JP3342253B2 - Failure recovery method for distributed node system - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a restart control method when a failure occurs in a distributed node system.
In particular, in service control stations that make up large-scale intelligent networks, a redundant configuration of distributed modules and module state control in a centralized control module using a backup mechanism can reduce service interruption time in the event of a failure. The present invention relates to a failure recovery method for a distributed node system that realizes processing.

[0002]

2. Description of the Related Art A switching system is applied to a service control station of a conventional intelligent network, and when a failure occurs, the service control station expands an initial setting range by a restart logic defined below. While trying to recover quickly from disability.

PH1: When a failure occurs, resources related to calls other than those in a stable state (busy state) are initialized and restarted. PH2: Initialize resources for all calls and resume.

PH3: Reload the system program, execute all initial settings, and restart. In the switching system, a duplex configuration is generally adopted, and each of the above restart logics is implemented in the 0 system / 1 system respectively. If the restart cannot be completed within the time (20 minutes), the failure is treated as a serious failure. During the execution of these restart logics in the affected module, the communication service cannot be accepted, so that the service is stopped.

[0005]

As described in the section of "Prior Art", in a service control station of an intelligent network using a conventional exchange, when a failure occurs, the service is provided by the module until restart is completed. The service stops. For this reason, the service quality of the network was degraded.

An object of the present invention is to minimize the service interruption time even when a module failure occurs, by coordinating a centralized control module, a failed module, and other modules in a distributed node system employing a distributed module configuration. Another object of the present invention is to make it possible to customize the service interruption time as needed.

[0007]

In order to solve the above-mentioned problems, the present invention provides a distributed node system having a distributed module group for load distribution or function distribution, a centralized control module for centrally managing these modules, and a maintenance terminal. In the configured communication control node, when the centralized control module detects a failure that occurred in the distributed module, the communication service access to the affected module is immediately stopped in cooperation with other modules, and the restart is completed in the affected module. To immediately restore communication service access.

Further, a database system is mounted on the function distribution module in the distributed node system,
When managing update data in a communication service, a backup relationship is specified between modules, and the centralized control module monitors the time of failure recovery when a failure occurs in the function distribution module, and within a specified time If not restored, take over the process to the backup side.

According to the above method, the following can be performed in the restart processing of the communication control node. (1) In the case of a failure of the load distribution module, the traffic of the failed module is assumed by the other module using the load distribution module configuration in cooperation with the central control module and other modules, and the failure without service interruption. A recovery procedure can be realized.

(2) Regarding the failure of the function distribution module, at least the interruption to take over the service processing to the module which has a mutual backup relationship with the failed module allows the continuation of the service to be performed in the module. .

However, both (1) and (2) increase the processing load as compared to before the occurrence of the module failure. In particular, in the function distribution module of (2), the processing load is doubled compared to before the backup, so the timer value is determined based on the value of the failure recovery monitoring timer, that is, the trade-off between the service interruption time and the processing load increase.

[0012]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 shows a configuration of a service control node using a distributed module configuration. In the figure, 101 is a transmission node, and 300 is a service control node adopting a distributed module configuration. A load distribution module group 201 (hereinafter, this type of module is referred to as a load distribution module A) that controls communication with the transmission node 101, and a load distribution module group 202 (hereinafter, this type of module that holds service control logic) is provided. A load distribution module B), 203 is a function distribution module group (hereinafter, this type of module is referred to as a function distribution module C) that holds update data updated by the service control logic on a real-time database, and 204 is a load distribution module. Groups 201, 202
And a centralized control module that centrally manages a module group of the function distribution module group 203. The service control node 300 is composed of these distribution modules. Reference numeral 400 denotes a maintenance terminal.

The transmission node 101 controls a communication path corresponding to a call between terminals. The transmission node 101 notifies the service control node 300 when detecting the intelligent network service call. The service control node 300 controls the call in the distributed module group 201 to
This is performed in cooperation with step 204.

FIG. 2 is a diagram showing a condition for allocating a function for failure recovery on each distributed module in the service control node 300. As shown in FIG. 2, the load distribution modules A and B and the function distribution module C have a failure detection / restart control mechanism 500, a failure / restart progress notification mechanism 501, and another module state management mechanism 502. The central control module 204 includes a failure detection / resumption control mechanism 500, an all module status monitoring / control mechanism 503, and a failure recovery monitoring mechanism 5.
04. The maintenance terminal 400 has a failure recovery monitoring timer setting function.

