KR20020093544A - Duplication method for real-time fault tolerance in wireless communication system - Google Patents

Abstract

PURPOSE: A duplication method for admitting fault in real time in a wireless communication system is provided to perform a successive service as to a call in progress as well as a set call when the fault is generated during the operation of a system by transmitting a changed item of a memory to a standby side while application tasks are operated by a message unit and logging a message for recovery when the application tasks are switched. CONSTITUTION: When performing an initialization, a registration message for registering its storage region is transmitted to a duplication module(201). An IPC(Inter Processor Communication) terminal receives an external message(202), and a corresponding application block processes the received message(203) and transmits backup data to the duplication module(204). It is judged whether the backup data are transmitted to the standby side(205). If the backup data are transmitted to the standby side, a check point message is transmitted to the duplication module(206).

Description

Duplication method for real-time fault tolerance in wireless communication system

The present invention relates to a duplication method for real-time fault tolerance in a wireless communication system and a computer-readable recording medium recording a program for realizing the method. Particularly, an application task operates in a message unit, By sending the changes to the standby side and logging messages for recovery during switchover, the system can provide continuous service not only for established calls but also for ongoing calls, even if a failure occurs during system operation. The present invention relates to a duplication method for real-time defect tolerance that improves stability and a computer-readable recording medium having recorded thereon a program for realizing the method.

The present invention, for example, to develop a radio network controller (RNC: Radio Network Control) of the next generation mobile communication (IMT-2000) system as a fault-tolerant system, which adopts a real-time fault-tolerant method, not just RNC. It can be used for all systems. In particular, it can be applied to the Node-B and the switch in the IMT-2000 system.

Conventional duplication methods are largely divided into a simultaneous operation method, a simultaneous write method, a state check pointing method.

First, in the simultaneous operation method, since the active processor and the standby processor execute the same code in separate hardware modules, only one external module needs to change the external communication path when a module needs to be replaced by a fault. Real time fault tolerance is possible. However, there is a limitation that the active processor and the standby processor must be started at the same time in the beginning, and if the standby operation is late, a process of collectively replicating the storage of the active processor is required.

In RNC, since the basic unit of execution is the call processing message, there may be a case where the context is not matched between tasks due to a temporary defect, which may cause a problem of system stability. Simultaneous operation method can service completely continuously when switching, but there is a problem that it is difficult to synchronize the synchronization between the two processors as the performance of hardware rapidly improves.

Next, the simultaneous write method should be applied to the storage of the standby processor at the same time as the active processor changes the contents of the storage. This requires special hardware that uses shared storage or supports concurrent writes. When switching is required due to a failure, only the context of the active processor is copied and applied to its own context, which is then activated to continuously perform the tasks performed by the active processor. This method has a dual down because the malfunction of the active side is passed to the standby side as it is, and the hardware configuration for simultaneous writing is not easy, and special hardware equipment is required, which adds cost. .

Lastly, the state check pointing method operates only the active processor and notifies the standby processor when a state change occurs. If an error or fault occurs during the operation of the active processor, the standby processor recovers based on the latest checkpoint data. Log information includes context status information, register values, and input queues. This method has high recovery rate because it has status information for each important task. However, this method is difficult to implement and has a large overhead of the redundant module.

The automatic checkpointing technique automatically generates log information at the operating system (OS) level that supports redundancy and delivers it to the standby processor. This is proposed to reduce the burden of implementing the status checkpointing method. However, there is a problem in that a cost increase to build a new OS that supports this, and the amount of data to be sent to the standby is rapidly increased, which is expensive to manage.

The present invention has been proposed to solve the above problems, the application tasks (Tasks) to operate in the unit of a message to transfer the changes of the memory to the standby side, logging the message (logging) for recovery when switching By providing a duplication method for real-time fault tolerance that enables continuous service for ongoing calls as well as calls established when a failure occurs during system operation, and a computer-readable recording medium recording a program for realizing the method. The purpose is.

