KR20040024068A - Method for controlling cross duplication of duplication processor - Google Patents

Abstract

PURPOSE: A method for controlling cross duplication of duplicated processors is provided to sense that an interworking cable between duplicated processors is abnormal or dismounted, and to quickly switch by searching a communicable path, thereby continuing a communication service without an interruption. CONSTITUTION: An MP(Main Processor) loads a program necessary for operating processors from an OS(Operating System)(S310). A control processor senses current states of each cable, and records the sensed current states in a bus state register(S320). The control processor continuously checks the current states of the cables, and decides whether the cables are dismounted or fail(S330). If so, the control processor informs each process that the cables are dismounted or fail(S340). Each processor decides whether communication is incapable(S350). If so, each processor re-decides whether the communication is capable(S370). If not, each processor outputs error messages(S380). If the communication is capable, the corresponding processors are switched, and the control processor switches contents stored in a memory block and duplicates the switched contents(S390). The communication is restarted through the corresponding processors(S395).

Description

Method for controlling cross duplication of duplication processor

The present invention relates to a cross-duplex control method of a redundant processor, and more particularly, in a communication system (eg, local exchange, private exchange, other communication equipment, etc.) having two or more redundant processors, interworking between arbitrary redundant processors. The present invention relates to a cross-duplex control method of a redundant processor that detects when a cable is disconnected or has an abnormality and searches for a communicable path and quickly switches to provide a communication service without interruption.

In general, various types of communication equipment, including a local exchange and a private exchange, are operated (interlocked) with major processors in a duplication manner to provide an uninterrupted service. Here, the duplication scheme is one of the processor configuration schemes applied to the system of the medium-large class (for example, the exchange) in order to configure the communication system without interruption. To implement the redundancy method, two identical processors are mounted (attached) in the system as a set, and the active processor among the two processes is' active processor 'and the waiting processor is' Standby processor (standby processor).

In general, if a communication failure (for example, a cable hernia) occurs during the operation of the active processor, and the service cannot be provided, the redundant processing is controlled so that the standby standby processor is transferred to the active processor to provide a service without interruption. When the system is initially powered up, the active processor is running. The standby processor is loaded into the active processor side memory but is not actually running. After that, when a communication failure occurs in the active processor, the standby processor is started, and the contents of the memory on the active processor side are copied to the memory on the standby processor side. Therefore, the operation of the active processor is reproduced as it is in the standby processor. In order to implement redundancy, software and hardware must be supported at the same time, which is implemented in a formal manner in most communication equipment.

However, such a conventional duplexing system has a problem in that it is difficult to be applied to a large communication system having a plurality of processors and requiring interworking between the processors. For example, when a connection cable between a pair of processors (that is, an active processor or a standby processor) is removed, the mutual switching operation cannot be performed. Recently, the interworking method between the dual processors is a trend made through the motherboard, but in the case of configuring a single system by combining the processors used in different systems, the interworking between the processors can be implemented through a cable, Such a cable interlocking method may cause obstacles such as cable sheathing and breaking. Accordingly, in implementing a processor interworking method through a cable, a standard for a switching operation should be prepared when a communication failure occurs. However, since the standard is not provided, it is difficult to perform a switching operation between processors.

In addition, in a communication system in which a plurality of pairs of processors are connected by cables, it is difficult to determine which cable is disconnected in the event of communication failure due to cable sheathing, and the standard for switching operation is ambiguous. There is a difficulty to perform.

Accordingly, the present invention has been made to solve the above-mentioned conventional problems, an object of the present invention is to provide a communication system in which at least two redundant processors are implemented and the redundant processors interface via a cable. The present invention provides a cross-duplex control method of a redundant processor capable of providing an uninterrupted communication service by detecting a disconnection or failure of an interworking cable between redundant processors and searching for a communication path.

1 is a connection state diagram for a redundant processor in an exchange system to which a cross redundancy control method according to the present invention is applied;

Figure 2a is a functional block diagram showing the internal function of the OMP according to the present invention divided into blocks,

Figure 2b is a functional block diagram showing the internal function of the MP according to the present invention divided into blocks,

3 is a flowchart illustrating the overall operation of the cross redundancy control method according to the present invention;

3A is an exemplary diagram illustrating a case in which a cable ④ has a hernia or a failure.

3b is an exemplary diagram showing a case in which a herniation or a failure occurs in the cable ①.

Figure 3c is an exemplary view showing a case in which a hernia or a failure occurs in the ① and ④ cable.

