KR100390403B1 - Method for activating software processor in dual processor - Google Patents

Abstract

The present invention provides a software processor switching method of a redundant processor that determines an active / standby processor through message exchange between redundant processors, and automatically switches the active / standby processor when a failure is detected. The processor switching method of the redundant processor of the present invention is a redundant processor including an active / standby processor configured as an active side and a standby side at initialization, and the processor configured as the standby side in the processor configured as the active side. Transmitting an active message at a predetermined cycle to the active side processor; and transmitting an alternating message from the standby side processor to the active side processor if the active message is not transmitted to the standby side at the preset period. The standby side processor is switched to the active side processor. Therefore, it is possible to prevent the occurrence of an error that operates on the active side or the standby side at the same time by applying the switching method of the active / standby processor of the redundant processor.

Description

Method for activating software processor in dual processor

The present invention relates to a software processor switching method of a redundant processor. In particular, an active / standby processor is determined through message exchange between redundant processors, and a software processor of a redundant processor capable of automatically switching an active / standby processor upon detecting a failure. It is to provide a transfer method.

In general, each processor (consisting of one or more boards) used in an electronic communication system has a redundant structure having two states of active / standby in consideration of reliability.

Redundant processors are two processors with the same function, one active and one in standby, and switching between states when the active processor fails or needs to activate a standby processor. do.

In the system having the structure as described above, the two redundant processors are configured to communicate the states of the self processor and the counter processor in order to obtain the information necessary for the active / standby state transition.

Hereinafter, with reference to the accompanying drawings will be described the initialization of the redundant processor of the general exchange system and the redundant processor of the exchange system.

1 is a diagram illustrating a redundant processor of a general exchange system.

As shown in FIG. 1, the redundant processor of the general switching system is a main processor 1 inside the switching system, and among the roles of switching, vocoder management, call processing, No.7 control, V5.2 control, etc. in the switching system. It is divided into one coprocessor, and if the processor is composed of several boards, it is composed of the main board which plays the main role of the processor, and the sub board (s) which supplies power and communication between the main board and other processors. Can be.

Each processor may be redundant, such as the A side 2a of the first coprocessor and the B side 2b of the first coprocessor, as shown in the first coprocessor 2. Of course, the main processor 1 may also be duplicated like the A side 1a and the B side 1b of the main processor.

Such a redundant processor is divided into two methods as follows. The first can be distinguished by the mounting positions of the processors. That is, the duplicated processors also have different addresses due to the division by the processor address allocated for the communication between the processors constituting the switching system.

Secondly, it distinguishes whether the processor plays a role. In this case, the active side refers to a processor that performs the role of an allocated processor among the redundant processors, and the standby side is a spare processor that performs internal data synchronization and state sensing of the active side to activate the active side. It acts as an active side by the activation process when a side failure occurs.

Such a redundant processor may constitute one or more boards, and the communication between the redundant processors is required for switching between the active / standby processor, which is the redundant control, and the hardware for detecting the mounting / dismounting of the redundant processor is configured. .

2 is a flowchart for explaining initialization of a redundant processor of a conventional switching system.

Initialization of the redundant processor of the conventional switching system is shown in Figure 2, the operation of the processor is generally the initialization of the board (or boards) by the power supply of the board (or boards) constituting the processor mounted in the switching system The process starts with performing (S1). This process is to prepare for interworking with other processors and to initialize the internal CPU or memory of the board (or boards) constituting the processor, and various devices. In addition, the process of checking access with the upper processor allocated to the processor is performed.

Subsequently, if a program constituting the processor is downloaded after the initialization process is executed (S2).

Then, the processor redundancy is determined (S3). In this case, the redundancy of the processor applies an interrupt to the main processor of the motherboard in hardware that detects the dismount / mounting of the processor, and calls the corresponding interrupt service routine from the program of the processor so as to register the status register of the chip related to the redundancy. ) And determine whether it is redundant.

If it is determined that the processor is not redundant, the redundant flag and the standby flag are reset, and the active flag is set (S4).

Then, the process proceeds to the active processor (S8).

However, if the determination result (S3) the processor is redundant is set to the dual flag (S5).

Next, an active / standby processor is set (S6).

If the active processor is set, the wait flag is reset and the active flag is set (S7).

