KR100298319B1 - Redundancy Device in Communication System_ - Google Patents

Abstract

The present invention relates to a duplexing device of a communication system, and in particular, in a communication system requiring high performance such as next generation mobile communication system (IMT-2000), etc., to improve the reliability and availability of the system and to provide continuous and stable service. The present invention relates to a redundancy device based on a log-based rollback-recovery method that can provide a. The present invention relates to a duplication device of a communication system, wherein a failure detection module for reporting a failure of a counterpart board to a higher processor and determining a changeover according to a duplication board, and matching a message scanning sequence and backing up data during a backup time. An operation module including a synchronization module for removing a message logged on the standby side, and a backup module which waits until the processing of the scanned message is finished and starts the backup of the backup target area when all processing is completed; A recovery module for processing the fault detection module, the synchronization module, the backup module, and an input message stored in an input queue up to a moment before the transfer from the standby side to the operating side and restoring to the state performed by the operating side before the transfer. And a standby module further comprising: a communication link providing a message transmission path between the operation module and the standby module.

Description

Redundancy Device in Communication System

The present invention relates to a duplexing device of a communication system, and in particular, in a communication system requiring high performance such as next generation mobile communication system (IMT-2000), etc., to improve the reliability and availability of the system and to provide continuous and stable service. Log-Based Rollback-Recovery (Log-Based Rollback-Recovery) based on the redundancy method to provide a.

Log-Based Rollback-Recovery is a duplication method commonly used in a typical message pass system. This log-base rollback-recovery method uses check point and message logging together, and the system performs the same operation when the message injection order to the system is the same. Characterized in that it is to achieve. This indicates that the message delivery system can be implemented as a finite state machine (FSM), so that the operation of the system is always constant according to the events that occur.

Looking at the duplication method according to the log-base rollback-recovery method described above, the system first performs the operation through the log-bass rollback-recovery protocol, in which logging and checking points of messages are stored in a stable storage device such as a hard disk. Save it. Then, the system recovers after a fault occurs in the system. In this case, the information of the logged message corresponding to the stored part is re-executed to recover the operation before the fault occurs. will be.

These log-basis methods are generally classified into pessimistic logging, optimization logging, and casual logging, each of which has three characteristics as shown in Table 1 below. Will have

Pessimistic logging Optimistic logging Casuallogging Over head Very high height height Output commit Very fast Slow speed Garbage Collection simple complication complication Recovery simple complication complication Orphans impossible possible impossible Rollback extent Last checkpoint Some previous checkpoint Last checkpoint

Meanwhile, a communication system that is commonly used is also operated by a message, which shows that the communication system may also be referred to as a message pass system. This indicates that even in a communication system, other redundancy may be achieved by applying the log-base rollback-recovery method as described above.

However, the communication system should have the characteristic that the service should be continuously available (availability) as soon as possible due to the changeover of the module generated when a failure occurs in the redundancy target module (Module), and before the transfer to the service state after the changeover. The inconsistency of the states must have a property that should be minimized (reliability). Ultimately, availability requires fast recovery times, and minimizing state inconsistencies requires accurate execution of output commits and garbage collection. And in order to minimize the overhead in the system during normal operation, it is required to perform the operation to minimize the overhead according to message logging.

However, the system redundancy according to the log-base rollback-recovery method as described above also coexists with the disadvantages as described in Table 1 above, so that these disadvantages require high-performance, next-generation communication. There were a lot of crowds in the application of the system. Therefore, the redundancy according to the log-bass rollback-recovery method has a problem that is not properly applied in the next generation communication system.

Accordingly, an object of the present invention is to provide a redundancy apparatus for applying the log-bass rollback-recovery method in redundancy in a next-generation mobile communication system, and also has disadvantages of the log-bass rollback-recovery methods. The present invention provides a redundant device.

More specifically, the present invention provides a redundancy apparatus employing a more reliable and stable rollback-recovery method by appropriately utilizing the advantages of pessimistic message logging and optimized message logging.

In the present invention to achieve the above object to achieve a redundancy device that achieves the following operation. First, the redundancy apparatus according to the present invention allows two boards to be duplicated on the hardware having the same structure to receive the same input. At this time, one board performs a normal operation with an active module and the other board stands by. By) Log input message to module. In addition, the message logging sequence in the standby module always operates by asynchronously providing information on the message input when the operation module inputs an external message to the standby module to the standby module through a separate communication link. Match the message execution order in the module. At this time, the operation module and the standby module are guaranteed to receive the same message without losing the mutual message.