♣ Explanation of symbols for the main parts of the drawing ♣

100: first OMP 200: second OMP

300: first MP400: second MP

The present invention for achieving the above object includes a first and a second MP to perform a call processing function and a first and a second OMP to process an operator command and perform a maintenance function, the first and second MP And a first and a second OMP are inter-processor redundancy control methods of a switching system connected to each other by different cables, the method comprising: loading an OMP and supporting loading of a corresponding program by a loading request signal of the MP; Sensing the current cable condition and writing it to a specific register; Determining whether a cable has failed and notifying the corresponding processors currently in the active state of the cable failure; Determining whether the inter-processor communication in the active state is incapable of maintaining the current communication path if it is not in the disabled state, and determining whether there is a possibility of resuming communication after switching if the incapacity state is not available; And outputting an error message if there is no possibility of communication resumption, performing a transfer operation to a communication path through which communication is possible, and copying the contents stored in the memory block of the processor before the transfer to the memory block of the processor after the transfer. Characterized in that.

Hereinafter, a preferred embodiment of a cross-duplication control method of a duplication processor according to the present invention will be described in detail with reference to the accompanying drawings.

1 is a diagram illustrating a connection state of a redundant processor in an exchange system to which a cross-duplex control method according to the present invention is applied, and a connection between a first operating maintenance processor (100) and a first main processor (200) is shown. ① cable is connected, and between the second OMP 300 and the first MP 200 is connected by cable ②, and between the second OMP 300 and the second MP 400 is connected to cable ③. Connected by the first OMP 100 and the second MP 400 are defined as connected by the cable ④. Here, the interworking between the OPM (100, 300) and the MP (200, 400) is made through a LAN, and has a structure in which a cable is mounted with a connection port to the OMP and MP. The first and second OMPs 100 and 200 have the same configuration and operation, and the first and second MPs 200 and 400 also have the same configuration and operation. The first and second OMPs 100 and 200 perform maintenance functions for initial loading, status management for processors and devices, statistics management, operator commands, and the like. Performs the main functions of processing state management, statistics management, call processing, etc. for the processor and device. Interworking and data communication are required for smooth work coordination between the OMPs 100 and 300 and the MPs 200 and 400. For example, when the OMP requires the operator's current MP status and information, when statistical data is transmitted from the MP to the OMP, and when the OMP continuously requests and responds to the MP status to monitor the MP status. Interworking and data communication between the two takes place.

FIG. 2A is a functional block diagram illustrating the internal functions of the OMPs 100 and 300 according to the present invention by block, and FIG. 2B illustrates the internal functions of the MPs 200 and 400 according to the present invention by block. One functional block diagram.

As shown in FIG. 2A, the OMP 100, 300 reads a corresponding operating system (OS) from a disk to perform an initial loading operation, and in response to a load request signal from the MP 200, 400, is suitable for the processor. Loading control block 110 to read and deliver the OS, the operator interfacing block 120 to receive the operator command to interface the command, whether the current state of the normal or abnormal and whether it is active or standby and lower A state management block 130 for identifying and managing the states of the processors and the like, a memory block 140 for storing various information necessary for the operation and management of the processor, a communication block 150 for exchanging information between the processors, and a state management block ( And a duplication processing block 160 for checking a communication failure state and then switching to a communication path after receiving a notification of a failure. There is. Interprocessor communication using the communication block 150 uses Inter Processor Communication (IPC). IPC is one of the communication methods and techniques between processors, and is a communication method on a transport (transmission) level such as, for example, TCP / IP. If data from one processor needs to be used by another processor, interprocessor communication is required. In order to implement IPC, a communication path between both processors to be communicated is required. Generally, files, sockets, signals, and databases are used as communication paths. It is preferable to use a signal method. However, it is obvious that files and databases may be implemented as communication paths.

As shown in FIG. 2B, the MPs 200 and 400 transmit a load request signal to the OMP, receive a corresponding OS, initialize a memory, and perform a specific program (for example, an application required for the operation of the processor). 210, a call processing block 220 for processing and managing various input / output calls, a statistics block 230 for collecting statistical data, analyzing and processing the corresponding statistical data, and managing the corresponding data for each statistical item, and a processor. After receiving the failure notification from the memory block 240 and the state management block 130 for storing various information necessary for operation and management, the redundant processing block 250 for checking a communication incompatibility state and switching to a communication path. It is configured to include.

In general, an exchange system has control processors such as OMP 100, 300 and MP 200, 400, and processing for a particular protocol is through a particular processor. For example, processing of the AAL5 layer of the ATM protocol is performed in the A5CC or the like. Each control processor has a register for recording alarm information such as cable disconnection or failure, and collects alarm information of each control processor through a specific board and displays the alarm and failure status in a specific location. Such a control processor includes a bus status register for recording the current state of the interprocessor communication path (i.e. cable), a clear register for initializing the bus status register value, a control register for instructing communication cable selection, and the like. An exemplary configuration of the above-described bus status register is shown in Table 1 below.

