SGER: Scalable adversary resistant routing

提案编号：0617883PI：Baruch Awerbuch Institution：Johns Hopkins University标题：SGER：可扩展的抗敌手路由摘要本研究倡导基于信誉的路由的概念，作为一种在不影响可扩展性的情况下实现安全路由的方法。现有的任何工作都不能保证如此强大的性质。为了使这项研究成功，新的数学和算法思想是必要的。实现我们的研究目标是一种极高风险和需要多学科的方法来实现许多领域的突破，如应用网络、人工智能和理论计算机科学。这一建议基于作者的一系列初步结果，这些初步结果表明，即使在完全对抗性的环境下，也可以在网络路由环境中进行高效协作。这项工作与其他已发表的类似主题的工作的关键区别在于我们坚持扎实的数学基础，不会使用不合理的假设。这是因为未经证明的启发式算法在对抗环境中并不可行。另一方面，迫切需要为容忍对手的路由领域奠定基础。没有这样的基础，我们就不可能对互联网在面对对手攻击时正常运行的能力有信心。对我们的网络结构缺乏信心，使人对我们整个社会的稳定产生怀疑。

Proposal Number: 0617883PI: Baruch Awerbuch Institution: Johns Hopkins University Title: SGER: Scalable adversary-resistant routing AbstractThis research advocates a concept of reputation-based routing as a way to achieve secure routing without compromising scalability. None of the existing work guarantees such strong properties.In order for this research to succeed, novel mathematical and algorithmic ideas are necessary. Accomplishing our research objective is an extremely high-risk and required multi-disciplinary approach accomplishing breakthroughs in many fields, such as Applied Networking, Artificial Intelligence and Theoretical Computer Science.This proposal is based on a sequence of preliminary results of the author that show that efficient collaboration in context of weeding out adversarial actions in the network routing setting is possibleeven under completely adversarial circumstances.The key differentiator between this work and rest of the published work on similar topics is our insistence of solid mathematical foundations, that will not use unjustified assumptions. The reason for this is that unproven heuristics are not a viable approach in an adversarial setting.On the other hand, there is an acute need laying the foundation for the area of adversary-tolerant routing. Without such a foundation, we cannot possibly have confidence in ability of Internet to properly function in face of adversarial attacks. Lack of confidence in our cyber-structure casts a doubt on the stability of our society as a whole.