Modeling for interest rate derivatives (such as bermudan swaptions) is said to have undergone significant changes since the crisis. Prior to the crisis, counterparty default risk was often ignored, allowing modelers to unify many curves that are now separated. Part of the story is in the Risk article Goldman and the OIS Gold Rush (paywall alert).
I am interested in articles or answers that characterize these changes, both in terms of what assumptions were changed and what that meant in terms of numerically modeling portfolios of interest rate derivatives.
