Electrical computers and digital processing systems: multicomput – Computer network managing – Computer network monitoring
Reexamination Certificate
2000-05-05
2004-10-19
Harvey, Jack B. (Department: 2142)
Electrical computers and digital processing systems: multicomput
Computer network managing
Computer network monitoring
C709S223000, C709S203000, C345S215000, C345S215000, C345S215000, C719S316000, C719S317000
Reexamination Certificate
active
06807575
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention is concerned with a distributed processing system in which several objects are distributed and arranged on several computers and relates to the performance monitoring method especially.
There is an equipment that measures the operation status of the computer system as a time series data for each of processes as shown in, for example, Japanese patent Laid-open 7-295864, as an equipment that measures the operation status of several processes in connection with a case that several processes execute a process on a computer cooperatively.
As an example of manufacturing the equipment, as shown in NIKKEI byte exception extra number, 1998.11-8, No.184, pp192-193, it is known that Visual Quantify, etc. of Rational Software company which is a profiler that analyzes a bottleneck of the performance of the application program that is executed on a computer with a module level, a function level and a source level.
On the other hand, a method for executing performance measurement without conscious of a remote procedure call is known, in case that several processes mutually cooperate a process in the distributed processing system that consists of several computers. For example, in Japanese patent Laid-open No. 5-274185, by having a means for detecting issue of a remote procedure call, a means for detecting end of a remote procedure call, and a central performance measurement control means, the mechanism that does association of a process has been realized. This method does association of a calling side with a called side by notifying all of remote procedure call to a central performance measurement control unit.
It is expected that, in the future, a distribution of processing over several computers is advanced and a distributed object system having several computers and a communication function between objects each of which is a program that is distributed and arranged and is executed on these several computers widely spreads.
By using a technique disclosed to the Japanese patent Laid-open No. 7-295864, it is possible to detect a bottleneck of performance in a computer. But, it is difficult to detect the bottleneck of the performance of a distributed object system because there is not association of performance data collected in each computer by only applying this technique to each computer of a distributed object system.
And, it is possible to associate the a send object and a receive object in communication between individual objects in the distributed object system by using a technique disclosed in the Japanese patent Laid-open No. 5-274185. But in case after going by way of processing of several objects through communication between objects, a process is completed, it is difficult to pursue a flow of this series of processing because there is not a means for associating related communication between objects.
In the prior art, while the performance data of a point like each computer and each object can be collected, there is a problem that the performance monitoring for the whole distributed object system cannot be made.
SUMMARY OF THE INVENTION
The present invention is provided on the basis of fore mentioned reason, and object of the present invention is to provide a method of the performance monitoring for the whole distributed processing system and a distributed processing system executing the method.
And, object of the present invention is to provide a method of the performance monitoring of the distributed processing system that can narrow down the bottleneck of performance.
A method for monitoring performance of distributed processing system in the present invention is to collect as performance data at least one of message receiving time from other program entity and message sending time to other program entity for program entity that is measurement target by monitoring means arranged in each of computers, to put the performance data collected by each monitoring means together and calculate processing time of each program entities consisting of a process flow, and to display the results on a display device.
And, a method for monitoring performance of distributed processing system in the present invention is (a) to measure processing time of each object consisting of a process flow, (b) to measure latency that after each object issues a request to other object until gets the result, for object group executed by a computer in which the relevant object is arranged, when processing time measured in the step (a) exceeds predetermined threshold of processing time of the relevant object, and (c) to execute the step (a) for processing flow having a process of the relevant object as the head, when processing time measured in the step (b) exceeds predetermined threshold of processing time of the relevant object.
REFERENCES:
patent: 5958009 (1999-09-01), Friedrich et al.
patent: 6584504 (2003-06-01), Choe
patent: 5-274185 (1993-10-01), None
patent: 7-295864 (1995-11-01), None
Nikkei, 1998.11-8, No. 184, pp. 192-193.
Akatsu Masaharu
Akifuji Shunsuke
Emaru Hironori
Kojima Gou
Nishikawa Norifumi
Harvey Jack B.
Hitachi , Ltd.
Le Hien C.
Mattingly Stanger & Malur, P.C.
LandOfFree
Performance monitoring method in a distributed processing... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Performance monitoring method in a distributed processing..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Performance monitoring method in a distributed processing... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3307514