Research On Business-user-oriented Process Modeling Method Supporting IoT Integration

Business Process Management(BPM)has a wide range of applications in intelligent telemedicine,logistics supply chain management,and other fields.Involving business users in the process modeling process can make BPM more flexible,and integrating event streams into business processes allows BPM to perceive changes in the physical world faster and more accurately.Due to the large,diverse,complex,and changeable business requirements in the enterprise,business users need to customize the IoT stream data according to business requirements during process modeling.However,the current business process modeling specification only supports access to pre-constructed events and does not support business users to define complex events that are personalized for IoT stream data.Constructing a process model that supports complex event processing requires higher requirements,therefore a business useroriented process modeling approach that supports IoT integration is needed.On the other hand,while supporting business users for personalized event customization,which brings flexibility to BPM,it also brings a greater load to the process engine to execute such business processes that integrate IoT event streams and personalize complex event processing.The resource consumption of the process engine is high and the execution efficiency is low.A process execution mechanism that supports the integration of diverse and complex event processing is required.Given the above problems,this paper studies the process modeling method for business users that supports IoT integration.The main work is as follows:(1)A business process model and modeling method that supports IoT integration for business users are proposed.First,in the business process model,a new class of"IoTService Task" node element is defined by extending the activity element in BPMN2.0 to access IoT services,and its custom attributes are added to store personalized event patterns.Secondly,to facilitate business users to customize business events of interest,guide business users to build models through wizard-style steps,and do not require business users to master professional programming capabilities,event description languages,or implementation language.(2)A process execution optimization mechanism that supports IoT event stream integration is proposed.Firstly,a business process execution support framework is proposed,which adopts the idea of separation of concerns,uses the business process engine and the stream processing engine to process the execution of business processes and event stream processing respectively,and communicates through the message middleware,thereby reducing the load of the process engine.Secondly,large-scale personalized event detection and the optimization algorithm are proposed by taking advantage of the phenomenon of repetition and shared subsequences in large-scale personalized events.By clustering the estimated execution time attributes of multiple IoTService Tasks,and then dividing and conquering,deduplicating and optimizing multiple personalized event patterns in each task cluster,the complexity of personalized event patterns in the cluster is reduced.Therefore,the repeated calculation of the stream processing engine is avoided,and the execution efficiency is improved.(3)Design and implement the prototype system.It mainly includes two parts:business process modeling tools and execution support tools.Business users build business processes that support IoT integration by business process modeling tools and complete the definition of personalized events for IoT event streams through wizardlike steps.After building a business process that supports IoT integration,deploy it to the engine in the execution support tool for execution.Finally,through the case modeling analysis in the field of intelligent telemedicine,the validity of the modeling method proposed in this paper is verified,and the execution optimization framework proposed in this paper can effectively reduce the execution load of the business process engine and improve the execution efficiency through experiments.