When I started my software journey, which is not very long
time ago, releases used to be anywhere from an year to 2 years. Anything less
than this is considered stressful. Fast forward to 2015 or 2016, we are
delivering releases every few weeks, with zero downtime for customers, all in
cloud with a switch.
I still have 20,000 tests that I need to execute all of them
and only way we are surviving is by distributing. I want to choose what tests
to run and run only 100 tests from the 20,000 tests that I have. Certainly
possible, all we need to do is to induce artificial intelligence into our
testing system.
In the current approach, QA teams runs the whole sets of
test suites or frameworks in order to find out bug(s). If we see this on a
bigger picture, there are cases where QA Cycle take more amount of time than
the development cycle. So, what can be done to fix this? How about Introducing
Artificial Intelligence (AI) to the Software Testing domain, where we can
reduce the time gap between the development and testing cycles and can release
the quality product into the market. Let see, how we can build a AI enabled
testing framework.
Every test is tagged with a file system, a feature, a sub
system. Whenever a file is changed, various triggers gets fired and different
permutations are done to choose the right tests to run. This system will be tested
and more intelligence added, just like any AI system. Once this is up, we are
then ready for daily releases, how cool will that be.
No comments:
Post a Comment