NETS: SMALL: Collaborative Research: Protocol Stacks Design and Evolution: The Role of Layering and Modularity

网络协议栈通常遵循多层分层架构。 什么决定了所需的层数，或者每层的协议数量？ 应用程序、服务和用户期望的空间（在顶层）以及基本功能的空间（在底层）都在不断发展--这种动态环境如何影响分层协议栈的组织？ 什么决定了协议栈在变化中的可演化性？ 可演化性与其他重要的系统性质，如鲁棒性（处理意外变化的能力），最优性和模块性有什么关系？ 这些都是一些高层次的问题，这个研究项目focusedon. Intellectual优点：这项研究旨在重新审视协议设计的基本原理和调查的作用和几个中心概念之间的依赖关系：分层，模块化，复杂性，鲁棒性，最优性。 然而，本研究的中心主题是进化--网络架构并不是设计用于静态环境中的。应用程序、服务、用户期望、通信和计算技术以及基础经济学都处于不断变化的状态。而不是依赖于一个单一的建模框架，这项研究开始与两个初步的模型，被称为地层和词汇。 这两种模式有一些共同的特点，但也有很大的不同和互补性。 它们都捕捉到了分层设计过程的时变特性，旨在从一组底层（也是动态的）基本功能支持一组动态的应用程序。Stratum比Lexis更受限制，因为它首先考虑可用构建块的特定顺序，其次考虑现有应用程序的“死亡率”的内生模型。 另一方面，Lexis更一般，更抽象，并捕获如何从时变字母表中分层构建时变正则表达式集，尽可能多地重用更简单的正则表达式。Stratum和Lexis模型是对影响协议可进化性和可持续性的因素进行基本理解的初始步骤。 有了这个基础，就可以进一步深入了解特定领域，并为未来网络协议的设计提供有针对性的指导。目前有四个大型的NSF资助的未来互联网架构（FIA）项目，追求一个“干净的石板”的设计方法，为未来的互联网，更强大，安全，进化，本研究的一个目标是既通知和学习的努力，四个国际汽联项目。更广泛的影响，这个项目还包括释放两个可扩展的模拟器（对应于Stratum和Lexis），这将允许学生和教师实验网络架构如何在动态环境中以及在优化目标和约束下发展。 目前还没有这样的工具。此外，该项目在其两个方面都有重要的跨学科组成部分。输入和输出。 具体来说，这项研究背后的大部分前期工作（投入）都起源于经济学、物理学和进化生物学。 相反，项目的结果（产出）可能与其他学科相关，如软件工程，智能制造和组织科学。

Network protocol stacks typically follow a multi-layer hierarchical architecture. What determines the required number of layers, or the number of protocols at each layer? The space of applications, services and user expectations (at the top layer) and the space of elementary functions (at the bottom layer) are constantly evolving -- how does this dynamic environment affect the organization of layered protocol stacks? What determines the evolvability of a protocol stack in the presence of changes? How does evolvability relate to other important system-wide properties such as robustness (the ability to deal with unexpected change), optimality and modularity? These are some of the high-level questions that this research project focuses on. Intellectual Merit: This research seeks to reexamine the fundamentals of protocols design and investigate the role of and dependencies between several central concepts: layering, modularity, complexity, robustness, optimality. The central theme in this research, however, is evolution -- networking architectures are not designed to work in a static environment. Applications, services, user expectations, communication and computing technologies, as well as the underlying economics, are all in a constant state of flux. Rather than rely on a single modeling framework, this research starts with two preliminary models, referred to as Stratum and Lexis. The two models share some features but are also significantly different and complementary. They both capture the time-varying character of a layered design process aiming to support a dynamic set of applications from an underlying set of (also dynamic) elementary functions. Stratum is more constrained than Lexis as it considers, first, a specific ordering for the available building blocks, and second, an endogenous model for the ``death rate'' of existing applications. Lexis, on the other hand, is more general and more abstract, and captures how a time-varying set of regular expressions can be constructed hierarchically from a time-varying alphabet, re-using simpler regular expressions as much as possible.The Stratum and Lexis models are initial steps towards developing a base understanding of what factors influence the evolvability and sustainability of protocols. With this base in place, further domain-specific insight can be developed and yield targeted guidelines for the design of future network protocols.Broader Impact: There are currently four large NSF-funded Future Internet Architecture (FIA) projects that pursue a ``clean-slate'' design approach for a future Internet that is more robust, secure, evolvable, etc. A goal of this research is to both inform and learn from the efforts of the four FIA projects.The broader impact of this project also includes the release of two extensible simulators (corresponding to Stratum and Lexis) that will allow students and instructors to experiment with how network architectures evolve in dynamic environments and under optimization objectives and constraints. No such tools are available today.Additionally, this project has a significant inter-disciplinary component in both its ?inputs and outputs.? Specifically, much of the prior work behind this research (inputs) has originated in economics, physics and evolutionary biology. Conversely, the project's results (outputs) are likely to be relevant to other disciplines such as software engineering, smart manufacturing and the science of organizations.