So far in our performance test blogs, we have had a look at performance requirements and test infrastructure, we will now look into an interesting aspect of performance testing, which has often been miscued – performance test data. Interestingly in many of the engagements that we have got engaged on in the past we have repeatedly zoned in on poor data quality as a cause for poor application quality. It is just not the DEV, SIT, UAT, DEMO, TEG that gets impacted, it encompasses performance test environments as well. This increases the cost of testing as there is a need to perform a separate exercise for data provisioning. Project teams at times, with an intent to adhere to schedule, have not looked at meaningful data for performance tests and leveraged meaningless data.
Some of the strategies are creating new data or leveraging existing data like Cartesian product of meaningless data available like A1, A2,...for names etc., having improper data distribution, etc. Often teams think that data volume is the only constraint to be looked for performance, completely ignoring aspects for meaningful data like right person/banking account data, data distribution requirements, data privacy, compliance aspects and regulations. Meaningless data can have impact on database operations like search queries, etc. Heard it all and sounds familiar in your experience, please read our blog Don't care for Meaningful Data? Congratulations you bought a ticket for disaster!!!"