Burn-in testing is the process the components of a system undergo before they are put into service (and, often, before the system has been completely assembled with those components).
The objective is to identify those particular components that would fail as a result of so-called infant mortality, that is, during the initial, high-failure part of the bathtub curve relating to the reliability of the components. If the burn-in period is long enough, it can be reliably assumed that, once completed, the tested system is almost totally free of any additional initial failures.