Travel and Registration Support for Third Bertinoro Workshop on Future of Distributed Computing

分布式系统越来越多地部署在多个管理域(MAD)上，在这些域中，没有一个机构控制所有参与节点。传统上，分布式系统中的节点偏离其规范是因为它们被破坏(例如，由于错误、硬件故障、配置错误，甚至恶意攻击)。MAD系统增加了一个新的维度：如果没有中央管理员确保每个未中断的节点遵守指定的协议，节点可能会偏离指定的协议，以自私地最大化其效用。拜占庭容错能够很好地处理中断节点。然而，拜占庭模型将所有偏差归类为故障，并要求对故障数量设定一个界限，而当节点可能是自私的时候，这个界限是不可能确定的。相反，传统的博弈论模型很好地处理了自私的节点，但即使是一个坏节点的行为不合理，也容易受到任意破坏。2007年6月4日至6月7日举行的FuDiCo III国际研讨会的目标是确定构建可靠的MAD分布式系统的概念和实践基础。研讨会是享有盛誉的分布式计算未来(FuDiCo)系列的第三个研讨会，聚集了来自不同领域(安全、分布式计算、人工智能、网络和经济)的领先高级研究人员和有前途的研究生，为一个跨文化？解决这一具有挑战性和多面性的问题。其目的是制作一系列书面报告(针对SIGACT和SIGOPS新闻)，提供该领域当前研究的快照，并确定未来研究的有希望的途径。

Distributed systems are increasingly deployed over Multiple Administrative Domains (MADs) in which no single authority controls all participating nodes. Traditionally, nodes in distributed systems deviate from their specification because they are broken (e.g., because of bugs, hardware failures, configuration errors, or even malicious attacks). MAD systems add a new dimension: without a central administrator ensuring that each unbroken node follows the assigned protocol, nodes may deviate to selfishly maximize their utility.Byzantine Fault Tolerance handles broken nodes well. However, the Byzantine model classifies all deviations as faults and requires a bound on the number of faults that is impossible to establish when nodes can be selfish. Conversely, traditional game-theoretic models handle selfish nodes well, but are vulnerable to arbitrary disruptions if even one broken node behaves irrationally.The goal of the FuDiCo III international workshop, held from June 4th to June 7th 2007, is to identify the conceptual and practical foundations on which to build dependable MAD distributed systems. The workshop, the third in the prestigious Future of Distributed Computing (FuDiCo) series, brings together leading senior researchers and promising graduate students from different areas (security, distributed computing, artificial intelligence, networking, and economics) for a ?multicultural? approach to this challenging and multi-faceted problem. The objective is to produce a series of written reports (targeted for SIGACT and SIGOPS News) that provide a snapshot of the current research in the area and identify promising avenues for future research.