Improving Performance of Key–Value Stores for High-Performance Storage Devices
Source abstract
Abstract
Key–value stores (KV stores) are becoming popular in both academia and industry due to their high performance and simplicity in data management. Unlike traditional database systems such as relational databases, KV stores manage data in key–value pairs and do not support relationships between the data. This simplicity enables KV stores to offer higher performance. To further improve the performance of KV stores, high-performance storage devices such as solid-state drives (SSDs) and non-volatile memory express (NVMe) SSDs have been widely adopted. These devices are intended to expedite data processing and storage. However, our studies indicate that, due to a lack of multi-thread-oriented programming, the performance of KV stores is far below the raw performance of high-performance storage devices. In this paper, we analyze the performance of existing KV stores utilizing high-performance storage devices. Our analysis reveals that the actual performance of KV stores is below the potential performance that these storage devices could offer. According to the profiling results, we argue that this performance gap is due to the coarse-grained locking mechanisms of existing KV stores. To alleviate this issue, we propose a multi-threaded compaction operation that leverages idle threads to participate in I/O operations. Our experimental results demonstrate that our scheme can improve the performance of KV stores by up to 16% by increasing the number of threads involved in I/O operations.
Source: crossref · Quality: baseline
Research context
OpenAlex topics: Advanced Data Storage Technologies · Cloud Computing and Resource Management · Distributed systems and fault tolerance
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
Affiliations
Department of Computer Science and Engineering, Seoul National University of Science and Technology, Seoul 01811, Republic of Korea
Department of Smart ICT Convergence Engineering, Seoul National University of Science and Technology, Seoul 01811, Republic of Korea
Funding
Basic Science Research Program
References (31)
Show source-provided references
1. Marmol, L., Sundararaman, S., Talagala, N., Rangaswami, R., Devendrappa, S., Ramsundar, B., and Ganesan, S. (2014, January 17–18). NVMKV: A Scalable and Lightweight Flash Aware Key-Value Store. Proceedings of the 6th USENIX Workshop on Hot Topics in Storage and File Systems, Philadelphia, PA, USA.
2. Drolia, U., Mickulicz, N., Gandhi, R., and Narasimhan, P. (2015, January 7). Krowd: A key-value store for crowded venues. Proceedings of the 10th International Workshop on Mobility in the Evolving Internet Architecture, Paris, France.10.1145/2795381.2795388 ↗
3. Chen, H., Ruan, C., Li, C., Ma, X., and Xu, Y. (2021, January 23–25). SpanDB: A fast, Cost-Effective LSM-tree based KV store on hybrid storage. Proceedings of the 19th USENIX Conference on File and Storage Technologies, Online.
4. Lu, Z., Cao, Q., Jiang, H., Wang, S., and Dong, Y. (2022, January 5–8). p2kvs: A portable 2-dimensional parallelizing framework to improve scalability of key-value stores on ssds. Proceedings of the Seventeenth European Conference on Computer Systems, Rennes, France.
5. InPlaceKV: In-place update scheme for SSD-based KV storage systems under update-intensive Worklaods10.1007/s10586-023-04031-9 ↗
6. Limon: A Scalable and Stable Key-Value Engine for Fast NVMe Devices10.1109/TC.2023.3283674 ↗
7. Yu, J., Noh, S.H., Choi, Y.R., and Xue, C.J. (2023, January 21–23). ADOC Automatically Harmonizing Dataflow Between Components in Log-Structured Key-Value Stores for Improved Performance. Proceedings of the 21st USENIX Conference on File and Storage Technologies, Santa Clara, CA, USA.
8. Bigtable: A distributed storage system for structured data10.1145/1365815.1365816 ↗
9. Cassandra: A decentralized structured storage system10.1145/1773912.1773922 ↗
10. Im, J., Bae, J., Chung, C., and Lee, S. (2020, January 15–17). PinK: High-speed In-storage Key-value Store with Bounded Tails. Proceedings of the 2020 USENIX Annual Technical Conference, Online.10.1145/3452846 ↗
11. Koo, J., Im, J., Song, J., Park, J., Lee, E., Kim, B.S., and Lee, S. (2021, January 14–16). Modernizing file system through in-storage indexing. Proceedings of the 15th USENIX Symposium on Operating Systems Design and Implementation, Online.
12. PetaKV: Building Efficient Key-Value Store for File System Metadata on Persistent Memory10.1109/TPDS.2022.3232382 ↗
13. Cai, T., Chen, F., He, Q., Niu, D., and Wang, J. (2019). The matrix KV storage system based on NVM devices. Micromachines, 10.10.3390/mi10050346 ↗
14. Lu, L., Zhang, Y., Do, T., Al-Kiswany, S., Arpaci-Dusseau, A.C., and Arpaci-Dusseau, R.H. (2014, January 6–8). Physical Disentanglement in a {Container-Based} File System. Proceedings of the 11th USENIX Symposium on Operating Systems Design and Implementation (OSDI 14), Broomfield, CO, USA.
