为了确保实施的正确性和一致性,需要对业务过程协同进行建模和分析,但现有的工作鲜有关注于从架构的视角和数据的层次来对其进行讨论。为此,首先提出了一种“代理端+盟主端”业务过程协同架构;然后通过对该协同架构进行分析,提出了一种业务过程协同的数据恢复通用模型;最后基于该通用模型,提出了数据恢复策略模型,以此为基础提出了五种数据恢复策略并采用随机Petri网对其进行建模与仿真。实验结果表明,采用最短队列随机恢复策略能够使得业务过程协同应用在协同中确保具有高的系统吞吐量、快的响应时间及低的拒绝率。
In order to guarantee the exactness and consistency of process implementation, to model and analyze business processes is necessary. However,the current research work rarely focuses on business process modeling and analysis from the view of architecture and the level of data. Firstly, this paper proposed a business process collaboration architecture called "stockholder end + agent end". Secondly, it proposed a general model for data recovery of business process collaboration by analyzing the collaboration architecture. Lastly, it proposed a data recovery strategy model based on this general model, and pro- posed five data recovery strategies based on this data recovery strategy model, then used stochastic Petri net to model and simu- late. The experimental results show that the shortest queue-random recovery strategy can ensure high system throughput, fast re- sponse time and low rate of using refusal in collaboration.