초록 ▼

Example methods, apparatus and articles of manufacture to characterize applications are disclosed. A disclosed example method includes collecting resource utilization trace data from the two or more applications simultaneously running on one or more computational devices, determining an intrinsic di

Example methods, apparatus and articles of manufacture to characterize applications are disclosed. A disclosed example method includes collecting resource utilization trace data from the two or more applications simultaneously running on one or more computational devices, determining an intrinsic dimensionality of the collected trace data, the intrinsic dimensionality representing a number of predominate features that substantially characterize the trace data, and characterizing each application's workload based on the determined intrinsic dimensionality.

대표청구항 ▼

1. A method to characterize two or more applications, the method comprising: collecting resource utilization trace data from the two or more applications simultaneously running on one or more computational devices;determining an intrinsic dimensionality of the collected trace data, the intrinsic dim

1. A method to characterize two or more applications, the method comprising: collecting resource utilization trace data from the two or more applications simultaneously running on one or more computational devices;determining an intrinsic dimensionality of the collected trace data, the intrinsic dimensionality representing a number of predominant features that substantially characterize the collected trace data;characterizing each application's workload based on the determined intrinsic dimensionality, wherein each application's workload comprises CPU utilization, memory usage, and device input/output usage;detrending the resource utilization trace data for one of the two or more applications to remove at least one of a plurality of characteristics of the corresponding application's workload; andconstructing a trace for the one of the two or more applications based on remaining characteristics of the application's workload. 2. A method as defined in claim 1, further comprising applying principal component analysis to the trace data to determine the predominant features that characterize each application's workload. 3. A method as defined in claim 2, further comprising performing dimensionality analysis on the trace data to determine which of the predominant features are to be determined for the applications. 4. A method as defined in claim 1, further comprising planning capacity usage of the one or more computational devices using the characterization of each application's workload. 5. A method as defined in claim 1, further comprising allocating the two or more applications to the one or more computational devices using the characterization of each application's workload. 6. A method as defined in claim 1, wherein the resource utilization trace data comprises at least one of a memory utilization, a disk usage, a network bandwidth usage or a device input/output usage. 7. A method as defined in claim 1, further comprising generating a synthetic workload to characterize a first application's workload. 8. A method as defined in claim 1, wherein the predominant features comprise at least one of spikiness, noisiness, or periodicity. 9. A method as defined in claim 7, wherein generating the synthetic workload comprises at least one of amplifying or suppressing a feature of the first application's workload. 10. An apparatus to characterize two or more applications, the apparatus comprising: a memory to store collected resource utilization trace data from the two or more applications simultaneously running on one or more computational devices; anda processor to determine an intrinsic dimensionality of the collected resource utilization trace data, the intrinsic dimensionality representing a number of predominant features that substantially characterize the collected resource utilization trace data, the processor to determine, for each application's workload, a value corresponding to an extent a first one of the predominant features is present in that application's workload based on the determined intrinsic dimensionality, and to compare the value to a threshold to characterize each application's workload, wherein each application's workload comprises CPU utilization, memory usage, and device input/output usage. 11. An apparatus as defined in claim 10, wherein the processor is to apply principal component analysis to the trace data to determine the predominant features that characterize each application's workload. 12. An apparatus as defined in claim 11, wherein the processor is to perform dimensionality analysis on the trace data to determine which of the one or more predominant features are to be determined for the applications. 13. An apparatus as defined in claim 10, wherein the processor is to plan capacity usage of the one or more computational devices using the characterization of each application's workload. 14. An apparatus as defined in claim 10, wherein the processor is to allocate the two or more applications to the one or more computational devices using the characterization of each application's workload. 15. An apparatus as defined in claim 10, wherein the resource utilization trace data comprises at least one of a memory utilization, a disk usage, a network bandwidth usage or a device input/output usage. 16. An apparatus as defined in claim 10, wherein the processor is to generate a synthetic workload to characterize a first application's workload. 17. A tangible machine readable storage device, excluding propagating signals, comprising machine-accessible instructions that, when executed, cause a machine to at least: collect resource utilization trace data from two or more applications simultaneously running on one or more computational devices;determine an intrinsic dimensionality of the collected resource utilization trace data, the intrinsic dimensionality representing a number of predominant features that substantially characterize the collected resource utilization trace data; anddetermine, for a workload of at least one of the applications, a value corresponding to an extent at least one of the predominant features is present in that workload based on the determined intrinsic dimensionality; andcompare the value to a threshold to characterize the workload, wherein the workload comprises CPU utilization, memory usage, and device input/output usage. 18. A storage device as defined in claim 17, wherein the machine-accessible instructions, when executed, cause the machine to apply principal component analysis to the trace data to determine the predominant features that characterize each application's workload. 19. A storage device as defined in claim 17, wherein the machine-accessible instructions, when executed, cause the machine to perform dimensionality analysis on the trace data to determine which of the predominant features are to be determined for the applications. 20. A storage device as defined in claim 17, wherein the machine-accessible instructions, when executed, cause the machine to plan capacity usage for the one or more computational devices using the characterization of each application's workload. 21. A storage device as defined in claim 17, wherein the machine-accessible instructions, when executed, cause the machine to allocate the two or more applications to the one or more computational devices using the characterization of each application's workload. 22. A storage device as defined in claim 17, wherein the machine-accessible instructions, when executed, cause the machine to generate a synthetic workload to characterize a first application's workload.