The aim of this work is to evaluate the impact of node churn -nodes leaving and rejoining the network- on the Bitcoin network. We provide a comprehensive analytical model for the churning process. We use a Continuous Time Markov Chain (CTMC) to describe the behavior of a node, and then apply the results to model the changes in connectivity and the impact on network performance. We analyze the time needed to resynchronize a node upon rejoining the network and find that sleep times of the order of hours require synchronization times limited by a minute. We estimate the impact of sleep and synchronization time on overall network connectivity and block/transaction distribution time. Our results show that networks with less than 4000 nodes are sensitive to churn. This occurs due to opposing impact of decrease in network size (and diameter) due to sleep time and increase of communication load per node. However, the impact of churn on network with more than 4000 nodes is noticeable but small enough to make a large Bitcoin network fairly resilient to churn.

中文翻译：

---

比特币网络中节点流失的影响


这项工作的目的是评估节点流失（节点离开和重新加入网络）对比特币网络的影响。我们为搅拌过程提供全面的分析模型。我们使用连续时间马尔可夫链（CTMC）来描述节点的行为，然后应用结果来建模连接的变化以及对网络性能的影响。我们分析了节点重新加入网络时重新同步所需的时间，发现数小时的睡眠时间需要限制为一分钟的同步时间。我们估计睡眠和同步时间对整体网络连接和块/交易分发时间的影响。我们的结果表明，节点数少于 4000 个的网络对流失很敏感。发生这种情况是由于睡眠时间导致网络规模（和直径）减小以及每个节点通信负载增加的相反影响。然而，流失对超过 4000 个节点的网络的影响是显而易见的，但小到足以使大型比特币网络对流失具有相当的弹性。