华南理工大学学报(自然科学版) ›› 2009, Vol. 37 ›› Issue (4): 24-30.

• 计算机科学与技术 • 上一篇    下一篇

基于数据流优化的BPEL流程分割方法

翟岩龙1  宿红毅1  张晗2  战守义1   

  1. 1. 北京理工大学 计算机科学技术学院, 北京 100081;  2. 导弹控制系统仿真国防科技重点实验室 北京仿真中心, 北京 100854
  • 收稿日期:2008-03-13 修回日期:2008-06-19 出版日期:2009-04-25 发布日期:2009-04-25
  • 通信作者: 翟岩龙(1981-),男,博士生,主要从事面向服务体系结构研究. E-mail:zhaiyl@bit.edu.cn
  • 作者简介:翟岩龙(1981-),男,博士生,主要从事面向服务体系结构研究.
  • 基金资助:

    “十一五”武器装备预研基金项目(513040102)

BPEL Process Decomposition Approach Based on Data Flow Optimization

Zhai Yan-long1  Su Hong-yi1  Zhang Han2  Zhan Shou-yi1   

  1. 1. School of Computer Science and Technology, Beijing Institute of Technology, Beijing 100081, China; 2. Key Laboratory of National Defense Science and Technology of Missile Control System Simulation, Beijing Simulation Center, Beijing 100854, China
  • Received:2008-03-13 Revised:2008-06-19 Online:2009-04-25 Published:2009-04-25
  • Contact: 翟岩龙(1981-),男,博士生,主要从事面向服务体系结构研究. E-mail:zhaiyl@bit.edu.cn
  • About author:翟岩龙(1981-),男,博士生,主要从事面向服务体系结构研究.
  • Supported by:

    “十一五”武器装备预研基金项目(513040102)

摘要: 集中式执行的BPEL流程性能不高并且不能满足动态流程管理的需求.为此,文中提出了一种基于数据流优化的BPEL流程分割方法,以支持BPEL流程的分散化协调.首先将BPEL流程变换成程序流图(PFG),然后提出数据流方程对PFG进行数据流分析,并扩展复制传播算法优化PFG的数据流,最后按照流程中的角色分割BPEL流程.实验结果表明,分割后流程的响应时间缩短了30%~50%,算法复杂度由O(en)降为O(n2).

关键词: 服务组合, 流程分割, 数据流分析, 分散化协调

Abstract:

As the centralized BPEL process can not provide sufficient supports for the dynamic process management, a BPEL process decomposition approach based on data flow optimization is proposed to support the decentralized coordination. In the approach, a BPEL process is represented as a program flow graph ( PFG), and data flow equations are deduced to analyze the data flow information. Moreover, the copy propagation algorithm is extended to optimize the data flow in the PFG, and the BPEL process is partitioned according to the optimized PFG. Experimental results show that, after the decomposition, the response time cost reduces by about 30% - 50% , and the computational complexity decreases from O( en) to O( n2 ).

Key words: service composition, process decomposition, data flow analysis, decentralized coordination