Distributed computing with mobile agents in unsafe networks

基于Internet的技术被广泛接受，基于事务的技术在许多系统(如电信系统)中的普及，以及网格系统的出现，所有这些都强烈地强调了分布式算法在当今计算世界中的至高无上的重要性。此外，在网络环境中进行计算时，移动代理的使用也变得越来越流行。但安全威胁的存在使解决所有问题的方案变得非常复杂，但不是微不足道的任务。也就是说，在支持移动代理的系统中，安全是最紧迫的问题，也可能是最难解决的问题。这项研究的总体目标在于开发高效和可测试的分布式算法，使用移动代理来解决不安全网络中的非平凡任务，即在存在安全威胁的情况下。所提出的研究被组织成三个大的目标：目标1：在存在有害主机的情况下调查分布式算法；目标2：使用移动故障模型研究不安全网络中的分布式算法。目标3：开发用于在不安全网络中验证分布式算法的可测试模型。目标1特别关注具有有害主机的网络。我们特别感兴趣的解决方案是尽可能限制对代理和网络的要求(以便最大限度地提高在支持代理的实际网络上实现此类解决方案的可行性)。更具体地说，我们的目标是算法尽可能高效，同时不需要松散约束(可能不现实)的机制。与目标1有关的所有工作都植根于处理器故障模型的使用，也就是将故障定位在网络节点中的模型。相比之下，目标2采用了更一般的移动故障模型。目标2有两个目标：a)研究如何使在不安全网络中使用处理器故障模型的分布式算法改为使用移动故障模型；b)在采用移动故障模型的不安全网络中开发新的高效分布式算法。最后，目标3解决了分布式算法在不安全网络中的测试问题。

The wide acceptance of Internet-based technologies, the pervasiveness of transaction-based techniques in a multitude of systems (such as telecommunication ones), as well as the emergence of Grid systems, all strongly emphasize the paramount importance of distributed algorithms in today's computing world. Also, the use of mobile agents is becoming increasingly popular when computing in networked environments. But the presence of security threats greatly complicates solutions for all but non-trivial tasks. That is, in systems supporting mobile agents, security is the most pressing concern, and possibly the most difficult to address. The overall objective of this research consists in developing efficient and testable distributed algorithms that use mobile agents to solve non-trivial tasks in unsafe networks, that is, in the presence of security threats. The proposed research is organized into three broad goals: Goal 1: Investigate distributed algorithms in the presence of harmful hosts;Goal 2: Investigate distributed algorithms in an unsafe network using the mobile faults model.;Goal 3: Develop a testable model for the validation of distributed algorithms in an unsafe network. Goal 1 specifically focuses on networks with harmful hosts. We are particularly interested in solutions that constrain as much as possible the requirements imposed on agents and network (in order to maximize the feasibility of the realization of such solutions on actual networks supporting agents). More specifically, we aim for algorithms that are as efficient as possible while not requiring loosely constrained (possibly unrealistic) mechanisms. All work pertaining to Goal 1 is rooted in the use of the processor failure model, that is, a model in which faults are localized in nodes of the network. In contrast, Goal 2 adopts the more general mobile faults model. Goal 2 is two-fold: a) investigate how to adapt distributed algorithms that use the processor failure model in an unsafe network to instead use the mobile faults model and b) develop new efficient distributed algorithms in unsafe networks that adopt the mobile faults model. Finally, Goal 3 addresses the testing of distributed algorithms in unsafe networks.