1 is an exemplary view illustrating a configuration of an interworking redundancy defect tolerance system to which the present invention is applied.

2 is a flowchart illustrating a duplication method for real-time fault tolerance in a wireless communication system according to the present invention.

3A is a flowchart illustrating an operation of an active-side duplication module of a duplication method for real-time fault tolerance according to the present invention.

3B is a flowchart illustrating an operation of a standby side duplication module in a duplication method for real time fault tolerance according to the present invention;

Figure 4 is a flow diagram of an embodiment of a recovery process of the standby side when switching among the duplication method for real-time fault tolerance according to the present invention.

5 is a flow diagram of an embodiment of an initial synchronization procedure of a duplication method for real time fault tolerance according to the present invention;

6 is an exemplary diagram illustrating a redundancy control algorithm at a logic level in accordance with the present invention.

* Explanation of symbols for the main parts of the drawings

11 Active Processor 12 Standby Processor

The present invention for achieving the above object, in a duplication method for real-time fault tolerance in a wireless communication system, in order to operate each of the application tasks (Tasks) loaded on the active side and standby side by message unit, active processor and standby A first step of initializing the processor; A second step of registering each application task in its memory for each application block and executing an active processor; And transmitting a change in memory to a standby processor while operating the application tasks in a message unit, and in the event of a transfer due to a fault on the active side, creating log information on a message for recovery and transmitting the changed log information to the standby side. Characterized in that comprises a.

The present invention also provides a wireless communication system including a processor, comprising: a first function of initializing an active processor and a standby processor to operate each application task loaded on the active side and the standby side in a message unit; A second function of executing the active processor after registering the respective application tasks in its own memory for each application block; And a third function of transmitting a change in memory to a standby processor while operating the application tasks in a message unit, and writing log information on a message for recovery and transmitting the changed information to a standby side when a changeover occurs due to a fault on the active side. A computer readable recording medium having recorded thereon a program for realizing the present invention is provided.

The present invention is based on a method of transferring application memory changes to the standby side while application tasks operate on a per message basis, and is an application block interworking duplexing scheme in which a message is logged for recovery when switching. Since the synchronization function between the active and standby processors is in units of messages, the contact status information and the register value of the application task are not necessary. Application tasks directly process the message, and then write log information about the message and send it to the standby, so synchronization is more accurate and easier to synchronize, and program implementation is easier. In addition, even if a fatal failure occurs due to a malfunction, it is immediately switched over and does not transmit wrong information to the standby. This results in a dual down prevention.

The above objects, features and advantages will become more apparent from the following detailed description taken in conjunction with the accompanying drawings. Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is an exemplary configuration diagram of an interworking redundancy defect tolerance system to which the present invention is applied.

As shown in FIG. 1, a path between an active processor and a standby processor for redundancy is constructed as 100base-T Ethernet because data and control signals must be transmitted reliably.

The processor in charge of call processing is divided into call setup and call release. After a call is set up, even if a fault occurs, the call is maintained for the set call. However, if a fault occurs during the call setup process, the call in progress is released. Therefore, if the status is matched by a message unit that changes the contents of a storage location, rather than a process for establishing and releasing a call, the service can be continuously performed even for an ongoing call.

2 is a flowchart illustrating a duplication method for real-time fault tolerance in a wireless communication system according to the present invention.

As shown in FIG. 2, upon initialization, a registration message for registering its storage area is transmitted to the duplication module (201).

Subsequently, an external message is received at the Inter Processor Communication (IPC) terminal (202), the message is processed in the corresponding application block (203), and then backup data (backup data) is transmitted to the duplication module (204).

Then, it is determined whether to transmit the backup data to the standby side (205). If the backup data is not transmitted, the process proceeds to the step 202 of receiving an external message from the IPC stage. Send 206.

That is, in the application block-based duplication method, each application task loaded on the active side and the standby side registers its storage area with the duplication module at initialization. This process must be performed by the same procedure on both the active and standby processors. The active processor and standby processor then receive the call message at the same time.