15. Kang, J., Zhang, B., Wo, T., Yu, W., Du, L., Ma, S., and Huai, J. (2015, January 8–10). SpanFS: A scalable file system on fast storage devices. Proceedings of the 2015 USENIX Annual Technical Conference, Santa Clara, CA, USA.
16. Son, Y., Kim, S., Yeom, H.Y., and Han, H. (2018, January 12–15). High-performance transaction processing in journaling file systems. Proceedings of the 16th USENIX Conference on File and Storage Technologies, Oakland, CA, USA.
17. Bhat, S.S., Eqbal, R., Clements, A.T., Kaashoek, M.F., and Zeldovich, N. (2017, January 28–31). Scaling a file system to many cores using an operation log. Proceedings of the 26th Symposium on Operating Systems Principles, Shanghai, China.10.1145/3132747.3132779 ↗
18. Liao, X., Lu, Y., Xu, E., and Shu, J. (2021, January 14–16). Max: A {Multicore-Accelerated} File System for Flash Storage. Proceedings of the 2021 USENIX Annual Technical Conference (USENIX ATC 21), Online.
19. Borthakur, D. (2024, July 16). Under the Hood: Building and Open-Sourcing RocksDB. Facebook Engineering Notes. Available online: https://engineering.fb.com/2013/11/21/core-infra/under-the-hood-building-and-open-sourcing-rocksdb/.
20. (2024, July 16). A Library That Provides an Embeddable, Persistent Key-Value Store for Fast Storage. Available online: https://github.com/facebook/rocksdb.
21. Wu, S.M., Lin, K.H., and Chang, L.P. (2018, January 19–23). KVSSD: Close integration of LSM trees and flash translation layer for write-efficient KV store. Proceedings of the 2018 Design, Automation & Test in Europe Conference & Exhibition, Dresden, Germany.10.23919/DATE.2018.8342070 ↗
22. Lee, C.G., Kang, H., Park, D., Park, S., Kim, Y., Noh, J., Chung, W., and Park, K. (2019, January 21–25). iLSM-SSD: An intelligent LSM-tree based key-value SSD for data analytics. Proceedings of the 2019 IEEE 27th International Symposium on Modeling, Analysis, and Simulation of Computer and Telecommunication Systems (MASCOTS), Rennes, France.10.1109/MASCOTS.2019.00048 ↗
23. Wisckey: Separating keys from values in ssd-conscious storage10.1145/3033273 ↗
24. Zuo, P., Hua, Y., and Wu, J. (2018, January 8–10). Write-optimized and high-performance hashing index scheme for persistent memory. Proceedings of the 13th USENIX Symposium on Operating Systems Design and Implementation, Carlsbad, CA, USA.
25. Chandramouli, B., Prasaad, G., Kossmann, D., Levandoski, J., Hunter, J., and Barnett, M. (2018, January 10–15). FASTER: A concurrent key-value store with in-place updates. Proceedings of the 2018 International Conference on Management of Data, Houston, TX, USA.10.1145/3183713.3196898 ↗
26. Liu, L., Wang, H., and Zhou, K. (2018, January 11–14). An optimized implementation for concurrent LSM-structured key-value stores. Proceedings of the 2018 IEEE International Conference on Networking, Architecture and Storage, Chongqing, China.10.1109/NAS.2018.8515730 ↗
27. Sun, X., Yu, J., Zhou, Z., and Xue, C.J. (2020, January 20–24). FPGA-based compaction engine for accelerating LSM-tree key-value stores. Proceedings of the 2020 IEEE 36th International Conference on Data Engineering, Dallas, TX, USA.10.1109/ICDE48307.2020.00113 ↗
28. Intel (2024, July 16). SPDK: Storage Performance Development Kit. Available online: https://spdk.io/.
29. Meta Platforms, Inc. (2024, July 16). RocksDB|A Persistent Key-Value Store|RocksDB. Available online: https://rocksdb.org/.
30. Guz, Z., Li, H.H., Shayesteh, A., and Balakrishnan, V. (2017, January 22–24). NVMe-over-fabrics performance characterization and the path to low-overhead flash disaggregation. Proceedings of the 10th ACM International Systems and Storage Conference, Haifa, Israel.10.1145/3078468.3078483 ↗
31. Zion, I.B. (2019, January 3–5). Key-value FTL over open channel SSD. Proceedings of the 12th ACM International Conference on Systems and Storage, Haifa, Israel.10.1145/3319647.3325853 ↗