arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1205.0477cs.DS
收藏

拜占庭故障下同步消息传递系统的保序重命名

Order-preserving Renaming in Synchronous Message Passing Systems with Byzantine Faults

Oksana Denysyuk, Luis Rodrigues

摘要:

重命名是分布式计算中的一个基本问题,它由一组进程从给定的命名空间中选取不同的名称组成。本文提出的算法,解决保持顺序重命名的同步消息传递系统与拜占庭进程。据我们所知,这项工作是第一个解决给定模型中的保序重命名。虽然这个问题可以通过使用共识来解决,但众所周知,重命名比共识“弱”,因此我们主要关注执行重命名的效率,并在这个方向上做出三个贡献。我们提出了一个保序重命名算法$N > 3t$与目标命名空间的大小$N+t-1$和对数步长复杂度(其中$N$是进程的数量和$t$是故障的数量上界)。类似于现有的耐撞解决方案,我们的算法采用的想法,从近似协议问题。我们表明,我们的算法具有恒定的步骤复杂性,如果$N>t^2+2t$,并实现紧凑的命名空间的大小$N$。最后,我们提出了一个算法,解决了保序重命名在短短的2个通信步骤,如果$N > 2t^2 + t$。

英文摘要:

Renaming is a fundamental problem in distributed computing, which consists of a set of processes picking distinct names from a given namespace. The paper presents algorithms that solve order-preserving renaming in synchronous message passing systems with Byzantine processes. To the best of our knowledge, this work is the first to address order-preserving renaming in the given model. Although this problem can be solved by using consensus, it is known that renaming is "weaker" than consensus, therefore we are mainly concerned with the efficiency of performing renaming and make three contributions in this direction. We present an order-preserving renaming algorithm for $N > 3t$ with target namespace of size $N+t-1$ and logarithmic step complexity (where $N$ is the number of processes and $t$ is an upper bound on the number of faults). Similarly to the existing crash-tolerant solution, our algorithm employs the ideas from the approximate agreement problem. We show that our algorithm has constant step complexity if $N>t^2+2t$ and achieves tight namespace of size $N$. Finally, we present an algorithm that solves order-preserving renaming in just 2 communication steps, if $N > 2t^2 + t$.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