State based adaptive compression of performance traces
AssigneeCA, Inc.
InventorsSreenivas Gukal, Sanjay Vyas
Life of this patent
Estimating the term…
Abstract
A method by a computer includes generating polylines based on a sequence of values contained in performance trace data from a software source on a host machine node. Segments of the polylines are generated. Separate ones of the segments are categorized based on patterns of the polylines of the separate ones of the segments. For the separate ones of the segments, compressed data is generated that approximates values contained in the performance trace data corresponding to the polylines of the segment while regulating a level of fidelity of the approximations based on the categorization of the segment. The compressed data is stored in a log repository.
Claims20
A method by a computer comprising:
- generating polylines based on a sequence of values contained in performance trace data from a software source on a host machine node;
- generating segments of the polylines;
- categorizing separate ones of the segments based on patterns of the polylines of the separate ones of the segments;
- for the separate ones of the segments, generating compressed data that approximates values contained in the performance trace data corresponding to the polylines of the segment while regulating a level of fidelity of the approximations based on the categorization of the segment;
- and
- storing the compressed data in a log repository.
A computer program product comprising:
- a non-transitory computer readable storage medium having computer readable program code embodied therewith, the computer readable program code comprising:
- computer readable program code to generate polylines based on a sequence of values contained in performance trace data from a software source on a host machine node;
- computer readable program code to generate segments of the polylines;
- computer readable program code to categorize separate ones of the segments based on patterns of the polylines of the separate ones of the segments;
- computer readable program code to, for the separate ones of the segments, generate compressed data that approximates values contained in the performance trace data corresponding to the polylines of the segment while regulating a level of fidelity of the approximations based on the categorization of the segment;
- and
- computer readable program code to store the compressed data in a log repository.
Dependencies are inferred from the claim text.