“…In [15] we have shown that our mechanism works equally well for H.263 videoconference traffic and excels, again, in comparison to the equivalent bandwidth approach. The accuracy of our video traffic models, for both MPEG-4 and H.263 traffic, helps alleviate the disadvantage that can be found in a number of efficient CAC mechanisms in the literature (e.g., [20,22,25]), which is the frequent use of bandwidth adapta- tion; as pointed out in [24], frequent bandwidth switching among different levels can consume a significant amount of resources, due to increased network signaling overhead. Regarding [25], which introduced a revenue-based CAC, it should also be pointed out that the linear change which the authors proposed in the number of guard channels (depending on the change in revenue) is not an efficient choice for video traffic, due to its bursty nature.…”