TABLE 1

0 One 2 3 4 5 6 7 ialarma ialarmb ialarmc ialarmd astclkf lastf - -

Here, 'ialarma to ialarmd' is a bit string for recording the current state of the cables (1 to 4), and 'astclkf' is a bit for recording the presence or absence of a transaction occurring for a predetermined time (for example, 2 ms). 'Lastf' is a bit string for recording a bus occupancy state of a predetermined byte (eg, 1024 bytes) or more. For example, if cable ① fails, the 'ialarma' bit string is changed to a high bit (ie, '1'), and if the transaction does not occur for a certain period of time (for example, 2 ms), the 'astclkf' bit The string is switched to a high bit (i.e., '1'), and the 'lastf' bit string is switched to a high bit (i.e., '1') if the bus is occupied by a certain number of bytes (e.g., 1024 bytes). Each item of the bus status register shown in Table 1 is merely exemplary, and other items may be added or changed. However, in the present invention, an 'ialarma to ialarmd' bit string entry that records the current state of the cable is essential. When the generated event is restored, the clear register initializes the bit value of each item to a low bit (ie, '0').

Hereinafter, the operation relationship of the cross-duplication control method of the redundant processor according to the present invention will be described in more detail with reference to the accompanying drawings.

Figure 3 is a flow chart illustrating the overall operation of the cross-duplex control method according to the present invention, Figure 3a is an illustration showing a case in which the cable ④ hernia or failure occurs, Figure 3b is a cable sheath ① or failure occurs 3C is an exemplary diagram illustrating a case in which a hernia or a failure occurs in cables ① and ④. Prior to describing the overall operation of the present invention, the present invention describes an embodiment applied to the exchange system using a medium-large-scale communication equipment (for example, a domestic exchange and a private exchange), the dual interprocessor interworking is implemented through a cable In this case, it is applied when a cable is disconnected or failed, and it is assumed that it does not apply to a failure inside a redundant processor.

First, while power is supplied to the exchange system, the OMPs 100 and 300 and the MPs 200 and 400 initialize the memory and perform loading of programs necessary for the operation of the processor from the OS (S310). Here, since the OMPs 100 and 300 are directly connected to the disc, the necessary programs can be loaded from the disc. However, since the MPs 200 and 400 are not directly connected to the disc, the OMPs 100 and 300 are not connected directly to the disc. Send to (100, 300) to receive the desired program to perform the loading. The communication between the OMP (100, 300) and the MP (200, 400) uses a signal communication IPC, the communication block 150 of the OMP (100, 300) and the state management block 240 of the MP (200, 400) ) Is responsible for interprocessor communication. The OMP 100, 300 receiving the loading request signal from the MP 200, 400 transmits data through a handshake method that transmits a corresponding program when the loading preparation completion signal is transmitted from the MP 200, 400. desirable. In addition to the data required for loading, it is preferable to use a handshake method for data exchange between the OMPs 100 and 300 and the MPs 200 and 400.

Next, the control processor detects the current state (ie, whether she is disconnected and whether a failure has occurred) of each cable that is redundantly connected and writes it to the bus state register (S320). In the event of a cable failure or failure, the 'ialarma to ialarmd' entry in the bus status register is written as a low bit (i.e. '0'). If a cable failure or failure occurs, the entry is a high bit (i.e. '1'). Thereafter, during operation of each processor, the control processor continuously checks the current cable state to determine whether a cable is disconnected or has a failure (S330). If cable hernia or failure does not occur, return to the initial stage and repeat the subsequent steps. If cable hernia or failure occurs, cable hernia to each processor (ie OMP (100, 300), MP (200, 400)). Or notifies of the fact that the failure occurs (S340).

Each processor that receives a notification about cable hernia or a failure occurs is determined to be in an incapable communication state due to a cable hernia or a failure (S350), and maintains the status as it is not in a communication state (S360). After re-determining whether the communication can be resumed (S370), if communication resume is not possible after the transfer, an error message is output (S380) .If communication is resumed after the transfer, the corresponding processor is transferred and stored in the memory block of the processor before the transfer. After the transfer, the contents are copied to the processor memory block (S390). After the transfer operation is performed, communication is resumed through the corresponding processor (S395), and then the process returns to the initial stage and the subsequent steps are repeated. The memory content copy after the changeover operation allows the operation before the changeover to be performed as it is after the changeover, and the memory content copy is performed through a specific channel (for example, C or D channel). Although the present invention is implemented so that the switching operation by copying the contents of the memory is implemented, the processor before the transfer and the processor after the transfer may share the same memory to share various information generated during communication. In such a switching operation using the shared memory, an operation such as copying the contents of the memory is unnecessary, and there is an advantage in that communication can be performed simultaneously with the switching.

