计算机科学 ›› 2016, Vol. 43 ›› Issue (11): 280-283.doi: 10.11896/j.issn.1002-137X.2016.11.054

• 软件与数据库技术 • 上一篇    下一篇

基于冲突率预测的自适应并发控制算法

范璧健,庄毅   

  1. 南京航空航天大学计算机科学与技术学院 南京210016,南京航空航天大学计算机科学与技术学院 南京210016
  • 出版日期:2018-12-01 发布日期:2018-12-01
  • 基金资助:
    本文受国家自然科学基金青年科学基金项目(61202351),国家博士后基金项目(一等)(2011M500124)资助

Adaptive Concurrency Control Algorithm Based on Conflict-rate Prediction

FAN Bi-jian and ZHUANG Yi   

  • Online:2018-12-01 Published:2018-12-01

摘要: 并发控制算法能够保证数据库事务集并发执行的正确性和一致性。为了提高并发事务的执行效率,提出了一种基于冲突率预测的自适应并发控制算法(ACC-PRC)。该算法将并发控制过程分为信息收集和策略选择两个阶段。信息收集阶段利用先验事务队列保证事务执行的可串行化,并且利用循环冲突队列收集系统的事务执行状态。策略选择阶段在循环冲突队列上运用改进的加权移动平均法预测下一阶段冲突率,并根据双向阈值决策下一阶段的并发策略。所提算法在事务到达率较高时能保持良好的事务执行效率,同时能够准确及时地感知冲突率的变化。对比实验表明ACC-PRC算法的综合性能优于HCC算法和ADCC算法。

关键词: 并发控制,冲突率预测,策略选择

Abstract: Concurrency control algorithm can guarantee the correctness and consistency of the database transaction.In order to improve the efficiency of concurrent transactions,an adaptive concurrency control algorithm based on conflict-rate prediction(ACC-PRC) was proposed.The algorithm is divided into two stages:information collection and strategy selection.The information collection stage uses a priori transaction queue PTQ to guarantee the serializable execution of the transaction,and a cyclic conflict queue CQR is used to collect the transaction execution state of the system.The strategy selection stage uses the improved weighted moving average method to predict the next stage of conflict using the cyclic conflict queue,and then chooses appropriate concurrency strategies by bidirectional threshold.The algorithm maintains good transaction efficiency while the transaction arrival rate is relatively high.The results show that the integrate performance of ACC-PRC algorithm is better than that of the HCC algorithm and ADCC algorithm.

Key words: Concurrency control,Conflict-rate prediction,Strategy selection

[1] Xu Shu-ting,Sun Yong-qiang.Performance Study of Real-Time Multiversion Concurrency Control Protocols in Parallel Database Systems[J].Chinese J.Computers,2002,5(2):173-180(in Chinese) 徐淑颋,孙永强.并行数据库实时多版本并发控制协议性能研究 [J].计算机学报,2002,25(2):173-180
[2] Nystrom D,Nolin M,Teoanovio A,et al.Pessimistic concurrency control and versioning to support database pointers in real-time databases;proceedings of the Real-Time Systems[C]∥ 16th Euromicro Conference on Real-time Systems.2004,6:261-270
[3] Makni A,Bouaziz R,Gargouri F.Formal Verification of an Optimistic Concurrency Control Algorithm using SPIN[C]∥Proceedings of the Thirteenth International Symposium on Temporal Representation and Reasoning.2006:160-167
[4] Demsky B,Lam P.Views:Synthesizing Fine-Grained Concurrency Control [J].Acm Transactions on Software Engineering & Methodology,2013,22(1):139-176
[5] Sheikhan M,Rohani M,Ahmadlueil S.A neural-based concurrency control algorithm for database systems [J].Neural Computing & Applications,2013,22(1):161-174
[6] Sheikhan M,Ahmadlueil S.An intelligent hybrid optimistic/pessimistic concurrency control algorithm for centralized database systems using modified GSA-optimized ART neural model [J].Neural Computing & Applications,2013,23(6):1815-1829
[7] Chen Y R,Zhuang Y.An Adaptive Decision Concurrency Control Algorithm [J].Advanced Materials Research,2014,1046(1):512-515
[8] Liu Ya-qi,Xie Bo.Iron ore CIF price forecasting based on improved weighted moving average method[J].Journal of Shanghai Maritime University,2015(2):55-59(in Chinese) 刘雅琪,谢波.基于改进加权移动平均法的铁矿石到岸价格预测[J].上海海事大学学报,2015(2):55-59

No related articles found!
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!