A Method For Slicing Distributed Business Processes Under Multiple Constraints And Its Implementation

In traditional Workflow and Business Process Management(BPM)technologies,the management and scheduling of business processes are handled by a single centralized BPM engine.However,in the context of cross-organizational workflows,a centralized BPM system is not the optimal choice due to issues such as data privacy requirements across different management domains,single point of failure,and performance bottlenecks.To better meet the needs of cross-organizational collaboration and improve system performance,a distributed approach to process management and execution can be used.To achieve the distributed operation of business processes,the key lies in slicing the business processes.By slicing,sub-processes can be deployed on different machines and run independently,realizing the distributed operation of the process and improving the efficiency of the process.Business process tasks in many fields have various business constraints,such as security constraints,time constraints,and resource constraints.The slicing of the process should meet the constraint requirements as much as possible in order to complete it correctly and efficiently.Therefore,this article proposes a multi-constraint process slicing method based on clustering algorithms for business processes with business constraints,and designs and implements a prototype system for business process slicing.The specific work includes the following three aspects:(1)A multi-constraint clustering slicing algorithm is proposed.This method first preprocesses the business process,grouping it based on the depth of the process tasks,and dividing it into clusters according to security constraints.Subtasks with the same security constraints are also added to the cluster.If a task has no security constraints,other business constraints are quantified,and the distance between clusters should be calculated using the clustering distance formula.The clustering result can be regarded as the process slicing result,ultimately achieving the business process slicing task.(2)To validate the slicing of business processes and their execution results,this study proposes a log-based slicing validation method.Firstly,the method adds a listening proxy class to collect execution logs of business process tasks,and then uses multiple middleware to transfer the logs to Elasticsearch for collection.The slicing results are validated and analyzed by designing validation rules.(3)Building on previous research on distributed business process systems,this article utilizes the open-source Flowable workflow software to design and implement a business process slicing system.The system provides a visual interface for users to set weights on business constraints,enabling the core function of slicing business processes under multiple constraint conditions.Additionally,the system also includes convenient features such as process file management,viewing,and validating slicing results.