Collaborative Research: Frameworks for NS-3

这个项目正在为ns-3离散事件网络模拟器开发新的模拟框架，ns-3离散事件网络模拟器是一个免费的开源研究工具，专为计算机通信网络研究人员设计。由于其易用性、可重复性、可用性、可扩展性和易于软件开发，仿真仍然大量用于网络研究和教育。 然而，一个典型的模拟研究的工作流程需要用户的仔细注意的方法中的一些最佳实践，以确保结果是可信的和可重复的第三方，并不会引入错误。 这个项目正在为ns-3模拟器建立一个自动化框架，提供一个更有利于创建严格研究的环境。 自动化框架引导用户通过组成和一致性检查以及实验设置的定义来生成有效的网络仿真模型。该框架对大规模模拟研究进行控制，并将统计方法应用于输出数据分析和执行控制。该框架还收集重现实验和报告结果所需的详细信息。 该框架还允许用户指定广泛的仿真网络拓扑结构，应用程序流量和其他参数，允许以最小的努力在各种场景上进行实验。 最后，继续进行ns-3软件的维护、文档编制和分发活动。 ns-3作为开源软件发布，其增加的内容正在改善未来基于模拟的网络研究的方法和可信度。

This project is developing new simulation frameworks for the ns-3 discrete-event network simulator, a free and open-source research tool designed for use by computer communications network researchers. Simulation remains in heavy use for network research and education, because of its ease of use, reproducibility, availability, scalability, and ease of software development. However, the workflow of a typical simulation study requires the user's careful attention to a number of best practices in methodology to ensure that the results are credible and reproducible by third parties and that errors are not introduced. This project is building an automation framework for the ns-3 simulator that provides an environment that is more conducive to the creation of rigorous studies. The automation framework guides the user to produce valid network simulation models by composition and consistency checking, and definition of experimental settings. The framework exercises control over large-scale simulation studies, and applies statistical methods for output data analysis and execution control. The framework also collect details required to reproduce the experiment and report results. The framework also allows users to specify wide ranges of simulation network topologies, application traffic, and other parameters, allowing experiments to be conducted on a wide range of scenarios with minimal effort. Finally, maintaining the ns-3 software, documentation, and dissemination activities continue. The additions to ns-3, disseminated as open source software, are improving the methodology and credibility for future simulation-based network research.