For example, as illustrated in FIG. 3A, when the cable ④ of the cable # 4 is generated when the first OMP 100 and the first MP 200 are in an active state, there is no problem in the communication path. That is, since the cable between the first OMP 100 and the first MP 200 in the active state (cable ①) is normally connected, the current state is maintained. However, as shown in FIG. 3B, when the first cable OMP 100 and the first MP 200 are in an active state, when the cable ① of the cable No. 1 is disconnected or a fault occurs, communication becomes inoperable. 200 is transferred to the second MP 400 (the first MP in the active state is switched to the standby state, and the second MP in the standby state is switched to the active state) and the first OMP (100) through the cable ④ ) And communication between the second MP 400. Since the switching operation supports software switching through a library at the OS level in most redundant processors, the present invention also performs a switching operation by utilizing a library provided by the corresponding OS. For example, in the present switching system, the transfer operation is performed by using a function 'er_restart' in the duplication processing blocks 160 and 260 of each processor.

In addition, as shown in Figure 3c, when the first OMP (100) and the first MP (200) is in the active state of the cable ① and ④ when the sheath or failure of the first MP (200) is removed. Even when switching to 2 MP 400, communication becomes inoperable state. At this time, the first OMP 100 is switched to the second OMP 300 (the first OMP in the active state is switched to the standby state, and the standby state is 2 OMP to the active state) so that the communication between the second OMP 300 and the first MP 200 through the cable ②. In addition, the switching operations for the cable sheath or failure of each processor in the active state may be as shown in Table 2 below.

TABLE 2

In the embodiment of the present invention, the switching operation may be automatically performed through software, but may also be performed manually by an operator's switching command. For example, as shown in FIG. 3A, when the first cable OMP 100 and the first MP 200 are in an active state, the operator may select the first MP 200 when a cable ④ of the cable ④ is interrupted. When a command to switch to the second MP 400 is input, the first OMP 100 and the second MP 400 become active. However, the current state of hernia or failure of cable ④ causes the communication to fall into an inoperable state. Therefore, in order to prevent this, when an incorrect transfer command is input, it is preferable to support an error message informing that the operator cannot perform the transfer command. In addition, even if a cable in which a hernia or a fault occurs after the transfer operation is restored to normal, it is preferable to operate the current communication state if the current state is a communication state.

The above description is only for explaining one embodiment, and the present invention is not limited to the above-described embodiment and can be variously changed within the scope of the appended claims. For example, the shape and structure of each component specifically shown in the embodiments of the present invention may be modified.

As described above, according to the cross-duplexing control method of the redundant processor according to the present invention, by detecting a disconnection or abnormality between any redundant processor between the redundant processor and detect the communication path and quickly switch, communication service without interruption There is an effect that can provide.

Claims (4)

First and second MPs that perform call processing functions and first and second OMPs that perform operator functions and perform maintenance functions, wherein the first and second MPs and first and second OMPs In the processor-to-processor redundancy control method of each switching system connected by different cables, (a) loading the OMP and supporting loading of the corresponding program by the loading request signal of the MP; (b) detecting the current cable condition and writing it to a specific register; (c) determining whether the cable has failed and notifying the corresponding processor currently in the active state of the cable failure; (d) determining whether the inter-processor communication in the active state is incapable of maintaining a current communication path if not in an inactive state, and determining whether there is a possibility of resuming communication after switching if incapable of incapacity; And (e) If there is no possibility of resuming communication in step (d), an error message is output, and if there is a possibility of communication resumption, the transfer operation is performed to a communication path that is capable of communication. Cross copying control method of a redundant processor comprising the step of copying to a memory block. The method of claim 1, wherein the switching operation of step (e) first switches any MP in an active state to a standby state, and then transfers another MP from a standby state to an active state, and is incapable of communicating even after switching of the MP. And converting any of the OMPs in the active state into the standby state and then switching the other OMPs from the standby state to the active state. The method of claim 1 or 2, wherein the switching operation is performed by either automatic switching using software or manual switching by an operator command. A method of cross redundancy control of a redundancy processor, comprising: outputting an error message and maintaining a current communication path when a transfer command from the operator to a failed cable is input when a cable having no error in a communication path occurs. The cross-duplication of the redundant processor of claim 1 or 2, wherein the current communication path is maintained if the current state is in a communicable state even when the failed cable is restored to normal after performing the switching operation. Control method.