Collaborative Research: CT-T: DoS Prevention in Shared Channels

提案编号：0524269标题：共享通道中的DoS预防PI：Carl GunterMethods用于分析和防止拒绝服务(DoS)威胁对于设计健壮的Internet协议具有基本价值。人们已经做了很多工作来开发针对协议特定DoS威胁的实用解决方案，但缺乏现实的理论模型来研究DoS，也缺乏设计具有DoS弹性的协议的广泛范例。该项目开发了基于“共享通道模型”的理论模型，该模型描述了攻击者和合法发送者如何共享攻击目标的网络带宽。它利用这一模型来设计基于范式的对策，在该范式中，被对手利用的协议工作负载中的不对称性被系统地转换为受信任方的优势。具体的项目目标包括开发(1)获得DoS弹性的通用技术，可用于修改现有协议或创建新的协议；(2)自动化协议的DoS分析的方法，以减少在协议部署之前从理论上确认实际可用性属性和发现意外攻击所需的工作量；以及(3)基于现有代数技术和新的概率技术的完整性、机密性和可用性的统一模型。

Proposal Number: 0524269Title: DoS Prevention in Shared ChannelsPI: Carl GunterMethods for analyzing and preventing Denial of Service (DoS) threats are of fundamental value for designing robust Internet protocols. Much work has been done to develop pragmatic solutions to protocol-specific DoS threats, but there is a lack of realistic theoretical models for studying DoS and of broad paradigms for designing DoS-resilient protocols. This project develops theoretical models based on a "shared channel model" which describes how adversaries and valid senders share the network bandwidth of attack targets. It exploits this model to design counter-measures based on a paradigm in which asymmetries in protocol workloads that are exploited by adversaries are systematically converted to the advantage of trusted parties. Specific project goals include developing (1) general techniques for obtaining DoS-resilience that can be used to adapt existing protocols or create new ones; (2) ways to automate DoS analysis of protocols to reduce the effort required to confirm practical availability properties theoretically and find unexpected attacks before protocols are deployed; and (3) a unified model of integrity, confidentiality, and availability based on both existing algebraic techniques and new probabilistic techniques.