一句话：提出软硬件协同的 CTXNL 系统，用面向事务数据弱一致性的混合一致性原语，提升 CXL 事务处理吞吐。
现代数据中心应用依赖事务处理系统，但现有多节点系统受网络开销拖累。本文提出以 Compute Express Link（CXL）作为网络替代方案，利用低延迟、缓存一致的共享内存访问。然而直接采用标准 CXL 原语会因维护跨节点缓存一致性成本过高而降低性能。为此，作者设计 CTXNL：一种软硬件协同系统，实现面向事务数据弱一致性特点的新型混合一致性原语，核心是让事务系统开发者能够按需选择性地获得数据一致性。评估在 OLTP 负载上进行，CTXNL 优于基于网络的系统，并在通用事务处理策略下相较原生 CXL 内存共享架构实现最高 2.08 倍吞吐提升。
Transaction processing systems are the crux for modern data-center applications, yet current multi-node systems are slow due to network overheads. This paper advocates for Compute Express Link (CXL) as a network alternative, which enables low-latency and cache-coherent shared memory accesses. However, directly adopting standard CXL primitives leads to performance degradation due to the high cost of maintaining cross-node cache coherence. To address the CXL challenges, this paper introduces CTXNL, a software-hardware co-designed system that implements a novel hybrid coherence primitive tailored to the loosely coherent nature of transactional data. The core innovation of CTXNL is empowering transaction system developers with the ability to selectively achieve data coherence. Our evaluations on OLTP workloads demonstrate that CTXNL enhances performance, outperforming current network-based systems and achieves up to 2.08x greater throughput than vanilla CXL memory sharing architectures across universal transaction processing policies.
@inproceedings{WangCLGNGDW025,
title = {C
<scp>T</scp>
X
<scp>N</scp>
L: A Software-Hardware Co-designed Solution for Efficient CXL-Based Transaction Processing},
author = {Zhao Wang and Yiqi Chen and Cong Li and Yijin Guan and Dimin Niu and Tianchan Guan and Zhaoyang Du and Xingda Wei and Guangyu Sun},
booktitle = {ASPLOS},
year = {2025},
doi = {10.1145/3676641.3716244}
}