|
Abstract : |
This short note is written as a submission to the DARPA Workshop on Validation of Large Scale Network Simulation Models. The purpose of the workshop is to identify issues critical to the validation of network models, and also to develop potential strategies for the research community to address these issues. 1 Current network simulation models or tools: I have used simulations as a key component of my network research since I began working in this area about ten years ago. While some of the simulation work (e.g., on synchronized routing messages, and on Scalable Reliable Multicast) has been done on special-purpose simulators that I wrote to investigate particular issues, most of my simulations have been done on the NS Network Simulator and its predecessor simulators (NS-1, and before that `tcpsim', and before that an earlier version) developed by Steve McCanne and myself at LBL. The NS simulator [NS95] evolved from a network simulator developed by Steve McCanne in 1990, based on the REAL network simulator. Steve and I were the only users of this simulator for some years. I used the simulator for research on issues such as TCP dynamics, RED queue management, explicit congestion notification, and CBQ scheduling, and was responsible for the details of those modules in the simulator. The LBL Network Simulator (written in C) evolved into the NS-1 network simulator (written in C++ and Tcl, and made publically available in 1995), and then into NS-2 (whose non-beta release was in 1997). For the last three years, the development of NS has been joint work at ISI/USC, Xerox PARC, LBNL, Xerox PARC, and UC Berkeley through the VINT project, and NS has become widely used in the network research community. In addition to using NS in my own research, I also have an interest, as a member of the VINT project, with the development of NS for more general use in the network research community., |
