探索面向分布式共享缓存架构的高性能数据库

尚碧筠; 魏星; 周士俊; 冀文冠; 董诗琦; 屠要峰; 董振江

doi:10.12263/DZXB.20240868

您当前的位置：

首页 >

文章列表页 >

探索面向分布式共享缓存架构的高性能数据库

云边端协同数据管理、分析与系统 | 更新时间：2025-12-08

- 探索面向分布式共享缓存架构的高性能数据库
- Exploring the High-Performance Database Based on Distributed Shared Cache Architecture
- 电子学报 2025年53卷第2期页码：314-328
- 作者机构：
  
  1.南京邮电大学计算机学院，江苏南京 210023
  2.中兴通讯股份有限公司，广东深圳 518057
- 作者简介：
  
  [ "尚碧筠女，2004年5月出生于江苏省南京市.本科.主要研究方向为数据库、大数据和机器学习. E-mail: shangbiyun@126.com" ]
  [ "魏星男，1994年6月出生于江苏省扬州市.博士.主要研究方向为数据库、大数据和机器学习. E-mail: wei.xing6@zte.com.cn" ]
  [ "周士俊男，1979年7月出生于江苏省扬州市.本科.主要研究方向为数据库. E-mail: zhou.shijun@zte.com.cn" ]
  [ "冀文冠男，1993年9月出生于山西省晋中市.本科.主要研究方向为数据库. E-mail: ji.wenguan@zte.com.cn" ]
  [ "董诗琦女，1997年12月出生于江苏省徐州市.硕士.主要研究方向为数据库. E-mail: ji.wenguan@zte.com.cn" ]
  [ "屠要峰男，1972年12月出生于河南省开封市.博士，研究员.主要研究方向为数据库、大数据和机器学习. E-mail: tu.yaofeng@zte.com.cn" ]
  [ "董振江男，1970年2月出生于河南省开封市.博士，教授，博士生导师.主要研究方向为人工智能、大数据、信息安全. E-mail: dongzhenjiang@njupt.edu.cn" ]
- 基金信息：
  
  国家重点研发计划(2021YFB3101101);江苏省重点研发计划重点项目(BE2023025)
- DOI：10.12263/DZXB.20240868
  中图分类号： TP311;
- 收稿：2024-09-23，
  
  修回：2025-01-26，
  
  纸质出版：2025-02-25
- 稿件说明：
移动端阅览
尚碧筠, 魏星, 周士俊, 等. 探索面向分布式共享缓存架构的高性能数据库[J]. 电子学报, 2025, 53(02): 314-328.

SHANG Bi-yun, WEI Xing, ZHOU Shi-jun, et al. Exploring the High-Performance Database Based on Distributed Shared Cache Architecture[J]. Acta Electronica Sinica, 2025, 53(02): 314-328.
尚碧筠, 魏星, 周士俊, 等. 探索面向分布式共享缓存架构的高性能数据库[J]. 电子学报, 2025, 53(02): 314-328. DOI：10.12263/DZXB.20240868

SHANG Bi-yun, WEI Xing, ZHOU Shi-jun, et al. Exploring the High-Performance Database Based on Distributed Shared Cache Architecture[J]. Acta Electronica Sinica, 2025, 53(02): 314-328. DOI：10.12263/DZXB.20240868

摘要

随着物联网和智能终端的普及，边侧产生的数据量远远超出了边缘节点的计算与存储能力，亟需云边协同处理以满足大规模数据的实时分析需求.共享缓存架构因其计算、内存与存储解耦的特点，成为满足边缘海量数据处理需求的关键方案.然而，在共享缓存架构中仍然存在一些亟待解决的问题.首先，在事务处理场景中，当热点缓存数据在节点间频繁迁移时，现有系统的日志落盘机制会产生大量日志写入操作，进而影响系统性能.此外，现有的缓存写失效机制会导致部分热点缓存数据频繁被淘汰，致使一些执行较慢的事务无法及时从共享缓存中读取目标数据，触发大量缓存重载而引发系统性能下降.针对这些问题，本文提出了一种基于依赖表的日志延迟写入机制，通过整合多条日志写盘操作，并推迟到日志缓冲区填满或事务提交时刻，降低了日志刷写频次和写盘开销；设计了一种异步缓存延迟失效机制，通过引入异步回放失效消息、页面可见性判断及优化的缓存替换策略，有效延长缓存数据服务时间，提升了缓存命中率和系统性能.基于这些机制，本文实现了一套高性能共享缓存数据库系统EBASE-T.实验结果表明：与优化前相比，EBASE-T的吞吐量提升了19.5%，时延降低了13.1%，在TPC-C（专门针对联机交易处理系统的规范）测试中，EBASE-T相较于大多数共享缓存数据库系统表现出了显著的性能优势.

