Access to good test data is critical to the success of any test function and testers spend a considerable amount of time preparing, managing & cleaning up this data. In an agile world it becomes even more imperative to be able to perform these test data activities in a faster, more automated manner.
For us as distributed agile teams, this was a big problem. We were struggling with creating bulk volume of data at short intervals for multiple teams distributed across a complex agile landscape, with different needs and prerequisites. The results of the tests setup by teams often exhibited erroneous results due to data being cross-used and contaminated. There was also a lot of duplication of test data as well as waste of effort creating data that could have been reused. We were manually creating data, thereby not reaching the needed coverage when data was required rapidly. We also needed to be able to provide test data to external partners, which added a lot of complexity in regards to communication, collaboration and simply sheer volume of data needed.
What we needed was a solution that would enable automated creation of data through multiple sources, that could be exposed to external partners as well as our internal teams. We needed to be able to seamlessly connect and extract the required test data in a way that was both performant, easy to use and allowed teams to reserve data to make sure their test results were unaffected by other tests. So we built the solution that fulfills all our needs!
Looking at where we are currently, we see an increase in velocity by 30% from when we started. We have happier teams,the test results have less noise and are able to release confidently into production.
In this session I will present the work we did around creating an efficient framework for test data management and how that has helped us prepare, manage and provision data across distributed agile teams, in an enterprise setting. We will cover both internal work and how we managed to allow external third party vendors and partners to interface into our TDM system through API’s, without losing performance or stability for our internal deliveries. You will hear about our learnings, current state and where we are planning to move in the future.