Method and apparatus for detecting outliers in a set of runs of software applications

Patent metadata

Field	Value
Publication number	US-12436870-B2
Application number	US-202217988937-A
Country	US
Kind code	B2
Filing date	Nov 17, 2022
Priority date	Nov 25, 2021
Publication date	Oct 7, 2025
Grant date	Oct 7, 2025

Abstract

A method for detecting outlier behavior in a set of executions of one or several applications on an information processing device, implemented by a computer and comprising steps of
triggering (S 1 ) said set of executions in collaboration with a profiling tool in order to collect, for each execution, at least one time series of measurement points assigning, for each measurement point, a value to a measured parameter; automatically formatting (S 2 ) the time series obtained for said set, by adjusting, for each time series, its length, its values, and its number of measurement points; calculating (S 3 ) a metric between two time series among the time series collected for said set of executions; detecting (S 4 ) an outlier based on said distance.

First claim

What is claimed is:

1. A method for detecting outlier behavior in a set of executions of one or several applications on an information processing device, implemented by a computer, the method comprising:
triggering said set of executions in collaboration with a profiling tool in order to collect, for each execution, at least one time series of measurement points assigning, for each measurement point, a value to a measured parameter;
automatically formatting the time series obtained for said set, by adjusting by normalization, for each time series, its length, its values, and its number of measurement points;
said adjusting of its length comprising projecting each measurement point of said time series toward a reference interval, said adjusting of its values comprising dividing each value of said numerical series by a total quantity corresponding to all values for said time series, and said adjusting of its number of measurement points comprises interpolating a set of additional measurement points so that the number of measurement points of said time series is equal to the number of measurement points of a longer time series of said set;
calculating a metric between two time series among the time series collected for said set of executions;
detecting an outlier based on a comparison of said metric with a threshold, said outlier being detected when said metric is greater than said threshold;
excluding, for one of said applications, executions of said set for which an outlier has been detected; and
determining a normal behavior of said application from the remaining executions.

2. The method according to claim 1 , wherein formatting is performed in pairs of time series within said set.

3. The method according to claim 1 , further comprising providing information corresponding to said outlier to a user via a human-machine interface.

4. The method according to claim 1 , further comprising automatically determining runtime optimization parameters for an application, from the remaining executions of said application, said automatically determining comprising repeating executions of said applications with different sets of executing parameters and evaluating performances of said application for said set of executing parameters by considering the normal behavior of said application.

5. The method according to claim 1 , further comprising selecting an acceleration module for an application, from the executions of the same application, excluding the executions for which an outlier was detected.

6. The method according to claim 1 , wherein said total quantity is an approximation of the integral of said time series with respect to time.

7. The method according to claim 1 , wherein said metric is calculated on the cumulative sums of the values of said time series.

8. A non-transitory computer-readable storage medium comprising instructions which, when executed by a computer, cause said computer to perform the method according to claim 1 .

Assignees

• Bull Sas
• Commisariat A Lenergie Atomique Et Aux Energies Alternatives

Inventors

• Vincent Lionel
• Pham Trong-Ton
• Mimouni Salim

Classifications

• G06F11/3452

  Performance evaluation by statistical analysis · CPC title

• G06F11/3419

  by assessing time · CPC title

• G06F2201/865

  Monitoring of software · CPC title

• G06F11/302

  where the computing system component is a software system · CPC title

• G06F11/3616 Primary

  using software metrics · CPC title