3A is a flowchart illustrating an operation of an active side duplication module in a duplication method for real time fault tolerance according to the present invention.

As shown in FIG. 3A, the operation process of the active-side redundancy module may register a registration message from each application block if the external communication channel (receive / transmit (Rx / Tx)) is enabled (301). Received (302) to manage the storage area for each application block (303).

Subsequently, the message information is received through the IPC terminal (304), the message information is transmitted to the redundant module on the standby side (305), and the backup message is received from the application block (306) to manage the backup data for each application block. (307).

Thereafter, a checkpoint message is received from the application blocks (308), and the backup data and message information managed for each application block are transmitted to the redundancy module on the standby side (309).

That is, the IPC side on the active side forwards this message to the corresponding application task to be processed and at the same time transmits information about the message to the standby side. If the content of the storage location is changed as a result of the message processing, the backup message (the changed details and information on the message) is transmitted to the duplication module. The redundancy module manages this for each application task. When the application task calls a checkpoint, the redundancy module sends it to the redundancy module on the standby side. Here, the reason for the checkpoint is to give applicability so that the application block can manage the backup message by itself.

3B is a flowchart illustrating an operation of a standby side duplication module in a duplication method for real time fault tolerance according to the present invention.

As shown in FIG. 3B, the operation process in the standby-side redundancy module may receive a registration message from respective application blocks when the external communication channel (transmission (Tx)) is disabled (311) (see FIG. 3B). 312, the storage area is managed for each application block (313).

Subsequently, the message information is received at the IPC stage (314), and the messages are sorted in order of messages processed by the active side for each application block and managed as a hash table (315).

Thereafter, the backup message is received from the duplication module on the active side (316), the backup data is applied to the storage area, and the message processed by the active side is removed using the message information (317).

That is, the standby side blocks the external transmission (Tx) at the IPC terminal and stores the message input from the outside in the form of a queue without transmitting the message to the corresponding application task. And, for efficient message management, a hash table with a task ID as a key is created and managed for each message. When the active side sends information about a message, it inserts it into a hash table managed by each application task, and when a data backup message is received, it is reflected in its own storage. Then, using the message information transmitted along with the backup message, the corresponding task is examined in the hash table, the matching message is found among them, and all previous messages are deleted.

Figure 4 is a flow diagram of an embodiment of a recovery process of the standby side when switching among the duplication method for real-time fault tolerance according to the present invention.

As shown in FIG. 4, the standby process recovery procedure is performed when the external communication channel (receive / transmit (Rx / Tx)) is enabled (401) and remains in the buffer of the IPC terminal while hashing. The message managed by the table is transmitted to the corresponding application block (402).

After all messages in the hash table have been transmitted to the corresponding application block, the role is changed to an active state (403).

That is, when a switchover occurs due to a defect in the active processor, the standby processor processes the remaining message and the newly input call processing message without being deleted from the input queue. In this process, the same message may be processed twice, which causes the generation of garbage resources and may cause the resource to be exhausted. However, in the application block interworking duplication method, the application task can solve the problem by checking whether the resource is useful or garbage.

5 is a flowchart illustrating an initial synchronization procedure of a duplexing method for real-time fault tolerance according to the present invention.

As illustrated in FIG. 5, when the standby side duplication module transmits a memory copy request message to the active side duplication module (501), the standby side duplication module receives a memory copy request message through the standby side duplication module (502).

Subsequently, the active side redundancy module holds a backup message of the corresponding application block of the memory area to be replicated (503), transmits the memory area of the corresponding application block to the standby side redundancy module (504), and then the corresponding application block. When releasing of (505), the standby side redundancy module reflects this to the memory area (506).

