一种适于周期性事务的二重调度策略  

A Dual Scheduling Strategy for Periodic Transaction Model

在线阅读下载全文

作  者:夏家莉[1] 王文乐[2] 曹重华[1] 韩增波[3] 

机构地区:[1]江西财经大学软件与通信工程学院,南昌330022 [2]江西师范大学软件学院,南昌330022 [3]中国软件与技术服务股份有限公司CEC移动技术实验室,北京102200

出  处:《小型微型计算机系统》2013年第12期2744-2748,共5页Journal of Chinese Computer Systems

基  金:国家自然科学基金项目(60763002)资助;江西省自然科学基金项目(2008GZS0021)资助

摘  要:支持替代的事务模型提供多条执行路径,提高了事务成功率.针对支持替代的事务模型,提出一种适于周期性事务的二重调度策略PT-DSS.外部调度中,PT-DSS采用固定优先级的可抢占调度策略,拒绝接纳不可调度作业,控制系统负载,以减少无效损耗,另外它还延迟放行与抢占作业相冲突的作业,避免无效的执行;内部调度中,PT-DSS尽可能选取与系统当前所有作业相容的可调度替代执行,不但保护了当前作业,也保护了低优先级作业.若无此类替代,则选择冲突最少的可调度替代,尽可能减小作业的重启率.实验证明,对于周期性事务,PT-DSS策略在提高事务成功率和CPU有效利用率等方面具有优越性.Transaction model provides multiple execution routes to improve the adaptability of the transaction itself, thereby increasing the transaction success rate. For the transaction model based on functional alternative, a new dual scheduling strategy for periodic transaction is proposed. The external scheduling of PT-DSS adopts the firm priority and preemptable strategy, and rejects those trans- actions that cant be scheduled, which controls the system load and decreases invalid loss. PT-DSS suspends those transactions which conflict with higher priority transaction, thereby decreases the transaction restart times. The internal scheduling of PT-DSS does its best to select the best schedulable alternative for executing. The best alternative compatibles with all the current active jobs. That not only protects the new job, but also secures those jobs with lower priority. If it is no the best alternative in the job's alternative set, PT- DSS would select the schedulable alternative whose conflicts with current active jobs is the least. Experiments proved that the dual scheduling strategy for the periodic transaction, PT-DSS can increase the successful-ratio of transaction and valid utilization of CPU better.

关 键 词:实时数据库 功能替代 周期性事务 二重调度 

分 类 号:TP18[自动化与计算机技术—控制理论与控制工程]

 

参考文献:

正在载入数据...

 

二级参考文献:

正在载入数据...

 

耦合文献:

正在载入数据...

 

引证文献:

正在载入数据...

 

二级引证文献:

正在载入数据...

 

同被引文献:

正在载入数据...

 

相关期刊文献:

正在载入数据...

相关的主题
相关的作者对象
相关的机构对象