I think they also addressed the performance issue in addition to saying it is not documented: "...evolving away from the global backlog concept in order to provide better support for different concurrency models"
It is still not clear to me that they actually really understood the performance issue, or how big it was going to be.