Distributed transactions on high-overhead TCP/IP-based networks were conventionally considered to be prohibitively expensive. In fact, the primary goal of existing partitioning schemes is to minimize the number of cross-partition transactions. However, with the new generation of fast RDMAenabled networks, this assumption is no longer valid. In this paper, we first make the case that the new bottleneck which hinders truly scalable transaction processing in modern RDMA-enabled databases is data contention, and that optimizing for data contention leads to different partitioning layouts than optimizing for the number of distributed transactions. We then present Chiller, a new approach to data partitioning and transaction execution, which aims to minimize data contention for both local and distributed transactions.

中文翻译：

---

冷水机

传统上认为，基于 TCP/IP 的高开销网络上的分布式事务非常昂贵。事实上，现有分区方案的主要目标是最小化跨分区事务的数量。然而，随着新一代快速 RDMA 网络的出现，这种假设不再有效。在本文中，我们首先说明了阻碍现代 RDMA 数据库中真正可扩展事务处理的新瓶颈是数据争用，并且针对数据争用进行优化会导致不同的分区布局，而不是针对分布式事务的数量进行优化。然后，我们介绍了 Chiller，这是一种数据分区和事务执行的新方法，旨在最大限度地减少本地和分布式事务的数据争用。