STAMP - Software Testing AMPlification

Project start: 
Thursday, 1 December, 2016
Project end: 
Saturday, 30 November, 2019
STAMP provides new tools for automatic software testing, focusing on amplifying existing test assets. These tools automatically generate and run variants of existing unit test cases and test configurations, to detect regressions . STAMP will deliver these tools as an Automatic Test Amplification a cloud service for DevOps.

 

Who is the service/solution designed for?: 

Software editors and integrators are the main users for STAMP. The test management solution market is segmented into the following three categories: 

  • Full-suite vendors.

  • Independent pure-play vendors.

  • Open-source tools.

  • The DevOps community in general is the first user of STAMP.

How will your solution/service benefit the end-user? 

The current state of practice for testing in DevOps relies on extensive manual effort to produce test cases that hardly cover all the code.

STAMP’s automatic amplification solutions will increase code coverage and the diversity of tested behavior. This will reduce the number of regression bugs in production, the cost of updates and enhance trust continuous delivery. This potential for radical innovations in DevOps relies on one key concept: deliver test amplification software services, which can increase test automation at multiple development steps that span from early development to operations in production and which can be integrated in various DevOps toolchains.

STAMP addresses test automation at unit level, configuration level and production stage, leveraging the human knowledge and the manual effort invested in the production of test assets (unit tests, APIs, manually defined test configurations, etc.) to increase their value through automatic amplification.

How can the solution/service help you become more efficient, more secure, faster or cost-effective?: 

STAMP test amplification techniques aim to reduce the accumulation of technical debt due to prohibitive refactoring costs, by significantly lowering the risk of opportunistic refactoring. The impact will be a decrease in associated long term maintenance costs and will :

1) Increase the diversity of execution paths covered by 40%.

2) Decrease by 20% the number of tests which fail once but not again if run several times.

3) Increase by 20% the number of lines of product code, which are executed for each second of time spent running tests.

4) Increase by 40% the number of unique invocation traces between services in a global perspective.

5) Increase by 30% the number of valid bugs detected during testing which are specific to the generated configurations.

6) Reduce by 30% the time on configuring and deploying products for testing purpose.

7) Reduce the size of log files by an order of magnitude, keeping all essential information.

8) Increase by 70% the number of crash replicating test cases.

9) Enhance existing test suites with 10% of production-level test cases.

10) three test amplification services integrated in two different toolchains.