When software (the Code) is properly developed it has adequate performance potential. Consequently it will always adequately perform as long as all other conditions are fulfilled. Of course application performance can become substandard nonetheless because of insufficient capacities or defects in the Implementation.
Insufficient performance potential of the code, entails by far the biggest threat that can materialize in a redesign, which can take much time ruining the time to value and cause great damage to a business. If the Performance potential of the code is insufficient your application will not perform no matter how much hardware capacity you provide. Understanding the impact of this is important:
Software development is not responsible for performance problems caused by
- insufficient capacity of any resource
- any defect in the implementation
So their job is done when they have delivered their code and assessed that its performance potential is OK.
They have delivered code with guaranteed performance. That does not mean that performance is guaranteed when the code is deployed. We can still experience performance problems but that is not the responsibility of Software Development but of Operations. It is an important merit of the mBrace Methodology and tooling that it, in contrast with LST, enables us to split the risk mitigating test process in stages that can be independently assigned to the responsible parties.