Error detection/correction and fault detection/recovery – Data processing system error or fault handling – Reliability and availability
Reexamination Certificate
1998-07-02
2001-04-17
Wiley, David A. (Department: 2781)
Error detection/correction and fault detection/recovery
Data processing system error or fault handling
Reliability and availability
Reexamination Certificate
active
06219801
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates to a system by which work processes are efficiently carried out by a plurality of processing units in a hot standby mode or a load sharing mode.
BACKGROUND OF THE INVENTION
In a system for which reliability is required, the system is duplicated. That is, two processing units are prepared, and a first processing unit is used as a current system and a second processing unit is on standby for a trouble occurring in the first processing unit. The second processing unit is provided with programs and data needed for processes to promptly inherit works when a trouble occurs in the first processing unit. Such a mode is called a hot standby mode.
In addition, as to a transaction process, although one of the transactions can be carried out in a short time, a large number of transactions which simultaneously occur can not be processed by a single processing unit in realtime. Thus, there is a load sharing mode in which, for distribution of load, the transactions which have occurred are distributed among a plurality of processing units and processed. In this load sharing mode, in a case where a trouble occurs in a processing unit, transactions are not distributed to the processing unit having the trouble.
To form a system in the hot standby mode or the load sharing mode, a large number of programs are needed.
In addition, in recent years, due to decreasing of the price of the processing unit and the necessity of processing many works, operations in a range between an operation in which two or three large general-purpose computers are used and an operation in which many processing units are used are desired. For example, the number of cases where a plurality of processing units each of which is mounted in a frame and named a cluster are connected to each other is increasing (a plurality of CPUs may be included in a single cluster). With the spread of such a case where a plurality of clusters are connected to each other, it is desired that works are efficiently inherited when a trouble occurs in one of the clusters.
SUMMARY OF THE INVENTION
Accordingly, a general object of the present invention is to provide a novel and useful work inheriting system which solves the above problems.
A specific object of the present invention is to provide a work inheriting system, without a great number of programs, in which works can be inherited among a plurality of processing units processing works in the hot standby mode or the load sharing mode when a trouble has occurred.
To achieve the above objects of the present invention, the present invention is a work inheriting system applied to a system in which processes for works are carried out by a plurality of processing units in a hot standby mode, said work inheriting system comprising: storage means for storing a table indicating that each of said processing units for a work is a current system or a standby system; means for, when a trouble occurs in a processing unit which is the current system for a work, causing a processing unit which is the standby system for the work to inherit processes for the work in said processing unit having the trouble with reference to said table stored in said storage means.
According to such a system, in the system in which a work is carried out by a plurality of processing units in the hot standby mode, a processing unit which is the standby system inherits the work processes in the processing unit of the current system in which the trouble occurs, with reference to the table indicating that each of the processing units for a work is the current system or the standby system. Thus, an algorithm for identifying a processing unit which should inherit the work processes is simplified. The work can be efficiently inherited without need of a program having a large size.
In addition, to solve the above problem, the present invention is a work inheriting system applied to a system in which processes for works are carried out by a plurality of processing units in a load sharing mode, said work inheriting system comprising: storage means for storing a table indicating a rate at which each of said plurality of processing units shares processes of the works; means for, when a trouble occurs in a processing unit, causing another processing unit to inherit the work which is shared at a rate by said processing unit having the trouble with reference to said table stored in said storage means.
According to such a system, in the system in which the work is carried out by a plurality of processing units in the load sharing mode, another processing unit inherits the work which is shared at a rate by the processing unit having the trouble with reference to the table indicating a rate at which each of the processing units shares processes of the work. Thus, the inheriting processes are simplified. The work can be efficiently inherited without need of a program having a large size.
Further, to solve the above problem, the present invention is a work inheriting system applied to a system in which processes for a plurality of works are carried out by a plurality of processing units in a hot standby mode and a load sharing mode, said work inheriting system comprising: storage means for storing a table indicating whether works to be processed in the hot standby mode and works to be processed in the load sharing mode are assigned to said respective processing units and whether each of processing units to which the works to be processed in the hot standby mode are assigned is a current system or a standby system; and means for, when a trouble occurs in one of said plurality of processing units, causing a processing unit which is the standby system for a work to be processed in the hot standby mode to inherit processes for the work out of the works assigned to said processing unit having the trouble and another process unit to which a work to be processed in the load sharing process is assigned to inherit processes of the work, with reference to the table stored in said storage means.
According to such a system, in the system in which a plurality of works are carried out by a plurality of processing units in the hot standby mode and the load sharing mode, another processing unit inherits the work processes of the processing unit in which the trouble occurs with reference to the table indicating whether works to be processed in the hot standby mode and works to be processed in the load sharing mode are assigned to said respective processing units and whether each of processing units to which the works to be processed in the hot standby mode are assigned is a current system or a standby system. Thus, the inheriting processes are simplified. The work can be efficiently inherited without need of a program having a large size.
REFERENCES:
patent: 5526492 (1996-06-01), Ishida
patent: 5774642 (1998-06-01), Flon et al.
patent: 5796937 (1998-08-01), Kizuka
patent: 5802265 (1998-09-01), Bressoud et al.
patent: 55-66049 (1980-05-01), None
patent: 07044413 (1995-02-01), None
patent: 7-44413 (1995-02-01), None
patent: 07334468 (1995-12-01), None
patent: 7-334468 (1995-12-01), None
patent: 97/49034 (1997-12-01), None
Amemiya Toshihiro
Matsumoto Satoshi
Suzuki Yasuhiro
Yuge Toshimichi
Fujitsu Limited
Staas & Halsey , LLP
Wiley David A.
LandOfFree
Work inheriting system does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Work inheriting system, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Work inheriting system will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2537210