1) Failure detection / restart control mechanism 500 The load balancing module B which performs call processing by the service control logic in conjunction with the call processing on the transmission node 101 has the restart described in the section of "Prior Art". Apply the logic as is. On the other hand, in the load distribution module A and the centralized control module 204 that perform communication control with the transmission node 101, the application of PH1 has no meaning since there is no management resource linked to the call processing of the transmission system. For this reason, the restart logic applies PH2 and PH3.

The function distribution module C for realizing data access processing from the service control logic also has processing only when accessing the database, and there is no management resource linked to call processing on the transmission node. Apply PH2 and PH3.

2) Fault / Restart Progress Notification Mechanism 501 Each distributed module detects a fault and performs autonomous restart logic, and the progress information during fault occurrence and restart processing is transmitted to the central control module 204. It has the function of notifying to. This function needs to be able to notify information in any case regardless of the state of the affected module, and therefore, this interface is realized by hardware logic, not software logic.

3) All module status monitoring / control mechanism 50
3. The centralized control module 204 receives the failure / restart progress notification from each module and manages the status of all modules in the service control node. When receiving the failure notification of the arbitrary module, the central control module 204 immediately sets the management state of the module to the failure state and notifies all other modules that the module state has changed to the failure state.

Thereafter, the centralized control module 204 monitors the notification of the recovery of the affected module, sets the status of the module to the normal state when the notification of the recovery is received, and changes the status of the module to all other modules. Notify that it has changed normally.

4) Other module status management mechanism 502 All distributed modules manage the status of other modules other than the own module. The other module that has received the module status of the affected module from the centralized control module 204 by the all module status monitoring / control mechanism 503 sets the management status of the affected module to the failed status.
This other module status is referred to in advance when access to another module is necessary in service processing performed by that module. If the module status is not normal, access to the module is stopped.
Select an alternative module and implement access.

As an alternative module selection method, when the affected modules are the load distribution modules A and B, another module of the same type is arbitrarily selected. On the other hand, when the affected module is the function distribution module C, a module having a mutual backup relationship with the module is selected after a certain waiting time based on the failure recovery monitoring mechanism 504 of the central control module 204.

5) Failure recovery monitoring mechanism 504 When a failure of the function distribution module C is detected in the central control module 204, the management state of the module is set to the failure state as shown in 3), and the maintenance terminal is set in advance. The time is monitored by the failure recovery monitoring timer according to the timer value designated by the failure recovery monitoring timer setting function of 400, and upon expiration of the timer, the service processing is taken over to the module which has a mutual backup relationship with the affected module.

6) Fault recovery monitoring timer setting function The maintenance terminal 400 has a function of setting the value of the fault recovery monitoring timer held on the centralized control module 204.

FIG. 3 shows an example of a procedure at the time of occurrence / recovery of a failure by function cooperation between modules. 3 is a module in which the load distribution modules A and B or the function distribution module C has failed, and the other modules 700 are other distribution modules except the failure module 600 and the centralized control module 204. It is.

The procedure when a failure occurs is shown in FIG.
It is as (5). (1) First, the fault detection /
The restart control mechanism 500 detects the failure.

(2) The failure / restart progress notification mechanism 501 notifies the central control module 204 of the failure. (3) The module status monitoring / control mechanism 503 of the central control module 204 sets the management status of the affected module 600 to failure.

(4) Then, the all module status monitoring / control mechanism 503 notifies all the other modules 700 of the occurrence of the fault in the faulty module 600. (5) Upon receiving the notification, the other module 700 sets the management status of the affected module 600 to failure. Thereafter, while the affected module 600 is in the failure state, the affected module 60
Access to 0 is prohibited.

The procedure at the time of recovery from a failure is shown in FIG.
(a) to (e). (a) In the affected module 600, the failure detection / restart control mechanism 500 executes a predetermined restart logic.

(B) The failure / restart progress notification mechanism 501 starts the failure recovery monitoring mechanism 50 of the centralized control module 204 when the restart by the failure detection / restart control mechanism 500 is completed.
4 is notified of the completion of restart.

(C) The failure recovery monitoring mechanism 504 notifies the all module status monitoring / control mechanism 503 of the recovery of the failed module 600. (d) The all module status monitoring / control mechanism 503 sets the status of the affected module 600 to a normal status,
The other module 700 is notified of the recovery of the affected module 600.

(E) The other module status management mechanism 502 in the other module 700 sets the status of the affected module 600 to a normal status. Thereafter, access to the restored affected module 600 becomes possible.

FIG. 4 shows a failure recovery procedure for a case where a failure has occurred in the function distribution module as an example. If a failure occurs in the function distribution module,
Immediately, a fault notification is sent to the central control module.

The centralized control module sets the state of the affected module to the failure state, starts measurement by the failure recovery monitoring timer, and starts failure recovery monitoring. Also, it notifies other modules of the module status of the affected module.

