----------------------- REVIEW 1 --------------------- PAPER: 13 TITLE: Placement du cœur d'un réseau mobile autonome AUTHORS: Jad Oueis, Vania Conan, Damien Lavaux, Razvan Stanica and Fabrice Valois ----------- Overall evaluation ----------- The paper “ Placement du Coeur d’un réseau mobile autonome” tackles the placement problem of a Local EPC in a network of multiple base stations with no backhaul connectivity. The authors propose a new centrality metric “the flow centrality” and show by comparison with other metrics from the literature that the proposed one permits the total amount of traffic that the local EPC is capable of receiving from all BSs. Overall the paper is well written, well-motivated and easy to follow. Some few comments have to be addressed for the camera ready version: - The parameter n in Section 4 is undefined, I suppose that is the total number of nodes in the network that was defined in Section 2 by | V|. - The authors mentioned in Section 4.1 that the computation time is in the order of minutes, does this time corresponds to one node or to the whole network? The authors have to clarify this point because if it corresponds to one node, the overall computation time will be in the order of hours and hence not time-efficient. ----------------------- REVIEW 2 --------------------- PAPER: 13 TITLE: Placement du cœur d'un réseau mobile autonome AUTHORS: Jad Oueis, Vania Conan, Damien Lavaux, Razvan Stanica and Fabrice Valois ----------- Overall evaluation ----------- This paper proposes a new centrality metric, named "flow centrality", defined as the node in the network graph which provides the maximum flow from all other nodes in the network. The flow centrality metric in this paper is used to choose the best placement of a Local EPC in a network composed of interconnectec base stations (BS). The paper is very well-written and easy to follow. The authors use CPLEX to solve a linear problem and find the node with largest flow centrality. The metric is then compared to other centrality mestrics, to show that the results differ and flow centrality is not the same as computing other metrics. The results show it. Yet, the paper should probably compare the defined problem, finding the node in the network with the maximum flow from all others, with a regular max flow problem. While the basic max flow problem is defined for one source and one sink, it can be easily extended to a multi-source multi-sink max flow problem. In the scenario analyzes in the paper, couldn´t the problem be viewed as N multi-source single-sink max flow problems, than choosing the one with the largest flow? How this solution compares to computing the max flow centrality? ----------------------- REVIEW 3 --------------------- PAPER: 13 TITLE: Placement du cœur d'un réseau mobile autonome AUTHORS: Jad Oueis, Vania Conan, Damien Lavaux, Razvan Stanica and Fabrice Valois ----------- Overall evaluation ----------- The article was well written and well presented. The problem modeling has been described including a state of the art on the subject. The authors presented a very interesting proposal (flow centrality) with promising results. Some mistypings: - Reference [Bor05]: 2015 -> 2005. - Some spaces after the comma.