With the proposal of a redundancy device that accomplishes this operation, it is possible to take advantage of the advantages of pessimistic message logging without the overhead of input messages in pessimistic message logging. Proposals can be made.

That is, implementing a duplexing device for performing the above operation is a technical problem to achieve the object of the present invention.

1 is a block diagram illustrating an internal block configuration of a duplication apparatus in a mobile communication system according to an exemplary embodiment of the present invention.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings. First, in describing the present invention through the reference numerals added to the components of each drawing, it should be noted that the same reference numerals are used even when the same reference numerals are shown on the other drawings to indicate the same components.

Also, in the following description, many specific details such as components of specific circuits are shown, which are provided to help a more general understanding of the present invention, and the present invention may be practiced without these specific details. It is self-evident to those of ordinary knowledge in Esau. In the following description of the present invention, if it is determined that a detailed description of a related known function or configuration may unnecessarily obscure the subject matter of the present invention, the detailed description thereof will be omitted.

1 is a block diagram of a duplexing apparatus in a communication system according to a preferred embodiment of the present invention.

As shown in FIG. 1, the duplexing apparatus according to the present invention is classified into a hardware configuration and a software configuration.

First, in the hardware configuration, the two boards to be redundant have a common configuration. This is because the external input can be commonly received. However, the output has a hardware configuration so that different outputs can be generated. Then, the hardware is configured to detect the normal operation and failure of the counterpart board. In addition to the existing communication path for input and output, a separate communication link is configured for communication between the two boards.

Next, in the software configuration, each board is divided into an active side and a standby side, so that the service can be supported only on the operating side, and the standby side logs and checks points of messages. ), Perform unnecessary cleanup. The operation side performs a recovery operation and a transfer operation at the time of failure inspection, and simultaneously operates as an operation module.

Each module of the software configuration has the following configuration.

Fault Detection Module (FDM) 110-1,2

Back Up Module (BM) 120-1,2

Synchronization Module (SM) 130-1,2

Recovery Module (RM) 140

First, look at the operation of the failure detection module (110-1,2),

Failure detection can be divided into hardware detection and software detection. Hardware failure detection detects board failures, which is supported in hardware.

The hardware failure detection is classified into two operations. One is to detect hardware failures in the board that may affect the operation of the system by itself, and notify the other board. The other is to detect hardware failure of the other board. In the present invention, when the failure of the counterpart board detects a counterpart board failure, the FDM reports the failure of the counterpart board to a higher processor and determines the transfer.

Software fault detection detects the fault of the other board by software, which detects the fault that cannot be detected by hardware, reports to the upper processor, and determines the switchover accordingly.

Second, looking at the operation of the synchronization module (130-1,2),

The synchronization module matches the message injection order to application tasks at the operation side and the standby side, and removes the message logged at the standby side during the data backup time. At this time, the operation side and the standby side receive the external input message at the same time at the physical layer. However, because of differences in throughput between the two sides, there is no guarantee that the two sides can receive messages in the same order at the same time through interprocessor communication (IPC). In addition, the internal message is generated only at the operation side, and as a result, the message at the operation side is serialized in a mixed form of the external message and the internal message and transferred to the application tasks. At this time, the message should be logged on the standby side in the same order as on the operation side for recovery. To this end, when the operator receives an external / internal input message, the operator transmits information about the message to the standby side. The standby side receives these messages and the information about them and logs the messages in the same order that they were delivered from the operating side to the application program. The standby side remembers the entire order for each input message and logs them separately for each application task. The reason for this is to facilitate the removal of backup-related messages from the input messages of the task when the data is backed up to the standby side, and the operation to deliver the input messages in order to each task during recovery. do. In addition, when a message loss occurs due to a failure in the internal message transfer process, an external message is generated during recovery.

A protocol is provided in the communication link between the operation side and the standby side so that logging information and backed up data are correctly transmitted. When backing up data from the operating side to the standby side, information about the last message received and processed is also delivered. This is to provide information about the messages to be removed when cleaning up unnecessary information. That is, the information about the message input just before the backup is needed for the task of removing the message that was involved in the setting of the checkpoint after the checkpoint is set. The information transmission on the message is performed by using existing information such as a header or a sequence number of the message, or a tag such as a check sum or a CRC. The backup timing and the backup data are determined by the application programmer. If multiple parts of data are updated by an input message in an application, each of these updated data must be backed up. Therefore, checkpoint setting is performed separately from backup. In this case, the checkpoint is an operation informing that a transaction has ended, and only the backup data for which the checkpoint is set is backed up to the standby module.

