Abstract: To achieve consistent health service measures, a method and an apparatus combine various metrics from both internal sources and external sources that relate to the service under observation. This service health information is generated independently from specific provider applications and performance monitoring tool sets, thereby allowing shorter time-to-market for service management solutions. The output of the method may be in the form of a programmatic or scriptable interface to be used by high-level performance monitoring tools that are capable of reporting status of many disparate computer services. The performance monitoring tools may reside on different systems and architectures and may be supplied by different vendors. As a result, the programmatic or scriptable interfaces are designed to be generic and flexible.

Abstract: Apparatus and method for enhancing performance of a computer system. The invention may comprise apparatus and methods for deriving relationships between system variables and the performance of the computer system, generating a number of rules based on the derived relationships, and adjusting at least one of the system variables based on the generated number of rules to enhance the performance of the computer system. Preferably, data is acquired for the system variables and the performance of said computer system by gathering the data over time, and logging the gathered data, wherein the relationships are derived based on the logged data.

Abstract: A method for the automatic selection of computer system parameter values to improve system performance as workload on the system changes. Methods are disclosed for automatically making minor adjustments to computer system parameters and comparing long-term associated performance changes in order to set parameter values so as to obtain improved system performance. The time frame over which a change in a system parameter and the associated system performance change is averaged is adjustable, thus permitting evaluation over shorter or longer periods of times as deemed appropriate. An objective measure of system performance is defined prior to implementation of the present methods. For transaction based systems, average system response-time for a given demand on the system could be, for example, such a measure. For systems executing batch-type workloads, system throughput may be used as the indicator of how well the system is performing.

Abstract: To achieve out-of-the-box performance management solutions in modern, complex computer environments, a method and an apparatus provide automatic discovery of applications and services resident on a monitored computer system and autonomous configuration of available performance management tools. The method includes inventorying installed applications, performance management tools, and active services. The inventory information is written to an ASCII-format file, and is used to generate performance management tool configuration files. The performance management software is then restarted to engage the new configuration customized to the computing environment.

Abstract: A system and method for generating and sharing consistent application definitions amongst multiple performance and resource management tools. A common interface accepts requests from performance or resource management tools and returns requested process-to-application mapping data. A parameters file contains a common syntax for application definitions, and a module uses the parameters file to determine mapping from processes into applications. A performance monitoring tool for senses current system processes' states and provides measurement data to the mapping module.

Abstract: A method for determining appropriate adjustments of computer system parameter values in order to improve system performance. The method for this determination is based on historical performance data. Methods are disclosed which involve measuring current system performance and retrieving paired stored values of measured system performance and a selected system parameter. The decision as to whether or not to adjust the value of the selected system parameter is based on performing a trend analysis on the paired values. When such changes are made, subsequent measurement of system performance is performed and beneficial changes are implemented.

Abstract: A method for obtaining an estimate of response time for a typical transaction request in a computer system under varying load conditions. This estimate can be compared to a predefined value considered by the system administrator to be the level at which the response time becomes significantly degraded, thereby providing a measure of the operating condition of the computer system. Preliminary measurements include measurement of response time for the typical transaction request while the computer system is lightly loaded, CPU usage, and response times for more than one concurrent typical transaction request. During computer system operation only the number of concurrent transaction requests and the total CPU usage need to be obtained and measured. A primary advantage of the embodiments as described in the present patent document over prior methods is that the present methods can provide a more accurate approximation of actual response-time under varying loads while consuming less system resources to do so.

Abstract: A method for the automatic selection of computer system parameter values to improve system performance as workload on the system changes. Methods are disclosed for automatically making minor adjustments to computer system parameters and comparing long-term associated performance changes in order to set parameter values so as to obtain improved system performance. The time frame over which a change in a system parameter and the associated system performance change is averaged is adjustable, thus permitting evaluation over shorter or longer periods of times as deemed appropriate. An objective measure of system performance is defined prior to implementation of the present methods. For transaction based systems, average system response-time for a given demand on the system could be, for example, such a measure. For systems executing batch-type workloads, system throughput may be used as the indicator of how well the system is performing.

Abstract: A method for the automatic selection of computer system parameter values to improve system performance as workload on the system changes. Methods are disclosed for automatically making minor adjustments to computer system parameters and comparing long-term associated performance changes in order to set parameter values so as to obtain improved system performance. The time frame over which a change in a system parameter and the associated system performance change is averaged is adjustable, thus permitting evaluation over shorter or longer periods of times as deemed appropriate. An objective measure of system performance is defined prior to implementation of the present methods. For transaction based systems, average system response-time for a given demand on the system could be, for example, such a measure. For systems executing batch-type workloads, system throughput may be used as the indicator of how well the system is performing.

Abstract: Apparatus and method for enhancing performance of a computer system. The invention may comprise apparatus and methods for deriving relationships between system variables and the performance of the computer system, generating a number of rules based on the derived relationships, and adjusting at least one of the system variables based on the generated number of rules to enhance the performance of the computer system. Preferably, data is acquired for the system variables and the performance of said computer system by gathering the data over time, and logging the gathered data, wherein the relationships are derived based on the logged data.

Abstract: To achieve consistent health service measures, a method and an apparatus combine various metrics from both internal sources and external sources that relate to the service under observation. This service health information is generated independently from specific provider applications and performance monitoring tool sets, thereby allowing shorter time-to-market for service management solutions. The output of the method may be in the form of a programmatic or scriptable interface to be used by high-level performance monitoring tools that are capable of reporting status of many disparate computer services. The performance monitoring tools may reside on different systems and architectures and may be supplied by different vendors. As a result, the programmatic or scriptable interfaces are designed to be generic and flexible.

Abstract: A method for determining appropriate adjustments of computer system parameter values in order to improve system performance. The method for this determination is based on historical performance data. Methods are disclosed which involve measuring current system performance and retrieving paired stored values of measured system performance and a selected system parameter. The decision as to whether or not to adjust the value of the selected system parameter is based on performing a trend analysis on the paired values. When such changes are made, subsequent measurement of system performance is performed and beneficial changes are implemented.

Abstract: A method for obtaining an estimate of response time for a typical transaction request in a computer system under varying load conditions. This estimate can be compared to a predefined value considered by the system administrator to be the level at which the response time becomes significantly degraded, thereby providing a measure of the operating condition of the computer system. Preliminary measurements include measurement of response time for the typical transaction request while the computer system is lightly loaded, CPU usage, and response times for more than one concurrent typical transaction request. During computer system operation only the number of concurrent transaction requests and the total CPU usage need to be obtained and measured. A primary advantage of the embodiments as described in the present patent document over prior methods is that the present methods can provide a more accurate approximation of actual response-time under varying loads while consuming less system resources to do so.