The process then proceeds to an active process (S8).

However, if the standby processor is set, the standby flag is set and the active flag and the other side are reset (S9).

Then, the process proceeds to the standby processor (S10).

In such a conventional redundant processor, a failure condition may occur in which the interrupt generated when the determination of the redundant processor and the determination of the active standby side through the hardware interrupt is not detected due to the CPU load. Both standby processors could act as active side or standby side, and if such problems would occur, the services provided by the switching system would be interrupted.

An object of the present invention has been made in view of the above-mentioned problems of the prior art, and determines an active / standby processor through message exchange between redundant processors, and automatically switches the active / standby processor when a failure occurs during operation. To provide a software processor switching method of a redundant processor that can be.

According to an aspect of the present invention for achieving the above object, in the redundant processor consisting of an active / standby processor set to the active side and the standby (Standby) side during initialization, the processor set to the active side Transmitting an active message to a processor set to a standby side at a predetermined cycle; and if the active message is not transmitted to the standby side at the preset cycle, a transfer message is transmitted from the standby side processor to the active side processor to the active side. The processor is configured to switch to a standby side processor, and the standby side processor is switched to an active side processor.

According to the present invention as described above, it is possible to provide a highly reliable redundant processor since the possibility of error occurrence when switching the active / standby processor of the redundant processor.

1 is a diagram illustrating a redundant processor of a general exchange system.

2 is a flowchart for explaining initialization of a redundant processor of a conventional switching system.

3 is a flowchart illustrating the initialization of the redundant processor of the exchange system according to the present invention.

4 is a flowchart for explaining message handling of a redundant processor of an exchange system according to the present invention.

5 is a flowchart illustrating a redundant message transmission of a redundant processor of a switching system according to the present invention.

6 to 7 are flowcharts for explaining a method of processing a redundant message of the redundant processor of the switching system according to the present invention.

8 is a flowchart illustrating a method of switching by not receiving active messages between redundant processors in a switching system according to the present invention.

9 is a flowchart illustrating a recovery switching method of a first failure state between redundant processors in a switching system according to the present invention.

10 is a flowchart illustrating a recovery switching method of a second failure state between redundant processors in a switching system according to the present invention.

Hereinafter, a configuration and an operation according to an exemplary embodiment of the present invention will be described with reference to the accompanying drawings.

3 is a flowchart illustrating the initialization of the redundant processor of the exchange system according to the present invention.

Initialization of the redundant processor of the switching system according to the present invention, as shown in Figure 3, performs the initialization process of the boards by the power supply of the boards constituting the processor mounted in the switching system (S11). Such an initialization process is a preparation process for interworking with another processor and initializes internal CPUs, memories, and various devices of the boards constituting the processor. In addition, the process of checking access with the upper processor allocated to the processor is performed.

Subsequently, if a program constituting the processor is downloaded after the initialization process is executed, the downloaded program is executed (S12).

Next, the processor is determined for duplication (S13). In this case, the redundancy of the processor applies an interrupt to the main processor of the motherboard in hardware that detects the dismount / mounting of the processor, and calls the corresponding interrupt service routine from the program of the processor so as to register the status register of the chip related to the redundancy. ) And determine whether it is redundant.

If it is determined in step S13 that the processor is not redundant, the redundant flag and the standby flag are reset, and the active flag is set (S14).

Then, the process proceeds to the active processor (S18).

However, if the determination result (S13) the processor is duplexed, the dual flag is set (S15).

Next, the active / standby processor is determined (S16).

Here, if it is determined as the active processor, the wait flag is reset and the active flag is set (S17).

Then, the process by the active process proceeds (S18).

If it is determined as the standby processor, the standby flag is set, and the active flag and other side flags are reset (S19).

Next, the standby processor maintains a standby state (S20).

4 is a flowchart for explaining message handling of the redundant processor of the switching system according to the present invention.

The message handling (S21) of the duplication processor of the switching system according to the present invention shows the processing for receiving a duplication related message of the processor which is the active side / standby side.

Upon receiving the duplication-related message of the processor that is the active side / standby side after process progress by the active processor and the standby processor (S18, S20), it is determined whether the received message is the active message (S22).

If the received message is an active message (S22), the current active message count is accumulated (S23). Then return to the message handling (S21).