In other words, when the active processor is initially started and the standby is started late, the memory contents of the active side and the standby side are different, so memory duplication is necessary to synchronize them. However, replicating them in bulk can result in the loss of messages for other purposes at one moment due to the large amount of data transfer. As a solution to this problem, a storage area managed by each application task is sequentially copied, and a technique of temporarily holding a backup message so as not to transmit a backup message for an application task related to the replicated storage area is used.

This prevents the loss of messages sent by other application tasks. Active, standby role determination is made through a failure check (see FIG. 6) at the hardware level.

The method of the present invention as described above may be implemented as a program and stored in a computer-readable recording medium (CD-ROM, RAM, ROM, floppy disk, hard disk, magneto-optical disk, etc.).

The present invention described above is not limited to the above-described embodiments and the accompanying drawings, and various substitutions, modifications, and changes are possible in the art without departing from the technical spirit of the present invention. It will be apparent to those of ordinary knowledge.

The present invention as described above, in order to develop a radio network controller (RNC) of the next generation mobile communication (IMT-2000) system as a fault-tolerant system, the application tasks are operated in the unit of messages, the change of memory to the standby side By using this method, a message logging method is used for recovery when switching. Therefore, even if a failure occurs during system operation, continuous service can be provided not only for the established call but also for the ongoing call. This can maximize the stability of the system to provide efficient service, and by reducing the maintenance cost, the reduction in service price can bring an economic effect.

In addition, the present invention has an excellent effect of preventing a large economic loss indirectly because continuous service can be performed even if a failure occurs for a call requiring call setup, a call in progress of call setup, or a call already established.

Claims (6)

In the duplication method for real-time fault tolerance in a wireless communication system, A first step of initializing the active processor and the standby processor to operate each application task loaded on the active side and the standby side in a message unit; A second step of registering each application task in its memory for each application block and executing an active processor; And The third step of transferring the changes of the memory to the standby processor while operating the application tasks in the message unit, and in the event of a transfer due to a fault on the active side, to write log information on the message for recovery to send to the standby side Redundancy method for real-time fault tolerance, including. The method of claim 1, The first step is, During initialization, if the active processor is started first and the standby is started late, the memory contents of the active side and the standby side are wrong and memory duplication is necessary to synchronize the memory areas. A method of redundancy for real-time fault tolerance, which employs a technique of duplicating and temporarily holding a backup message so as not to send a backup message for application tasks related to the area of memory being replicated. The method according to claim 1 or 2, The second step, Through the IPC (Inter Processor Communication) terminal on the active side, the message is delivered to the corresponding application task to be processed and the information on the message is transmitted to the standby side. If a change is made, a backup message (change history and information on the message) is transmitted to the redundancy module, and the redundancy module manages it for each application task, and when the application task calls a checkpoint, Redundancy method for real-time fault tolerance, characterized in that the transmission to the redundancy module. The method of claim 3, wherein The third step, Blocking the external transmission (Tx) in the IPC stage, and stores the message input from the outside in the form of a queue without transmitting to the application task, and task ID for each message for efficient message management It creates and manages a hash table with keys, and inserts it into the hash table managing the application task when information on the message is delivered from the active side, and reflects it in its own storage when the data backup message is received. Investigate the corresponding task in the hash table using the message information transmitted with the message, find a matching message among them, and delete all previous messages up to the message. . The method of claim 4, wherein The transfer recovery process due to the defect of the active side, When a transfer occurs due to a defect in the active processor, the standby processor processes the remaining message and the newly input call processing message without being deleted from the input queue, and checks whether the resource is useful or garbage in the application task. Redundancy method for real-time fault tolerance, characterized in that the recovery. In a wireless communication system having a processor, A first function of initializing the active processor and the standby processor to operate each application task loaded on the active side and the standby side in a message unit; A second function of executing the active processor after registering the respective application tasks in its own memory for each application block; And A third function of transmitting a change in memory to a standby processor while operating the application tasks in a message unit, and writing log information on a message for recovery and transmitting the changed information to a standby side when a changeover occurs due to a fault on the active side; A computer-readable recording medium having recorded thereon a program for realizing this.