Industrial Consultancy & Sponsored Research (IC&SR) , IIT Madras
Method for Consensus Prioritization of Regression Test- Cases Supported by Parallel Execution Windows
Technology Category/Market
Computer Science, Software Testing
Applications – Software testing solutions in enterprise setting, BFSI, IT, Govt. & Public sector
Market – Software testing market is valued at USD 40 B in 2021 and is anticipated to record a CAGR of more than 6% between 2022 – 2030.
Targeted Industries
Watermarking IP design,
Electronic circuits
Image Gallery
Problem Statement
- Presently, software regression testing techniques employ single heuristic factor for testing purposes. This may not be desirable as diversity and sophistication varies in different software.
- The current software testing techniques lack efficiency and delay in testing which may lead to non-compliance in software release timelines.
- Existing technology fail to account for resources expended in technique and there is a need to provide test cases according to business and resource requirements.
- The present invention use two approaches: Hybrid (priority-aware / on-the-fly), and Consensus (priority blind / post-individual) regression test prioritization.
Technology
- Testing initiated; Heuristics employed to compute individual scores based on software tested.
- Individual heuristics compute & generate individual priority scores from which a weighted score is assigned; test case prioritization performed in accordance.
- A cost cognizant metric EPL is used to quantify the effectiveness of the prioritization, when execution is driven by size-varying test parallelization windows of unequal load distribution.
- Distance between two prioritizations (with and without ties) to measure the quality of the final consensus prioritization.
Key Features/Value Proposition
- The weighted score assignment and hybridization function is not performed by other software testing technique.
- The employment of parallelization windows helps in division of labor and efficient utilization of computer resources.
- The present prioritization method outperformed existing techniques by showcasing an effectiveness of 55.22%. (Fig. 2).
Questions about this Technology?
Contact for Licensing
Research lab
Prof. Rupesh Nasre
Department of Computer Science & Engineering
Intellectual Property
- IN 386511
- PCT/IN2022/050354
- IITM IDF Ref. 2106
Technology Readiness Level
TRL 3/4
Early-stage validation has been carried.
The method is evaluated on 20 open-source subjects including source code, 69,305 test-cases, and with parallelization support of up to 40 logical CPUs.