 |
|
|
|
 |
 |
 |
|
 |
|
 |
|
|
曲云尧,施伯乐.一个基于SQL的并发控制机制.软件学报,1995,6(10):582-592 |
一个基于SQL的并发控制机制 |
A CONCURRENCY CONTROL MECHANISM BASED ON SQL |
投稿时间:1994-04-23 修订日期:1994-11-22 |
DOI: |
中文关键词: 数据库,并发控制,幽灵问题,SQL语言,两段锁 |
英文关键词:Database,concurrency control,phantom, SQL language, two-phase locking. |
基金项目:本项目受到国家863高技术项目资助. |
|
摘要点击次数: 3565 |
全文下载次数: 3315 |
中文摘要: |
传统的读写事务模型(由read(x)和write(x)序列组成)不能使调度机制充分利用应用程序的语义信息对事务进行灵活调度,从而不能有效提高系统的并发度.本文根据SQL语言的操作语义,给出了基于SQL的事务模型.利用这种事务模型并结合2PL方法,设计了并发控制机制:Condition—locking.这个机制可以:(1)避免数据库中的幽灵(phantom)问题,(2)利用应用程序的语义信息和完整性约束提高系统的并发度,(3)减少发生死锁的机会.因此,这是一个实用的并发控制机制. |
英文摘要: |
By the standard transaction model(composed of read and write operations),the scheduler can not exploit the semantics of application programs,thus not enhancing concurrency effectively.In this paper,a new transaction model based on SQL language are presented, which can exploit the semantics of high-level operations.Depending on the the transaction model and Two-Phase Locking method,the authors design a concurrency control mechanism-Condition-Locking, which can,(1)prevent phantoms from database,(2)exploit semantic information of application programs and the knowledge of the integrity constraints of the database system to enhance. cononrrency,(3)reduce the chances of producing deadlock.Hence,it is a practical concurrency control mechanism. |
HTML 下载PDF全文 查看/发表评论 下载PDF阅读器 |
|
|
|
|
|
|
 |
|
|
|
|
 |
|
 |
|
 |
|