Currently, multicast-related studies can be classified into four specific classes: multicast forwarding state aggregability for network-layer multicast routing; content distribution network replica placement; application-level and router-assisted hierarchical reliable multicast schemes; and application-level multicast routing. In these fields, analytical work is needed to help understand theoretical upper bounds and performance.
In this paper, the authors present the calculation (from the analytical point of view) of blocking probabilities in multicast trees with dynamic membership. The paper is well organized. The outcome of the work is extremely clear, and in spite of the complexity of the topic, the authors provide a readable paper, thanks to the presence of clear notations and definitions, a precise evaluation of the context, a clear definition of the model and solution, and finally, a complete and detailed description of their algorithm. The authors make available a precise positioning in the existing literature, and also provide the reader with an exhaustive analysis of related works.
The paper is clear and well written, and represents a worthwhile read for those researchers interested in the analytical modeling of complex algorithms in the multicast network field. A simulation study would be useful in helping readers to better understand the applicability of the achieved results. Furthermore, in order to enlarge the number of user population models analyzed, a coupled analysis, based on analytical and simulative study, would help readers to envisage some of the dependencies between the achieved results and user models. Finally, some other remarks on real applications of the work (with respect to the four classes depicted above), as well as an investigation of the blocking analysis, in the case of unicast traffic patterns different from Poisson distributions (introducing self-similar processes, and the long range dependence of persistence phenomena), could also be added.