Evaluating reliability of a software module using development life cycle

• US 9,372,748 B2
• Filed: 07/15/2015
• Issued: 06/21/2016
• Est. Priority Date: 12/05/2012
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Bird, John J., Ulmer, Terry R.
• Primary Examiner(s): GUYTON, PHILIP A
• Application Number:
  US14/800,190
• Publication Number:
  US 20150331737A1
• Time in Patent Office: 342 Days
• US Class Current: 1/1
• CPC Class Codes: G06F 11/008 Reliability or availability analysis
• CPC Class Codes: G06F 11/0721 within a central processing unit [CPU]
• CPC Class Codes: G06F 11/079 Root cause analysis, i.e. error or fault diagnosis in a hardware test environment G06F11/22; in a software test environment G06F11/36
• CPC Class Codes: G06F 11/3495 for systems
• CPC Class Codes: G06F 8/70 Software maintenance or management

Abstract

Reliability of one or more software modules is projected according to a current state in a development life cycle of the software modules and any of various additional indicators. Preferably, a data processing support provider separate from the service-providing enterprise maintains historical field support data concerning significant field defect events with respect to various resources, and uses this data for projecting reliability of the resources. Preferably, software module reliability projections are used to support an analysis of risk of degradation of a service specified in a service requirements specification when provided by a configuration of data processing resources specified in a configuration specification.

Claims

1. A computer program product embodied as a plurality of computer-executable instructions recorded on non-transitory computer readable media, said instructions, when executed by at least one computer system, causing the system to perform a method for evaluating reliability of a first software module, the method comprising:

accessing data recording a state in a software module development life cycle of said first software module;

accessing recorded historical reliability data with respect to said first software module;

using said state in a software module development life cycle of said first software module and said historical reliability data with respect to said first software module to automatically determine a reliability of said first software module corresponding to at least one of (a) a current state in a software module development life cycle of said first software module;

or (b) a future state in a software module development life cycle of said first software module; and

providing results of said using said state in a software module development life cycle of said first software module and said historical reliability data with respect to said first software module to automatically determine a reliability of said first software module to a user;

wherein said computer program product further includes an degradation of service risk evaluation function which uses said results of said using said state in a software module development life cycle of said first software module and said historical reliability data with respect to said first software module to automatically determine a reliability of said first software module to evaluate degradation of service risk for a service provided by a set of data processing resources, said set of data processing resources containing said first software module, and provides results of evaluating degradation of service risk to said user.

2. The computer program product of claim 1,

wherein said historical reliability data comprises data concerning failure events occurring in production installations of at least one of:

(a) said first software module, and (b) at least one previous version of said first software module.

3. The computer program product of claim 2,

wherein said historical reliability data further comprises a respective cause for each of at least some of said plurality of failure events;

wherein said state in a software module development life cycle of said first software module includes data concerning any fixes applied to said first software module to correct at least one cause of failure; and

wherein using said state in a software module development life cycle of said first software module and said historical reliability data with respect to said first software module to automatically determine a reliability of said first software module comprises filtering each failure event of said plurality of failure events for which a respective cause has been identified and has been corrected by said any fixes applied to said first software module, and computing a reliability of said first software module from the filtered results.

4. The computer program product of claim 2,

wherein using said state in a software module development life cycle of said first software module and said historical reliability data with respect to said first software module to automatically determine a reliability of said first software module comprises using said data concerning failure events occurring in production installations of at least one of:

(a) said first software module, and (b) at least one previous version of said first software module, to automatically derive, using said data processing system, at least one coefficient of a mathematical model projected software reliability as a function of time by deriving a best fit of said data concerning failure events occurring in production installations to a mathematical function.

5. The computer program product of claim 1,

wherein using said state in a software module development life cycle of said first software module and said historical reliability data with respect to said first software module to automatically determine a reliability of said first software module comprises:

(a) using said historical reliability data to derive at least one coefficient of a mathematical model projecting software reliability as a function of time in a software module development life cycle; and

(b) predicting a reliability of said first software module from said mathematical model and said state in a software module development life cycle of said first software module.

6. The computer program product of claim 1,

wherein said historical reliability data includes a plurality of event records, each record corresponding to a respective instance of software unavailability, and indicating a respective length of unavailability as a result of the corresponding event.

