I haven't read that paper but it seems like it's fixing a different problem of Byzantine fault tolerance. Most consensus systems that are internal for an organization don't have the Byzantine issue so it simplifies the problem.