However, if it is determined that the received message is not an active message (S22), it is determined whether the received message is a standby message (S24).

If it is determined that the received message is a waiting message, the current waiting message count is accumulated (S25). Then return to message handling (S21).

However, the determination result (S24), if the received message is not a wait message, it is determined whether the (Activate) message (S26).

As a result of the determination (S26), if the received message is a transfer message, a dual flag and a standby flag are set, and the active flag and another side flag are reset (S27).

However, if the received message is not a transfer message (S26), a message other than the active / standby message and the transfer message (for example, a call processing message) is processed, and the message is processed and the message handling continues.

5 is a flowchart illustrating a redundant message transmission of the redundant processor of the switching system according to the present invention.

The duplication message transmission method of the duplication processor of the switching system according to the present invention shows a process of transmitting a duplication-related status message (called a duplication message) to the other processor. Repeat the cycle set by the running processor.

The redundancy message is transmitted when the A side processor or the B side processor is in an active or standby state, and the processor operating as the active side of the A side and B side processors is in a state in which it is operating as the active side (that is, the active flag is set). ) Is determined in units of set cycles (S28), and if the active flag is set (S28), an active message is transmitted to the other standby processor (S29) and returned.

However, if it is not operating on the active side, it is a state of operating on the standby side (that is, when the standby flag is set), so it is determined whether to set another flag that determines the case where the active side is determined (S30). ).

As a result of the determination (S30), if another flag set state, that is, the active side, is determined, it is the standby side. The standby processor then determines whether the standby message transmission count is equal to or greater than the set threshold (S31).

As a result of the determination (S31), if the wait message transmission count is equal to or greater than the threshold value, it returns.

However, if it is determined in step S31 that the waiting message transmission count is not yet at the limit, the waiting message is transmitted to the active side (S32) and returned.

The reason for transmitting the standby message as described above (S32) or returning it without transmitting is to reduce the load caused by exchanging a message between the active side and the standby side processor when processing a message between duplicate processors. It is for redundancy control.

6 to 7 are flowcharts for describing a method of processing a duplication message (duplicated state related message) of a duplication processor of a switching system according to the present invention.

The duplication message processing method of the duplication processor of the switching system according to the present invention is performed at twice the period of the duplication message transmission method described in FIG. In this case, the reason for performing the duplication message processing in more than twice the period is to prepare for the possibility of the loss of the message related to the duplication state between the duplicated processors, and to prevent the service degradation of the processor due to the active side / standby side transfer whenever there is a loss. It is to execute (Activate).

6 illustrates a method of processing a message related to the active side duplication state, and FIG. 7 illustrates a method of processing a message related to the standby side duplication state. First, the method of processing a message related to the active side duplication state checks whether there is a duplication message (S33).

Next, it is determined whether the active flag is set (S34).

If it is not the active flag set (S34), it is determined whether the previous active message count is equal to the current active message count (S43) (see Fig. 7). That is, a process of determining whether an active processor has failed is performed by comparing the number of currently received active messages / queue messages with the number of received message frames in a previous cycle to check the duplicated message.

However, if it is determined in operation S34 that the active flag is set, it is determined whether the current waiting message count is greater than the minimum number of received waiting messages in order to determine whether the standby side is operating normally (S35).

If it is determined that the current waiting message count is not greater than the minimum number of received waiting messages (if small), it is determined whether the current active message count is greater than zero (S38).

However, if it is determined in operation S35 that the current waiting message count is greater than the minimum number of waiting messages received, another side flag indicating that the current waiting message count is equal to the minimum number of waiting messages received (S36) and indicating that the current status is a redundant state. Set is performed (S37).

Next, it is determined whether the current active message count is greater than zero (S38).

As a result of determination (S38), if the current active message count is not greater than 0, the previous active message count / previous storing the number of currently received active messages and waiting messages for the next received duplicated message check and the number of received message frames in a previous cycle. It stores in a waiting message count (S50) (refer FIG. 7).

As a result of the determination (S38), if the active message count is greater than zero, the active message is received by the processor operating as the active side. Therefore, whether the A side flag of the processor is set and both the A side and the B side processors are the active side. This is the case. Then, it is determined whether the number of currently received active message counts is greater than the limit of the number of received active messages set by the A side processor (Limit of A-Side Act_Msg Count: LAAC) (S39).

