Blockchain is rapidly emerging as an important class of network application, with a unique set of trust, security and transparency properties. In a blockchain system, participants record and update the `server-side' state of an application as blocks of a replicated, immutable ledger using a consensus protocol over the Internet. Mining blocks has become lucrative in recent years; e.g., a miner receives over USD 200,000 per mined block in Bitcoin today. A key factor affecting mining rewards, is the latency of broadcasting blocks over the network. In this paper, we consider the problem of topology design for optimizing mining rewards in a wide-area blockchain network that uses a Proof-of-Work protocol for consensus. Contrary to general wisdom that a faster network is always better for miners, we show a counter intuitive result where a slower network is actually beneficial to some miners. This is because competing miners must choose neighbors that not only decrease their own latency to others, but also ensure that the latency between other miners do not decrease because of itself. We formalize this problem, and provide both theoretical analysis and experimental results to support our claim.
翻译:屏障链正在迅速成为一个重要的网络应用类别,具有一套独特的信任、安全和透明度特性。在一个链条系统中,参与者记录并更新了一个应用程序的“服务器侧面”状态,将其作为利用互联网的协商一致协议进行复制的、不可改变的分类的块块。采矿区块近年来已经变得有利可图;例如,一个矿工今天在比特币每块雷区得到20多万美元。影响采矿收益的一个关键因素是广播区块在网络上的延缓性。在本文中,我们考虑了在使用“工作验证协议以达成共识”的广域链网中优化采矿收益的地形设计问题。与一般的智慧相反,一个更快的网络对矿工来说总是更好,我们表现出一种反直觉的结果,因为网络速度放慢实际上对一些矿工有利。这是因为,竞争的矿工必须选择邻居,而不仅减少他们自己的耐久性,而且确保其他矿工之间的耐久性不会因为其本身而减少。我们把这个问题正式化,并且提供理论分析和实验结果来支持我们的要求。