Third, look at the operation of the data backup module (120-1,2),

The data backup occurs when the operating side backs up data to the standby side during normal execution, and when the operating side dumps the redundant data to the standby side after the transfer. During the normal operation, each application program on the operating side knows the data to back up, and if the data has changed, it should request a backup. The request method uses the provided application program (API), which is supported at the function level. This API takes relevant data and input messages as parameters and queues them to the redundant server. The duplication server then reallocates memory for writing the queue, and copies the data area to the reallocated memory area. When copying, this API uses Critical Sections. The reason for using this hazardous area is to eliminate data contradictions caused by interference that may be caused by other tasks when this area is a shared area. The reason for writing the queue is to serialize the information on the input message and the backup data in the duplication server. When the replication server receives a backup request for the replication target data from the application program, the replication server generates a sequence number and transmits it to the standby side through the communication link. In this case, since the redundant server regards the backup data and the input message information as the same data to be transmitted without distinguishing them, the processing for them is performed the same. As mentioned above, since the request for data backup is made at the functional level, the overhead of the execution time of the duplication related functions is borne by the application program. When an application requests a backup, the data to be backed up is copied and enqueued, so it is responsible for minimizing the overhead of copying. Therefore, the overhead should be minimized through data structure definition.

When dumping data stored in the operating side to the standby side while the standby side is prepared, the operating side knows the area of data to be backed up. The area is already determined at the time of initial logging. At this time, care must be taken to ensure the authenticity of the data being backed up, and the data authenticity is as follows. First of all, if there is a message received from the outside after the logging is finished, the waiting side transmits information on the message to the operation side and requests a backup of the data. When the operation side receives a data backup request from the standby side, it examines whether there is a message received at the standby side and scans the message immediately before the message to the application task. If not, it scans up to the message it currently has and does not inject any more messages into the application tasks. After the processing of the scanned message is finished, if all processing is completed, the backup of the backup target area is started. At this time, the contents of the message of the operation side and the standby side are the same. Once all the data has been backed up, both sides perform normal operations.

Finally, looking at the operation of the recovery module 140,

The module that is newly operating when the transfer from the standby side to the operating side needs to perform a recovery. Restoration is an operation to keep track of where the previous operation was performed. If data backup is not completely performed on the previous operating side, the newly operated side is stored in the input queue until the moment when it is transferred based on the currently backed up data. It refers to an operation of recovering up to the state performed by the operating side before the transfer by processing the input message.

In the normal operation state, the transferred operating side separates and logs the messages for each task by the synchronization operation, and also knows the order of all the logged messages. Therefore, the recovery module scans each message to the corresponding task in this order. At this time, each application task operates in a normal state, so an output message can be generated. The output message generated may be an outgoing message or an internal message. In case of internal message, it checks whether it is already logged in the logged message buffer of the task that is the destination, and in case of logged message, discards this message and inserts it at the end of the task message buffer. .

As described above, the present invention can achieve a redundancy device of a communication system through a log-based rollback-recovery method, and in particular, pessimism without overhead for input messages in pessimistic message logging. You can take advantage of the benefits of message logging.

Accordingly, there is an advantage of achieving a proposal of a duplication device suitable for a next generation communication system requiring relatively high performance.

Claims (5)

In the redundant device of the communication system, A failure detection module that reports a failure of the other board to a higher processor and determines a transfer according to the detection of the other board; A synchronization module that matches the order of message scanning and removes messages logged to the standby during data backup; An operation module including a backup module which waits until the processing of the scanned message is finished and starts the backup of the backup target area when all processing is finished; The fault detection module, the synchronization module, the backup module, A standby module further comprising a recovery module for processing an input message stored in an input queue up to the instant of the transfer from the standby side to the operating side to recover the state performed by the operating side before the transfer; And a communication link providing a message transmission path between the operation module and the standby module. The method of claim 1, wherein the operation module and the standby module, Redundancy device of a communication system, characterized in that it receives the same message input. The method of claim 1, wherein the operation module and the standby module, Redundancy device of a communication system, characterized in that for performing each message output. The method of claim 1, And transmitting the input message information from the operation module to the standby module through the communication link, and matching the order of messages between the two modules through the transmission of the message information. The method of claim 1, wherein the operation module and the standby module, Redundancy apparatus of a communication system, characterized in that for transmitting the input message information from the operation module to the standby module through a communication link installed between the two modules, thereby matching the message order.