In the parlance of software testing and verification, a green build is a software build that passes tests on all reference devices. A green build is typically determined by a centralized test-scheduler. The centralized test-scheduler has a database of parameters, e.g., build-artifacts, build-branches, etc., corresponding to each device. The centralized scheduler uses the database to efficiently schedule tests. Centralized scheduling is computationally intensive, and maintenance of the database is a significant burden.
Per the techniques of this disclosure, devices in a pool collaboratively pick a new build to test. The first device to start within a given scheduling interval picks a build, and the remaining devices pick the same build. The devices independently test the selected build. The first device to finish testing, either due to pass or fail, picks another build. The remaining devices follow the newly picked build. The process continues until the devices converge upon a green build. The distributed manner of test scheduling, as described herein, enables efficient determination of the green build.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Yim, Keun Soo, "Distributed and Collaborative Test Scheduling to Determine a Green Build", Technical Disclosure Commons, (September 20, 2018)