7. A computer program product embodied as a plurality of computer-executable instructions recorded on non-transitory computer readable media, said instructions, when executed by at least one computer system, causing the system to perform a method for evaluating reliability of a first software module, the method comprising:

accessing data recording a state in a software module development life cycle of said first software module;

accessing recorded historical reliability data with respect to said first software module;

using said state in a software module development life cycle of said first software module and said historical reliability data with respect to said first software module to automatically determine a reliability of said first software module corresponding to at least one of (a) a current state in a software module development life cycle of said first software module;

or (b) a future state in a software module development life cycle of said first software module; and

providing results of said using said state in a software module development life cycle of said first software module and said historical reliability data with respect to said first software module to automatically determine a reliability of said first software module to a user;

wherein said historical reliability data comprises data concerning failure events occurring in production installations of at least one of;

(a) said first software module, and (b) at least one previous version of said first software module.

8. The computer program product of claim 7, wherein said computer program product further includes an degradation of service risk evaluation function which uses said results of said using said state in a software module development life cycle of said first software module and said historical reliability data with respect to said first software module to automatically determine a reliability of said first software module to evaluate degradation of service risk for a service provided by a set of data processing resources, said set of data processing resources containing said first software module, and provides results of evaluating degradation of service risk to said user.

9. The computer program product of claim 7,

wherein said historical reliability data further comprises a respective cause for each of at least some of said plurality of failure events.

10. The computer program product of claim 9,

wherein said state in a software module development life cycle of said first software module includes data concerning any fixes applied to said first software module to correct at least one cause of failure;

and

wherein using said state in a software module development life cycle of said first software module and said historical reliability data with respect to said first software module to automatically determine a reliability of said first software module comprises filtering each failure event of said plurality of failure events for which a respective cause has been identified and has been corrected by said any fixes applied to said first software module, and computing a reliability of said first software module from the filtered results.

11. The computer program product of claim 7,

wherein using said state in a software module development life cycle of said first software module and said historical reliability data with respect to said first software module to automatically determine a reliability of said first software module comprises:

(a) using said historical reliability data to derive at least one coefficient of a mathematical model projecting software reliability as a function of time in a software module development life cycle; and

(b) predicting a reliability of said first software module from said mathematical model and said state in a software module development life cycle of said first software module.

12. The computer program product of claim 7,

wherein said historical reliability data includes a plurality of event records, each record corresponding to a respective instance of software unavailability, and indicating a respective length of unavailability as a result of the corresponding event.

13. A computer program product embodied as a plurality of computer-executable instructions recorded on non-transitory computer readable media, said instructions, when executed by at least one computer system, causing the system to perform a method for evaluating reliability of a first software module, the method comprising:

accessing data recording a state in a software module development life cycle of said first software module;

accessing at least one of;

(a) recorded data of development activity with respect to said first software module, and (b) recorded field event data with respect to said first software module, to automatically derive at least one coefficient of a mathematical model projecting software reliability as a function of time in a software module development life cycle of said first software module;

using said state in a software module development life cycle of said first software module and said mathematical model projecting software reliability as a function of time in a software module development life cycle of said first software module to automatically determine a reliability of said first software module corresponding to at least one of (a) a current state in a software module development life cycle of said first software module;

or (b) a future state in a software module development life cycle of said first software module; and

providing results of said using said state in a software module development life cycle of said first software module and said mathematical model projecting software reliability as a function of time in a software module development life cycle of said first software module to automatically determine a reliability of said first software module to a user;

wherein using said data arising from at least one of;

(a) development activity with respect to said first software module, and (b) field event data with respect to said first software module, to automatically derive, using said data processing system, at least one coefficient of a mathematical model comprises at least one of;

deriving a best fit of field event data to a mathematical function; and

using at least one of a measure of software quality during development of said first software module and a measure of size of said first software module to device said at least one coefficient.

14. The computer program product of claim 13, wherein said computer program product further includes a degradation of service risk evaluation function which uses said results of said using said state in a software module development life cycle of said first software module and said mathematical model projecting software reliability as a function of time in a software module development life cycle of said first software module to automatically determine a reliability of said first software module to evaluate degradation of service risk for a service provided by a set of data processing resources, said set of data processing resources containing said first software module, and provides results of evaluating degradation of service risk to said user.

15. The computer program product of claim 13,

wherein using said state in a software module development life cycle of said first software module and said mathematical model projected software reliability as a function of time in a software module development life cycle of said first software module to automatically determine, with said data processing system, a reliability of said first software module corresponding to at least one of (a) a current state in a software module development life cycle of said first software module;

or (b) a future state in a software module development life cycle of said first software module, comprises automatically determining a reliability of said first software module as a weighted average of reliability determined using (a) said state in a software module development life cycle of said first software module and said mathematical model, and (b) field event data with respect to said first software module.

16. The computer program product of claim 15,

wherein said weighted average is adjusted over time to diminish the weight given reliability determined using said mathematical model with respect to the weight given reliability determined using said field event data.