As a result of determination (S39), if the number of active message counts currently received is greater than the limit of the number of received active messages set by the A side processor, the B side processor transmits a transfer message to maintain the standby side (S40). That is, the A side processor operates as the active side.

However, if it is determined in operation S39 that the number of active message counts currently received is not greater than the limit of the number of received active messages set by the A side processor, whether the B side flag of the processor is set and the number of currently received active message counts is the B side. It is determined whether the number of received active messages is greater than the limit set by the processor (Limit of B-Side Act_Msg Count: LABC) (S41).

As a result of determination (S41), if the number of active message counts currently received is greater than the limit of the number of received active messages set by the B side processor, the B side processor transmits a transfer message to the A side processor so that the A side processor maintains the standby side. (S42). That is, the B side processor is operated as the active side.

At this time, in order to give priority to one of the redundant processors, LAAC is set to have a smaller value than LABC when both the A side and B side processors operate as the active side. Therefore, it increases the probability of the A side acting as the active side. Of course, you can also do the opposite.

However, if it is determined in step S41 that the number of active message counts currently received is not greater than the limit of the number of received active messages set by the B side processor, the number of currently received active messages and waiting messages is transferred to check the next received duplicated message. The number of message frames received in a cycle is stored in a previous active message count / previous wait message count (S50) (see FIG. 7).

When the above-mentioned active flag is reset, that is, when the standby processor among the redundant processors (S34) determines that the previous active message count is not equal to the current active message count (S43), the current active message count is the upper limit of the current active message count value. If it is determined that the value is larger (S44), and the determination result (S44), if the current active message count is not greater than the upper limit of the current active message count value, the period of the previous received active message and the waiting message is checked to check the next received duplicated message. In operation S50, the received message frame count is stored in a previous active message count / previous wait message count.

However, if the current active message count is equal to the current active message count value upper limit, the determination result (S44) sets the current active message count to the lower limit of the current active message count value, sets the active wait count to 0, and sets another side flag (S46). In order to check the received duplicated message, the current received active message and the number of waiting messages are stored in a previous active message count / previous waiting message count storing the number of received message frames in a previous cycle (S50).

However, if the active message count before the determination result S43 is equal to the current active message, the active standby message count is accumulated (S45).

Subsequently, it is determined whether the active message is not received during the time multiplied by the limit of Act_Msg Same Count (LASMC), which determines whether the active wait message count has not been received (S47).

As a result of determination (S47), if a failure occurs in the active side, the standby side transfers to the active side by not receiving the active message (S48).

However, if it is determined in operation S47 that the active wait message count is not greater than the limit cycle number LASMC, it is determined whether the current wait message count is greater than zero by counting current wait messages of the A side and B side processors (S49).

As a result of the determination (S49), if the current waiting message count is not greater than zero, the previous active message count / storage storing the number of currently received active messages and the number of waiting message frames in the previous cycle to check the next received duplicated message. Stored in the previous wait message count (S50).

However, if the current standby message count is greater than 0, the determination result (S49) corresponds to a case where both the A side and the B side processors are the standby side because the standby message is received by the processor operating as the standby side. Then, it is determined whether the A side flag of the processor is set and whether the number of waiting message counts currently received is greater than the limit of the number of waiting messages set by the A side processor (Limit of A-Side Std_Msg Count: LASC) (S51). ).

As a result of the determination (S51), if the number of currently received waiting messages is greater than the limit of the number of waiting messages set by the A side processor, the transfer message is transferred to the B side processor so that the B side processor maintains the standby side (S52). That is, the A side processor operates as the active side.

However, if it is determined in operation S51 that the number of received wait message counts is not greater than the limit of the number of received wait messages set by the A side processor, whether the processor B side flag is set and the number of currently received wait message counts is B side. It is determined whether the number of messages waiting to be received is greater than the limit set by the processor (Limit of B-Side Std_Msg Count: LBSC) (S53).

As a result of determination (S53), if the number of currently received waiting messages is greater than the limit of the number of waiting messages set by the B side processor, the B side processor transmits a transfer message to the A side processor so that the A side processor maintains the standby side. (S54). That is, the B side processor is operated as the active side.