In other modules, access processing to the affected module is prohibited. In the affected module, an autonomous restart process is attempted. If the restart process is completed before the monitoring timer expires, the centralized control module detects the restart notification and returns the status of the module to other modules normally. Give instructions to return. Case 1 shown in FIG. 4 shows a case where the completion of the resumption is achieved by resuming the PH2.

On the other hand, if the restart cannot be completed even if the monitoring timer expires, the processing is taken over by the backup module which has a mutual backup relationship with the affected module. Case 2 shown in FIG. 4 indicates a case where the restart has been completed by the restart of PH3, but the monitoring timer has already expired.

As shown in the figure, in the case of Case 2, the service to be provided by the affected module is interrupted between the occurrence of the failure and the execution of the backup processing. On the other hand, if the failure recovery monitoring period, that is, the monitoring timer value is shortened, the service interruption time can be reduced to the time required for the backup activation process.

When only the restart logic described in the section of "Prior Art" is applied to the function distribution module, the application phases are PH2 and PH3. Assuming that the initial setting time of the system is 30 seconds and the load time of the system program is 100 seconds, PH2 requires 30 seconds and PH3 requires 130 seconds.

As described above, the function distribution module is equipped with a database system, and it is necessary to reconstruct the database system when a failure occurs. Assuming that the time for this reconstruction (that is, equal to the backup processing execution time) is 30 seconds, P
H2 requires 60 seconds and PH3 requires 160 seconds.

Therefore, for example, if the set value of the failure recovery monitoring timer value is within 0 to 30 seconds, the service can be started earlier than the restart of PH2, and the service interruption time can be shortened.

In general, even in the conventional exchange, the restart is not always completed by one restart of PH2. Therefore, this method of starting backup by using the failure recovery timer has an advantage that the service interruption time can be fixed.

By setting the timer value from the maintenance terminal 400, the service interruption time can be customized by the maintenance person. This function is
For example, this is effective in a case where the service interruption allowable time is set as a contract condition for a user who uses the function distribution module.

[0042]

According to the restart processing method of the present invention,
In a communication control node that employs a distributed module configuration, by detecting a faulty module quickly and controlling the status of the module, the service interruption time can be shortened compared to the past, and the service interruption when a failure occurs By setting the time to the failure recovery monitoring timer value, the maintenance person can customize the time.

[Brief description of the drawings]

FIG. 1 is a diagram showing a connection configuration of a service control node to which the present invention is applied.

FIG. 2 is a diagram showing a condition of function deployment for realizing failure recovery on each distributed module constituting a service control node.

FIG. 3 is a diagram showing exchanges at the time of failure occurrence / recovery between functional elements arranged on each distributed module.

FIG. 4 is a diagram illustrating a specific failure recovery procedure when a failure occurs in a function distribution module.

[Explanation of symbols]

 101 Transmission Node 201 Load Balancing Module Group 202 Load Balancing Module Group 203 Function Balancing Module Group 204 Centralized Control Module 300 Service Control Node 400 Maintenance Terminal 500 Fault Detection / Restart Control Mechanism 501 Fault / Restart Progress Notification Mechanism 502 Other Module State Management Mechanism 503 All module status monitoring / control mechanism 504 Failure recovery monitoring mechanism 600 Affected module 700 Other module

────────────────────────────────────────────────── ─── Continuing on the front page (72) Inventor Osamu Osamu 5-7-1 Shiba, Minato-ku, Tokyo NEC Corporation (72) Inventor Yasuo Okamoto 5-7-1 Shiba, Minato-ku, Tokyo NEC (56) References JP-A-7-168778 (JP, A) JP-A-6-209367 (JP, A) JP-A-6-37783 (JP, A) JP-A-63-214842 (JP, A A) (58) Field surveyed (Int. Cl. ⁷ , DB name) H04L 29/14 H04L 12/24 H04L 12/26 G06F 13/00 353

Claims (1)

(57) [Claims] 1. A distributed module group for load distribution and function distribution , a centralized control module for supervising and managing these modules,
A failure recovery method for a communication control node configured by a distributed node system having a maintenance terminal, wherein when the centralized control module detects a failure that has occurred in a distributed module, the failure of the affected module is coordinated with another module. stop the communication service access immediately, as well as restoring the communication service access immediately if resuming process has been completed by the Kakasawa module, the function distribution module in the distributed nodes system
Equipped with a database system for communication services
When managing updated data, the centralized control module is
When a failure occurs, monitor the time of failure recovery, and
If the failure does not recover within the time set from the terminal
The backup function between the modules specified in advance is
A failure recovery method for a distributed node system , wherein the processing is taken over to a backup side based on a relationship.