Abstract

With the widespread adoption of internet of things (IoT) and smart devices

the volume of data generated at the edge has far exceeded the computational and storage capabilities of edge nodes. This creates an urgent need for cloud-edge collaborative processing to meet the real-time analysis demands of large-scale data. With the decoupling of computation

memory

and storage

the shared-cahe architecture has become a critical solution for addressing the processing requirements of massive edge data. However

there are still several issuses remained in shared-cache architecture. First

in transactional processing scenarios

when hotspot cached data frequently migrates between nodes

the log persistence mechanisms of existing databases will generate a large number of log write operations

thereby impacting system performance. Secondly

the existing cache write-invalidation mechanism can lead to frequent eviction of some hotspot cached data

causing slower transactions to fail in retrieving target data from the shared cache in time. This could trigger a large number of cache reloads

resulting in system performance degradation. To address these issues

this paper proposes a dependency-table-based delayed log flushing mechanism. By consolidating multiple log write operations and deferring them until the log buffer is full or a transaction is committed

the mechanism reduces the frequency of log flushing and the overhead of disk writes. In addition

this paper also introduces a cache delayed invalidation mechanism that incorporates asynchronous replay of invalidation messages

page visibility determination

and an optimized cache replacement. This approach effectively extends the service time of cached data

improving cache hit rates and overall system performance. Based on these mechanisms

this paper implements a high-performance shared-cache database system called EBASE-T. Experimental results show that

compared to its pre-optimized version

EBASE-T achieves a 19.5% increase in throughput and a 13.1% reduction in latency. In TPC-C (online transaction processing system benchmarks) tests

EBASE-T demonstrates significant performance advantages over most shared-cache database systems.

关键词

Keywords

references

CORBETT J C , DEAN J , EPSTEIN M , et al . Spanner: Google’s globally distributed database [J ] . ACM Transactions on Computer Systems , 2013 , 31 ( 3 ): 1 - 22 .

YANG Z K , YANG C H , HAN F S , et al . OceanBase: A 707 million tpmC distributed relational database system [J ] . Proceedings of the VLDB Endowment , 2022 , 15 ( 12 ): 3385 - 3397 .

LYU Z H , ZHANG H H , XIONG G , et al . Greenplum: A hybrid database for transactional and analytical workloads [C ] // Proceedings of the 2021 International Conference on Management of Data . New York : ACM , 2021 : 2530 - 2542 .

HUANG D X , LIU Q , CUI Q , et al . TiDB: A raft-based HTAP database [J ] . Proceedings of the VLDB Endowment , 2020 , 13 ( 12 ): 3072 - 3084 .

PAZ J R G . Microsoft Azure Cosmos DB Revealed: A Multi-Modal Database Designed for the Cloud [M ] . Beach Park : Apress , 2018 .

VERBITSKI A , GUPTA A , SAHA D , et al . Amazon aurora: Design considerations for high throughput cloud-native relational databases [C ] // Proceedings of the 2017 ACM International Conference on Management of Data . New York : ACM , 2017 : 1041 - 1052 .

PANDIS I . The evolution of Amazon redshift [J ] . Proceedings of the VLDB Endowment , 2021 , 14 ( 12 ): 3162 - 3174 .

DAGEVILLE B , CRUANES T , ZUKOWSKI M , et al . The snowflake elastic data warehouse [C ] // Proceedings of the 2016 International Conference on Management of Data . New York : ACM , 2016 : 215 - 226 .

ANTONOPOULOS P , BUDOVSKI A , DIACONU C , et al . Socrates: The new SQL server in the cloud [C ] // Proceedings of the 2019 International Conference on Management of Data . New York : ACM , 2019 : 1743 - 1756 .

LAHIRI T , SRIHARI V , CHAN W , et al . Cache fusion: Extending shared-disk clusters with shared caches [C ] // Proceedings of the 27th International Conference on Very Large Data Bases . New York : ACM , 2001 : 683 - 686 .

