimed in claim 2 is characterized in that the network recovery step comprises:

Define recovery time in the configuration file;

Network supervisor uses independently thread, regularly attempts connecting the fault network segment;

When network supervisor detects successful connection, then safeguard the information of this connection automatically, and the initial work of being correlated with again;

It is normal that fault network recovers.

Images