Which two options could you use to optimize the invalidation criteria and yet achieve the same goal to invalidate the multiple objects?