JPH04295937A - Stand-by redundancy type system - Google Patents

Abstract

PURPOSE:To offer the stand-by redundancy type system which is easily designated and altered while the delay time for the takeover of processed to a stand-by element in case of main system element trouble is made shorter than before. CONSTITUTION:Each system element is composed of a main processing part 10 which performs a process characteristic to the element, a main system state monitoring procedure part 11 which monitors the main system state and switches its element to the main system immediately when trouble occurs to the main system element, a main system determining and arbitrating procedure part 12 which performs arbitration so that more than a specific number of main systems are not generated, etc.

Description

[Detailed description of the invention]

[Object of the invention]

0002

[Industrial Field of Application] The present invention provides system elements such as programs and resources that are multiplexed for the purpose of improving reliability, and that these system elements have a group of main system elements that they normally process, and a group of main system elements that these system elements normally process. It relates to a system consisting of a group of standby elements that take over processing in the event of a failure.

0003

2. Description of the Related Art In recent years, as requirements in various fields of computer systems have diversified and technology has become more sophisticated, systems have tended to become larger and more complex. For this reason, improving system reliability has become important.

[0004] As methods for improving system reliability, there are two methods: a parallel multiprocessing method in which the same process is executed in parallel by multiple system elements;
There is a standby redundant processing method in which processing is executed by a main system element (main system element), and a standby system element continues processing when the main system element currently in execution fails.

[0005] The parallel multiprocessing method does not require operations such as starting up standby elements in the event of a failure, and processing can be continued without stopping the system. However, even under normal conditions, the same process is processed in parallel by a plurality of multiplexed elements, so the processing load on each computer is large.

On the other hand, in the standby redundant processing method, operations must be performed to maintain the same internal state of the main element and the standby element, but under normal conditions, the standby element does not perform actual processing. Therefore, the processing load on each computer is small, and computer resources can be used effectively.

In a standby redundant system employing such a standby redundant processing method, a plurality of standby system elements may be provided to improve reliability. In this way, in a standby redundant system where multiple standby elements exist, if the main element fails, which standby element will become the main element and take over the processing is determined in advance among the standby element groups. A method of assigning priorities and a method of arbitrating which standby element becomes the main element among a group of standby elements have been proposed.

[0008]

[Problems to be Solved by the Invention] However, as mentioned above, there are standby redundant systems in which priorities are assigned in advance between standby element groups, and which standby element among the standby element groups becomes the main element. In a standby redundant system that performs arbitration work, when the main system fails, it is necessary to check the priorities of the standby system elements and perform arbitration work to determine a new main system. For this reason, there is a problem in that the delay time until a new main system element is determined and processing is restarted after a main system element failure becomes long. Also, the multiplicity of such system elements,
A system must be constructed taking into account priorities, etc.
There is a problem in that the design of the system becomes complicated and it is difficult to change it.

The present invention has been made in response to such conventional circumstances, and it is possible to reduce the delay time for handing over processing to a standby system element in the event of a main system element failure compared to the conventional system, and also to reduce system design. The present invention aims to provide a standby redundant system that can be easily modified.

[Configuration of the invention]

[0011]

[Means for Solving the Problems] That is, the standby redundant system of the present invention allows at least some of the system elements constituting the system to be used as a main unit that executes a given task while exchanging messages during normal times. In a standby redundant system that is multiplexed as a standby system element group and a standby system element group that switches to the main system and continues processing when the main system element fails, each standby system element updates the state of the corresponding main system element. The present invention is characterized by comprising a monitoring means that monitors the main system element and immediately switches its own element to the main system element when an abnormality is detected in the main system element.

0012

[Operation] In the standby redundant system of the present invention having the above configuration,
Each standby system element monitors the status of the corresponding main system element, and when an abnormality in the main system element is detected, it is equipped with a monitoring means that immediately switches its own element to the main system element. Sometimes, the corresponding standby element switches to the main system immediately without making any adjustments such as keeping the number of main elements constant.

[0013] Therefore, the delay time for handing over processing to the standby system element in the event of a failure of the main system element can be minimized, and the system designer does not have to be aware of the multiplicity of each system element. , changes, etc. can be easily made.

Furthermore, after switching its own element to the main element, the number of said main elements in the system element group is determined, and if a predetermined number or more of standby elements have been switched to the main element, It is also possible to always maintain a constant number of main system elements by providing each element group with an arbitration means that maintains the number of main system elements at a predetermined number by returning the main system element to a standby system element. be.

[0015]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, embodiments of the standby redundant system of the present invention will be described in detail with reference to the drawings.

FIG. 1 shows the configuration of a standby redundant system according to an embodiment of the present invention. In this embodiment, program modules are system elements.

In FIG. 2, 1a, 1b, . . . are processors as processing execution means in which system elements are registered, and these processors 1a, 1b, . . . are message exchange devices 2a, 2, respectively, as message exchange means.
b..., and these message exchange devices 2
a, 2b... are connected to the - column. These processors 1a, 1b... execute the functions of the system elements, and communication between the system elements is carried out by the message exchange device 2a.
, 2b...

Furthermore, in this embodiment, for reasons such as improving reliability, individual system elements are multiplexed and distributed at each site, as shown in FIG. 3, for example. In Figure 3, system element A (A1, A2, A
3) is registered in processors 1a, 1b, and 1c, but only system element A1 registered in processor 1a is the main system element that normally operates, and system element A1 registered in processors 1b and 1c is A2, A
3 is a standby system element that does not normally perform actual processing but stands by to become the main system when the processor 1a fails. If there are two or more standby elements like this, one of the standby elements becomes the main element.

