The adversary's goal in mounting Long Range Attacks (LRAs) is to fool potential victims into using and relying on a side chain, i.e., a false, alternate history of transactions, and into proposing transactions that end up harming themselves or others. Previous research work on LRAs on blockchain systems have used, at a high level, one of two approaches. They either try to (1) prevent the creation of a bogus side chain or (2) make it possible to distinguish such a side chain from the main consensus chain. In this paper, we take a different approach. We start with the indistinguishability of side chains from the consensus chain -- for the eclipsed victim -- as a given and assume the potential victim will be fooled. Instead, we protect the victim via harm reduction applying "short leashes" to transactions. The leashes prevent transactions from being used in the wrong context. The primary contribution of this paper is the design and analysis of leashes. A secondary contribution is the careful explication of the LRA threat model in the context of BAR fault tolerance, and using it to analyze related work to identify their limitations.
翻译:长距离袭击(LARC)的对手目标是欺骗潜在受害者,使其使用和依赖一个侧链,即虚假的、不同的交易历史,并提议最终伤害自己或其他人的交易。以前对链链系统上帝军的研究工作在高层次上使用了两种方法之一。它们或者试图(1) 防止建立一个假的侧链,或者(2) 使这种侧链与主要共识链区分开来。在本文中,我们采取不同的做法。我们从侧链与共识链的不可分性开始 -- -- 即对日蚀的受害者 -- -- 开始,作为给定的,并假定潜在受害者会被骗。相反,我们通过对交易应用 " 浅利沙 " 来减少伤害来保护受害者。皮带防止交易被错误地利用。本文的主要贡献是设计和分析皮带。一个次要贡献是,在BAR错误容忍的背景下,仔细地展示上帝军的威胁模式,并利用它来分析相关工作以确定其局限性。