Conclusion
In this paper, we propose a fast high-availability in-memory key-value store based on RDMA network and Optane DCPMM named FaHA. FaHA proposes RDMA persist PRC that enables remote data persistence for fast log shipping with minimal round-trip and persistence overhead. FaHA further designs append-only storage with pipeline batching scheme in NVM and a hotness-aware differential hash index in DRAM to relieve read/write amplification of Optane DCPMM. Evaluations show that FaHA supports up to 2–3 synchronous backups without significantly lowering the primary’s performance, and outperforms existing works.
References
Wang T, Johnson R, Pandis I. Query fresh: log shipping on steroids. Proceedings of the VLDB Endowment, 2017, 11(4): 406–419
Yang J, Kim J, Hoseinzadeh M, Izraelevitz J, Swanson S. An empirical guide to the behavior and use of scalable persistent memory. In: Proceedings of the 18th USENIX Conference on File and Storage Technologies. 2020, 169–182
Cooper B F, Silberstein A, Tam E, Ramakrishnan R, Sears R. Benchmarking cloud serving systems with YCSB. In: Proceedings of the 1st ACM Symposium on Cloud Computing. 2010, 143–154
Author information
Authors and Affiliations
Corresponding author
Electronic supplementary material
Rights and permissions
About this article
Cite this article
Qi, X., Hu, H., Guo, J. et al. High-availability in-memory key-value store using RDMA and Optane DCPMM. Front. Comput. Sci. 17, 171603 (2023). https://doi.org/10.1007/s11704-022-1123-8
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s11704-022-1123-8