Mechanisms for Secure and Robust Agent-Based Distributed Computing

安全、鲁棒的基于代理的分布式计算机制

• 批准号: 9813703
• 负责人: Anand Tripathi
• 金额: $ 20.11万
• 依托单位: University of Minnesota-Twin Cities
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 1998
• 资助国家: 美国
• 起止时间: 1998-10-01 至 2001-09-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=9813703&HistoricalAwards=false
• 关键词: Mechanisms; Secure; Robust; Agent; Based

A mobile agent is a program capable of migrating autonomously fromnode to node, to perform computation on behalf of a human user. Themobile agent paradigm can play a significant role in current and futurenetwork-centric applications, especially in the context of the NextGeneration Internet. Security and robustness concerns about mobileagents are currently identified as the biggest obstacle preventing thewidespread deployment and use of agent-based applications.The primary objective of the proposed research is to investigate securityand robustness issues in building a mobile agent programming system,in an open distributed environment, such as the Internet. The proposedresearch addresses issues related to programming primitives andsystem-level mechanisms for building agent-based applications. Theapproach presented in this proposal is based on supporting the mobileagent paradigm using mobile objects. This raises a number ofresearch challenges related to protection of host resources, protectionof mobile agents, and secrecy and authenticity of information carriedby agents. It also requires mechanisms for remote monitoring andcontrol of agents, and recovery from failures. Another related challengeto be addressed by this effort is in building scalable mechanisms fornaming and locating objects in the network. The proposed research will bebased on the Java language and will use cryptographic techniques to designprotocols to solve several of the security problems. The facilitiesprovided by Java (such as object serialization, reflection, securityfeatures, etc.) will be used in devising mechanisms for agent mobility andsupporting protection domains for agent execution on remote hosts.A prototype implementation of a mobile agent system will be built toserve both as a research infrastructure as well as a demonstrationof the mechanisms to be developed during this effort. The systemwill be evaluated using both theoretical and experimental analyses.

移动代理是一种能够自主地从一个节点迁移到另一个节点，代表人类用户执行计算的程序。移动代理范式可以在当前和未来以网络为中心的应用中发挥重要作用，特别是在下一代互联网的背景下。目前，移动代理的安全性和健壮性问题被认为是阻碍基于代理的应用广泛部署和使用的最大障碍。本研究的主要目的是研究在开放的分布式环境(如互联网)中构建移动代理编程系统时的安全性和健壮性问题。提出的研究解决了与构建基于代理的应用程序的原语编程和系统级机制相关的问题。该方案中提出的方法是基于使用移动对象支持移动代理范例的。这就提出了一些与主机资源保护、移动代理保护以及代理携带的信息的保密性和真实性相关的研究挑战。它还需要远程监视和控制代理以及从故障中恢复的机制。这项工作要解决的另一个相关挑战是为网络中的对象命名和定位建立可扩展的机制。这项拟议的研究将基于Java语言，并将使用密码技术来设计协议，以解决几个安全问题。Java提供的便利(如对象序列化、反射、安全特性等)将用于设计代理移动机制和支持代理在远程主机上执行的保护域。将构建移动代理系统的原型实现，作为研究基础设施以及在这一努力中将开发的机制的演示。该系统将通过理论分析和实验分析进行评估。