At this time, in order to give priority to one of the redundant processors, the LASC is set to have a value smaller than that of the LBSC when both the A side and the B side processors operate as the standby side. Therefore, it increases the probability of the A side acting as the active side. Of course, you can also do the opposite.

8 is a flowchart illustrating a switching method by not receiving active messages between redundant processors in a switching system according to the present invention.

Here, FIG. 8 performs a switchover from the standby side (B side) to the active side (A side) through a procedure (S43 to S47) for monitoring the operation of the active side of the redundant message processing method shown in FIG. 7.

That is, as shown in Fig. 8, the active side (A side) transmits the active message (Act-Msg) at the message sending interval set to the standby side (B side). At this time, the standby side (B side) receives the active message is transmitted, but if no active message is transmitted during the set Active Side Fault Detect Interval, the standby side (B side) goes to the active side (A side). The transfer message (Activate Message) is transmitted to switch the active side (A side) to the standby side (A side), and the standby side (B) is transferred to the active side (B). After such a transfer, the active message Act-Msg is transmitted at the message sending interval set from the active side (B side) to the standby side (A side).

9 is a flowchart illustrating a recovery switching method of a first failure state between redundant processors of an exchange system according to the present invention.

9 illustrates whether the number of active message counts currently received is greater than the limit of the number of received active messages set by the A side processor when both the A side processor and the B side processor are the active side in the redundant message processing method illustrated in FIG. 6. (Smit Limit A-Side Act_Msg Count: LAAC) and transmitting the transfer message (S40), and whether the number of active message counts currently received is greater than the limit of the number of received active messages set by the B side processor. In step S41, a limit of B-Side Act_Msg Count (LBAC) and a transfer message (S42) are applied. As shown in FIG. 9, an active message is transmitted to each other at a message sending interval from an A-side processor and a B-side processor to a counterpart side processor. If it continues during the Active_Side Decision Interval, the A-Side processor sends an Activation message to the B-Side processor so that the A-Side processor remains active. The B-side processor keeps the standby side.

10 is a flowchart illustrating a recovery switching method of a second failure state between redundant processors in an exchange system according to the present invention.

FIG. 10 illustrates whether the number of waiting message counts currently received is greater than the limit of the number of waiting messages set by the A side processor when both the A side processor and the B side processor are standby sides in the redundant message processing method illustrated in FIG. 7. Limit of A-Side Std_Msg Count (LASC) step (S51), sending a transfer message (S52), and the number of currently received waiting message count is greater than the limit of the number of waiting messages set by the B side processor Step S53 of determining whether a limit of B-Side Std_Msg Count (LBSC) is transmitted and step S54 of transmitting a transfer message are applied. As shown in FIG. 10, the A-side processor and the B-side processor transmit the waiting messages to each other at a message sending interval, and set the active side in advance. If it continues during the Active_Side Decision Interval, the A-Side processor sends an Activation message to the B-Side processor so that the A-Side processor remains active. The B-side processor keeps the standby side.

As described above, the present invention performs active / standby switching of the redundant processor through message exchange between the A side processor and the B side processor and data value change in the processor, so that both sides operate as an active side or a standby side. It is easy to detect the occurrence of the active / standby side switching, thereby improving the processor reliability of the switching system.

Claims (3)

A redundant processor consisting of an active / standby processor configured as an active side and a standby side at initialization, Transmitting an active message at a predetermined period from a processor set as the active side to a processor set as the standby side; If the active message is not transmitted to the standby side processor at the predetermined period, the standby side processor transmits a transfer message to the active side processor to switch the active side processor to the standby side processor, and the standby side processor is an active side. A software processor switching method of a redundant processor comprising the step of switching to a processor. The processor of claim 1, wherein the processor configured as the standby side monitors transmission of an active message at a predetermined period from the processor configured as the active side, and determines whether the processor configured as the active side is faulty, And transferring the configured processor to a standby side processor and transferring the standby side processor to an active side processor. The processor of claim 1, wherein the processor configured as the active side and the processor set as the standby side transmit or receive active messages to each other, or if a failure state transmits and receives standby messages to each other, gives priority to the processor mounting location and transmits a transfer message. The software processor switching method of the redundant processor, characterized in that for recovering the failure state.