(In this SmartBits, Jawahar Sabapathy outlines “Should I know the architecture to test?“. The video is at the end of this blog)
I think it is a given. You cannot validate if you do not how it is constructed and what it is doing. One has to validate the normal functionality, in the case of vehicle, when the road is good and when it is real bumpy and bad. Hence I need to know under what operating conditions that the vehicle can actually be performing correctly and how it is built. For example if the vehicle has tubeless tyres, we may have to test it differently.
The same is applicable to software construction, so when we deliver on a high availability or scaling, the mechanism that is used must be known so that we can actually simulate those minimum- maximum outlier conditions and see that it is working in those conditions or not. At the least we should know and document it. So a deeper understanding of architecture in deployment is probably more necessary for today’s validation teams.