BARSHAI V . Delivering Continuity and Extreme Capacity with the IBM DB2 Purescale Feature [M ] . New York : IBM , 2012 .

CAI Q C , GUO W T , ZHANG H , et al . Efficient distributed memory management with RDMA and caching [J ] . Proceedings of the VLDB Endowment , 2018 , 11 ( 11 ): 1604 - 1617 .

KATSARAKIS A , GAVRIELATOS V , SIAVASH KATEBZADEH M R , et al . Hermes: A fast, fault-tolerant and linearizable replication protocol [C ] // Proceedings of the Twenty-Fifth International Conference on Architectural Support for Programming Languages and Operating Systems . New York : ACM , 2020 : 201 - 217 .

LENOSKI D , LAUDON J , GHARACHORLOO K , et al . The stanford dash multiprocessor [J ] . Computer , 1992 , 25 ( 3 ): 63 - 79 .

WANG Q , LU Y , XU E , et al . Concordia: Distributed shared memory with in-network cache coherence [C ] // Proceedings of the 19th USENIX Conference on File and Storage Technologies (FAST' 21) . San Jose : USENIX , 2021 : 277 - 292 .

CAO W , LI F F , HUANG G , et al . PolarDB-X: An elastic distributed relational database for cloud-native applications [C ] // 2022 IEEE 38th International Conference on Data Engineering (ICDE) . Piscataway : IEEE , 2022 : 2859 - 2872 .

RUAN C Y , ZHANG Y Q , BI C , et al . Persistent memory disaggregation for cloud-native relational databases [C ] // Proceedings of the 28th ACM International Conference on Architectural Support for Programming Languages and Operating Systems , Volume 3 . New York : ACM , 2023: 498 - 512 .

YANG X J , ZHANG Y Q , CHEN H , et al . PolarDB-MP: A multi-primary cloud-native database via disaggregated shared memory [C ] // Companion of the 2024 International Conference on Management of Data . New York : ACM , 2024 : 295 - 308 .

LI G L , TIAN W G , ZHANG J Y , et al . GaussDB: A cloud-native multi-primary database with compute-memory-storage disaggregation [J ] . Proceedings of the VLDB Endowment , 2024 , 17 ( 12 ): 3786 - 3798 .

BOUTIN E , ABRAHAM S . Amazon aurora multi-master: Scaling out database write performance [EB/OL ] . ( 2019-08-13 )[ 2024-09-23 ] . https://d1.awsstatic.com/events/reinvent/2019/REPEAT_1_Amazon_Aurora_Multi-Master_Scaling_out_database_write_performance_DAT404-R1.pdf https://d1.awsstatic.com/events/reinvent/2019/REPEAT_1_Amazon_Aurora_Multi-Master_Scaling_out_database_write_performance_DAT404-R1.pdf .

DEPOUTOVITCH A , CHEN C , LARSON P A , et al . Taurus MM: Bringing multi-master to the cloud [J ] . Proceedings of the VLDB Endowment , 2023 , 16 ( 12 ): 3488 - 3500 .

LI C , MARKL V , MUKHERJEE N , et al . Distributed architecture of oracle database in-memory [J ] . Proceedings of the VLDB Endowment , 2015 , 8 ( 12 ): 1630 - 1641 .

WANG T Z , JOHNSON R . Scalable logging through emerging non-volatile memory [J ] . Proceedings of the VLDB Endowment , 2014 , 7 ( 10 ): 865 - 876 .

CUBUKCU U , ERDOGAN O , PATHAK S , et al . Citus: Distributed postgresql for data-intensive applications [C ] // Proceedings of the 2021 International Conference on Management of Data . New York : ACM , 2021 : 2490 - 2502 .

KOPYTOV A . Sysbench: A system performance benchmark [EB/OL ] . [ 2024-09-23 ] . http://sysbench sourceforge net/ http://sysbenchsourceforgenet/ .

LEUTENEGGER S T , DIAS D , LEUTENEGGER S T , et al . A modeling study of the TPC-C benchmark [C ] // Proceedings of the 1993 ACM SIGMOD International Conference on Management of Data . New York : ACM , 1993 : 22 - 31 .

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

GaussDB：智能云原生分布式数据库

SKYLINE查询解析

计算机及现代通信中的多媒体技术