Error detection/correction and fault detection/recovery – Data processing system error or fault handling – Reliability and availability
Reexamination Certificate
1999-02-23
2002-10-22
Beausoleil, Robert (Department: 2184)
Error detection/correction and fault detection/recovery
Data processing system error or fault handling
Reliability and availability
C713S100000, C702S186000, C707S793000
Reexamination Certificate
active
06470464
ABSTRACT:
The present invention is also related to “Method and System for Graphically Displaying Trend and Range Data for a Variety of Systems” as described in a pending patent application Ser. No. 09/088,525 filed Jun. 2, 1998 by Randal Bertram et al. This patent is sometimes referred to herein as the Performance Display Patent.
A concurrently-filed-patent application (RAL9-99-00) entitled “System and Method for Monitoring and Analyzing Computer System Performance and Making Recommendations for Improving it”, by Randal Bertram and Gregory McKnight, two of the inventors of the present patent, may also be relevant to the subject matter of this patent. This patent is sometimes referred to as the System Recommendation Patent.
A concurrently-filed patent application (RAL9-99-010) entitled “System and Method for Identifying Latent Computer System Bottlenecks and for Making Recommendations for Improving Computer System Performance”, by Randal Bertram and Gregory McKnight, two of the inventors of the present patent, may also be relevant to the present invention. This patent is sometimes referred to as the Latent Bottlenecks Patent.
The Performance Data Patent, the Performance Display Patent, the System Recommendation Patent and the Latent Bottleneck Patent are all assigned to the assignee of the present invention and the specification and drawings of each patent are specifically incorporated herein by reference.
1. Field of Invention
The present invention relates to capacity management in a computer system such as a network or server and, more particularly, to a method and system for analyzing the past performance of components of a computer system and projecting the results into the future, applying rules to the results to identify bottlenecks which have not yet occurred but are likely to occur and to make recommendations for ameliorating the bottlenecks.
2. Background Art
Managing a computer system which includes a plurality of devices such as networks or servers is of special interest to data processing (or information technology) personnel. The computer systems typically include a plurality of diverse devices such as memory, disks, local area network (LAN) adapters and central processing units (CPUs) which interact in various interrelated ways when a variety of data processing applications are used in the computer system. As the systems get larger and more complex, these interactions become hard to define, model or predict the relationships between the devices, and hence the capacity of the system becomes difficult to manage. These systems are quite expensive to install and changes to the system involve a significant investment, so, while an investment is desirable which will improve the performance of the computer system, some investments in improvements to the computer system would not be worthwhile since the performance of the overall system would not improve.
Frequently, the performance of the computer system or network is less than it could be because only or more of the components is not appropriate for application loading of the computer system (or a network or server). It is desirable to know what changes to the computer system would be worthwhile in improving the capacity of the computer system and making those changes while avoiding changes which would not have a significant benefit to the performance of the computer system. The System Recommendations Patent describes the inventors' solution to this problem. However, the System Recommendation Patent does not make any projections as to how the computer system is likely to operate in the future.
One way to address the proper components for the loading of the system is to provide a model of the load and simulate the system to provide an optimum (or desirable) combination of elements. While there are numerous simulation techniques, they all rely on approximations of the loading and the components, and, as the systems become larger and more complex and the loading becomes more complex, the simulations are approximations whose accuracy and reliability is subject to significant doubt.
Another approach to predicting performance of a complex computer system involves active monitors, or adding a known load to an existing system and measuring the resulting output and effect of the load. This requires that the system be available for experimentation and that the added load operate in a known manner, both of which are assumptions that may work in some instances but not in others. For example, BlueCurve Dynameasure by BlueCurve, Inc. Intentionally induces an artificial workload to determine performance characteristics of a computer system. Such an active monitor is disruptive to the network (in that it interferes, at least to some extent, with the ongoing work of the computer system and the artificial load on the network may not accurately reflect the real world actual work of the computer system, either now or in the future.
Another way to manage the capacity is described in the Performance Management Patent and involves sampling of indicators of system activity. These indicators can be displayed as described in the Performance Display Patent, if desired. In any event, the data must be interpreted by a professional who has experience in looking at the results and interpreting the data to make recommendations. Unfortunately, these experts are in demand and not enough exist, so it is unlikely that a network expert would be available to analyze the results and to make suggestions for improvement at any given time and having the local expert is an inefficient use of his time and expertise.
Accordingly, the prior art systems for capacity management are limited and have undesirable limitations and disadvantages.
SUMMARY OF THE INVENTION
The present invention overcomes the limitations and disadvantages of the prior art systems by providing an improved capacity management system which is easy to use and which provides an indication of the bottleneck(s) in the system, in an ordered list, along with recommendations on how to improve the computer system, based on the use of passive monitors. The system projects the performance into the future using conventional algorithms and uses rules to determine potential bottlenecks which have not occurred but are projected as likely to occur in the further. Based on the predictions, the present invention projects when a bottleneck is likely to occur and makes recommendations as to how to ameliorate the system.
The present invention also has the advantage that analysis of the performance data of the system allows one to extend the trends into the future, with a confidence range based on statistical analyses, and allows for the inclusion of appropriate warnings about the reliability of the projections.
The improved capacity management system uses data which is typically available from hardware and software and uses software tools which are typically available. Thus, it is not necessary to find some obscure data on the computer systems or to add additional overhead (such as additional hardware or new software) to the computer system in order to obtain the necessary data to make recommendations on improving the computer system.
The present invention also allows use of statistical techniques for projections and for the use of confidence ranges, if desired, using conventional statistical methods. As the statistical tools become better defined and/or subject to further analysis, the substitution of new techniques for the previous projection and analysis techniques can be accomplished.
The present system has the advantage that it does not require a simulation program and it does not require that the user find or create unusual sets of data (like artificial loads of an active monitoring system like BlueCurve Dynameasure) which other prior art capacity management systems may require for analysis.
The present system avoids the need for consulting with an expert in the field of analyzing computer system performance to look at the various data which is available on the performance of the computer system and make judgment
Bertram Randal Lee
Hunter Krauss Frederick Scott
McKnight Gregory J.
Beausoleil Robert
Munoz-Bustamante Carlos
Ziemer Rita A
LandOfFree
System and method for predicting computer system performance... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with System and method for predicting computer system performance..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and System and method for predicting computer system performance... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2925187