HOME未踏/セキュリティ・キャンプ2008年度下期 採択案件概要



2008年度下期 採択案件概要


 David J. Farber PM(Distinguished Career Professor of Computer Science and Public Policy Carnegie Mellon University)


 チーフクリエータ:中野 賢(University of California, Irvine Department of Computer Science)



 学校法人国際大学 (グローバル・コミュニケーション・センター)




 Development of an Evolutionary Agent-based Network Simulator




 Computer networks today are rapidly growing in size and complexity. A recent network potentially contains millions of nodes with different computing and networking capabilities that may dynamically leave and join such a network. A key challenge here is to design a network that is scalable, adaptive and robust, and that operates autonomously to minimize operational and maintenance costs. In this project, the creator develops a network simulator that can be used to design an agent-based autonomous network. The specific goal is to develop a scalable network simulator that can simulate 10,000 to 100,000 interacting agents on a large scale network. The creator also develops an easy-to-use graphical user interface for designing agent-based networks. The creator also develops built-in functions to simulate typical network conditions (e.g., typical network topologies and workload patterns) as well as a set of libraries for extending the functions. The simulator will be made publicly available on the Internet, and in the long run, it is expected that the simulator is used to design new network services and solutions.


 I have very mixed feelings toward this effort. While the idea of gent vased systems is very interesting (and has been since the 1970s in local environments) I have serious concerns about building a simulator. It is not that simulators are not good but I have a concern that the effort will not lead anyplace. It is just too easy to create a real environment on today’s networks using today’s large-scale distributed environments.
 That being said I would not veto this effort. tIt is well thought out and would contribute to this area but I still am very concerned what real contribution it will make as oppoded to real experiments.