Vigil-KV: Hardware-Software Co-Design to Integrate Strong Latency Determinism into Log-Structured Merge Key-Value Stores
Miryeong Kwon, Seungjun Lee, Hyunkyu Choi, Jooyoung Hwang, Myoungsoo Jung
Abstract
We propose Vigil-KV, a hardware and software co-designed framework that eliminates long-tail latency almost perfectly by introducing strong latency determinism. To make Get latency deterministic, Vigil-KV first enables a predictable latency mode (PLM) interface on a real datacenter-scale NVMe SSD, having knowledge about the nature of the underlying flash technologies. Vigil-KV at the system-level then hides the non-deterministic time window (associated with SSD's internal tasks and/or write services) by internally scheduling the different device states of PLM across multiple physical functions. Vigil-KV further schedules compaction/flush operations and client requests being aware of PLM's restrictions thereby integrating strong latency determinism into LSM KVs. We implement Vigil-KV upon a 1.92TB NVMe SSD prototype and Linux 4.19.91, but other LSM KVs can adopt its concept. We evaluate diverse Facebook and Yahoo scenarios with Vigil-KV, and the results show that Vigil-KV can reduce the tail latency of a baseline KV system by 3.19× while reducing the average latency by 34%, on average.
Ask about this paper
Your agent reads all of it.
Lune indexed this paper to the last equation, along with the top-tier papers that cite it. Ask a question and the answer quotes them.
Your agent calls
Luneget_paper_fulltext
Free to start. No credit card required.
Terminal
Install the CLIlune papers fulltext 8f24b754-31ee-4fd4-8872-9c617a844d2aCited by top-tier papers1
Ask how each one uses itBuilds on12
- SpanDB: A Fast, Cost-Effective LSM-tree Based KV Store on Hybrid StorageHao Chen, Chaoyi Ruan, Cheng Li, Xiaosong Ma et al.FAST 2021 · 120 citations
- Evolution of Development Priorities in Key-value Stores Serving Large-scale Applications: The RocksDB ExperienceSiying Dong, Andrew Kryczka, Yanqin Jin, Michael StummFAST 2021 · 110 citations
- SplinterDB: Closing the Bandwidth Gap for NVMe Key-Value StoresAlexander Conway, Abhishek Gupta, Vijay Chidambaram, Martin Farach-Colton et al.USENIX ATC 2020 · 90 citations
- PinK: High-speed In-storage Key-value Store with Bounded TailsJunsu Im, Jinwook Bae, Chanwoo Chung, Arvind et al.USENIX ATC 2020 · 85 citations
- Differentiated Key-Value Storage Management for Balanced I/O PerformanceYongkun Li, Zhen Liu, Patrick P. C. Lee, Jiayu Wu et al.USENIX ATC 2021 · 79 citations
Related papers
- lODA: A Host/Device Co-Design for Strong Predictability Contract on Modern Flash StorageHuaicheng Li, Martin L. Putra, Ronald Shi, Xing Lin et al.SOSP 2021 · 36 citations
- Analysis and Evaluation of Using Microsecond-Latency Memory for In-Memory Indices and Caches in SSD-Based Key-Value StoresYosuke Bando, Akinobu Mita, Kazuhiro Hiwada, Shintaro Sano et al.SIGMOD 2026
- Revisiting Log-Structured Merging for KV Stores in Hybrid Memory SystemsZhuohui Duan, Jiabo Yao, Haikun Liu, Xiaofei Liao et al.ASPLOS 2023 · 25 citations
- PartitionKV: Redesigning LSM-tree KV Stores on NVMs with Adaptive Partitioning for Reducing Write Stalls and AmplificationXingye Huang, Jinyu Wu, Xiaofang Xia, Jiangtao Cui et al.SIGMOD 2026
- Holistic and Automated Task Scheduling for Distributed LSM-tree-based StorageYuanming Ren, Siyuan Sheng, Zhang Cao, Yongkun Li et al.FAST 2026 · 1 citation