System element B (B1, B2) is registered with processors 1a and 1b, and processor 1b
The system element B1 of the processor 1a is the main system, and the system element B2 of the processor 1a is the standby system. Similarly, system element C (C
1, C2, C3) are processors 1a, 1b, 1
c and is registered in system element C of processor 1c.
1 is the main system, system element C2 of processors 1a and 1b
, C3 are standby systems.

As shown in FIG. 1, each system element has a main processing unit 10 that executes processing specific to each system element, and a main system state that is monitored, and when a failure occurs in the main system element, the main system element is Main system status monitoring procedure section 1 that immediately switches the main system to the main system
1, and a main system determination arbitration procedure unit 12 for arbitrating between elements that have newly become main systems so that more than a predetermined number of main systems do not occur.
It is composed of etc.

[0021] The main system status monitoring procedure section 11 monitors the main system status by, for example, determining that the main system has failed when the main system periodically transmits an "alive" signal and the standby system no longer receives the message. This can be achieved by a method in which the standby system inquires about the status of the main system and when there is no response, it is determined that the main system has failed.

The operation of the main system status monitoring procedure section 11 will be explained below with reference to FIG.

As shown in FIG. 4, the main system state monitoring procedure unit 11 of the standby system elements (for example, A2, A3) monitors the main system state (step 401), and if the main system is operating normally, Otherwise, it will continue monitoring without taking any action.

When a failure in the main system is detected (step 402), other standby system elements (A3 for A2,
For A3, the own element is made the main system regardless of the state of A2) (step 403).

After this, a plurality of standby elements (A2, A3
) may become main system elements at the same time, so in order to return the number of main system elements to the original one, the main system determination arbitration procedure unit 12 in its own element and in other elements is activated (step 40
4) End the process.

Next, the operation of the master decision arbitration procedure section 12 will be explained with reference to FIG.

When the main system determination arbitration procedure section 12 of the standby system element or the standby system element that has already been switched to the main system (for example, A2) receives the activation message from the main system status monitoring procedure section 11 (step 501), It is determined whether the request message is requested from within the own element (A2) or from outside (A3) (step 502).

[0028] When activated from within its own element (A2), the element (A3) which has already become the main system before its own element
) exists (step 503), and if the element that became the main system first does not exist, the own element (A2
) broadcasts the fact that it has become the main system to the main system determination arbitration procedure unit 12 of the other multiplexing element (A3) (step 504).
, it continues to operate as the main system.

[0029] Also, the element that is the main system first (A3
) exists, the own element (A2) is returned to the standby element (step 505).

On the other hand, in step 502 described above, the relevant (
The main system decision arbitration procedure unit 12 (of A2) is the external element (A3
), set the number of main system elements +1 to count the elements that have already become main system (Step 5
06).

[0031] To determine whether or not there is an element that has become the main system first, the number of main system elements for counting the elements that have already become the main system is incremented by 1 in step 506. This is done by reference.

As described above, in the standby redundant system of this embodiment, when the standby elements (for example, A2, A3) detect a failure in the main element (for example, A1), the Because it immediately switches to the main element without performing any processing, the standby element (for example, A2, A3)
The delay time for taking over processing from a failed main system element (for example, A1) is reduced, and the processing interruption time can be shortened compared to the conventional method. Additionally, when multiplexing system elements to improve system reliability, system designers need to be aware of how many layers are being used, as information is exchanged using broadcasting between the elements that are intended to become the main system. There is no need to worry about system design and changes, making it easy to design and change the system. Furthermore, by performing arbitration in the main system determination arbitration procedure unit 12 after switching the own element to the main system element, a constant number of main system elements is always maintained without prolonging the processing interruption time. can do.

[0033] In the above example, an example in which the number of main system elements is one has been described, but the present invention can be similarly applied to a case where the number of main system elements is two or more. When the number of main system elements is two, even if one of the main system elements fails, processing will continue without any interruption, and one of the standby system elements will become the new main system and will continue to be the main system. 2 system elements
maintained as one.

[0034]

As explained above, according to the standby redundant system of the present invention, the delay time for handing over processing to the standby element when the main element fails can be reduced compared to the conventional system. System design and changes can be easily performed.

[Brief explanation of drawings]

FIG. 1 is a diagram showing the configuration of system elements.

FIG. 2 is a diagram showing the configuration of an embodiment of the present invention.

FIG. 3 is a diagram showing a state in which standby redundant elements are distributed among processors.

FIG. 4 is a diagram showing the operation of the main system state monitoring procedure section.

FIG. 5 is a diagram showing the operation of the main system decision arbitration procedure section.

[Explanation of symbols]

10　Main processing section 11 Main system status monitoring procedure department 12 Main Decision Conciliation Procedures Department

Claims (2)

[Claims] Claim 1: At least some of the system elements constituting the system are connected to a main system element group that executes a given task while exchanging messages during normal times, and to a main system element group when a failure occurs in the main system element. In a standby redundant system that is multiplexed as a group of standby elements that continue switching processing, each standby element monitors the state of the corresponding main element, and when an abnormality in the main element is detected, A standby redundant system characterized by comprising a monitoring means for immediately switching its own element to a main element. 2. The standby redundant system according to claim 1, wherein each standby element determines the number of main elements in the system element group after switching its own element to the main element. , an arbitration means is provided for maintaining the number of main system elements at a predetermined number by returning one of the main system elements to a standby system element when a predetermined number or more of standby system elements have been switched to the main system element. A standby redundant system featuring: