Data processing: database and file management or data structures – Database design – Data structure types
Reexamination Certificate
1999-12-28
2002-08-20
Coby, Frantz (Department: 2171)
Data processing: database and file management or data structures
Database design
Data structure types
C707S793000, C707S793000, C709S201000, C709S203000
Reexamination Certificate
active
06438553
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates generally to a distributed job integrated management system which manages a plurality of distributed job management systems and a method thereof. More particularly, the present invention relates to an integrated management system and a method thereof which do not require dividing network resources for different distributed job management systems and in which network resources are efficiently utilized to realize high speed data processing.
BACKGROUND OF THE INVENTION
FIG. 12
shows a structure of a conventional job processing system in which two different distributed job management systems are operating on a network. Network resources
23
are composed of a plurality of host systems, for example, a host system A and a host system B. Also, network resources
24
are composed of a plurality of host systems, for example, a host system C and a host system D. Each of the host systems comprise two CPU's. There are two different distributed job management systems utilizing these host systems. The network resources
23
having the host system A and the host system B are managed or administered by the distributed job management system
21
. The network resources
24
having the host system C and the host system D are managed or administered by the distributed job management system
22
.
A distributed job management system is a system which allocates a CPU_A
1
to a job “a”, when the job “a” is to be performed, and allocates a CPU_A
2
to a job “b”, when the job “b” is to be performed, thereby having both jobs processed in parallel. That is, the job “a” and the job “b” are processed by using the CPU_A
1
and the CPU_A
2
, respectively. In this case, when processing of the job “b”, for example, is finished earlier, the CPU_A
2
becomes an idle condition. Therefore, it becomes possible to allocate the CPU_A
2
to a new job “c” to perform it. Thereby, the distributed job management system enables a whole processing to be performed quickly.
Concretely, the distributed job management system itself monitors the rate of operation of each CPU of the host systems which the distributed job management system administrates, and grasps the condition of execution of each job. Also, the distributed job management system grasps which job is allocated to which CPU, the condition of processing, the condition of execution of jobs, and the like. That is, the distributed job management system performs administration, monitoring, processing and analysis on the resources. By performing the analysis, it is possible for the distributed job management system itself to intelligently manage or administrate the CPU's of the host systems, i.e., the resources, such that, for example, when any CPU is idle or when processing load of any CPU is light, the distributed job management system makes the processing load of that CPU full.
An example of a distributed job management system is described in Japanese patent laid-open publication No. 2-224169. The Japanese patent laid-open publication No. 2-224169 discloses an invention of a processing distributing system having a plurality of computer systems connected to a network. In the processing distributing system, a task management control portion divides a job injected to the system into tasks. The tasks are transferred to other computer systems to perform processing thereof according to parameters of the divided jobs, and processing result of the tasks processed by the other computer systems are integrated with the processing result of the tasks processed by its own computer system. By distributing the load within the network, resources in the network can be effectively utilized, processing load of each resource can be decreased, and waiting time of each job for processing can be shortened. Thereby, a turn around time (TAT) can be reduced.
However, in any of the above-mentioned conventional technique, the distributed job management systems respectively perform mutually different processing and are independent of each other. Therefore, it is impossible to administrate and monitor all the jobs which are inputted from one distributed job management system, from the other distributed job management system. Also, resources such as hardware and the like are completely separated every distributed job management system. As a result, even when, for example, the CPU_C
2
of the distributed job management system
22
is not so fully utilized but the distributed job management system
21
is fully utilizing the CPU_A
1
, the CPU_A
2
, the CPU_B
1
and the CPU_B
2
and there is always no idle time in these CPUs, the distributed job management system
21
can not utilize the CPU_C
2
, because the CPU's are completely separated every system.
As mentioned above, in case there are a plurality of different distributed job management systems on a network, since the distributed job management systems respectively use different job management method, it is impossible to administrate and monitor all the jobs given by one distributed job management system, from the other distributed job management system. It is necessary to statically separate the network resources every distributed job management system. Therefore, in the conventional technique, there is a disadvantage that jobs introduced or injected from every distributed job management systems can not share all the network resources.
Also, since the jobs injected can not share all the network resources, even when load of jobs injected from one distributed job management system is low, the turnaround time (TAT) of jobs introduced from any other distributed job management system is not shortened.
SUMMARY OF THE INVENTION
Therefore, it is an object of the present invention to obviate the disadvantages of a conventional job management system.
It is another object of the present invention to provide a distributed job integrated management system in which it is not necessary to separate network resources every different distributed job management system and, thereby, the network resources can be effectively utilized.
It is still another object of the present invention to provide a distributed job integrated management system in which jobs can be efficiently processed at high speed.
According to an aspect of the present invention, there is provided a distributed job integrated management system used in a networked computer system which comprises a plurality of host computer systems coupled to a common network, and a plurality of heterogeneous distributed job management systems each processes inputted jobs by using the host computer systems, wherein one of the distributed job management systems is used as a master distributed job management system and the rest of the plurality of distributed job management systems are used as slave distributed job management systems, and the master distributed job management system determines allocation of resources of the host computer systems to the master distributed job management system and to the slave distributed job management systems, based on the condition of usage of the resources of the host computer systems.
In this case, it is preferable that inputted jobs are allocated to the slave distributed job management systems or to the master distributed job management system via the master distributed job management system.
It is also preferable that the master distributed job management system has a master job queue for jobs to be processed by the master distributed job management system and a slave job queue for jobs to be processed by the slave distributed job management system, and inputted jobs are queued into the master job queue or into the slave job queue via the master distributed job management system.
It is advantageous that jobs are inputted from each of the plurality of host computer systems.
It is also advantageous that jobs can be inputted from each of the host computer systems with a command designating whether the jobs are to be processed by the slave distributed job management system or by the master distributed job managemen
LandOfFree
Distributed job integrated management system and method does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Distributed job integrated management system and method, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Distributed job integrated management system and method will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2896680