Fascinating attempt to increase resilience.
Do you use a particular standard or description language to model the process? How is the interface between the model (guarantees, assumptions,…) and the CEP engine solved?
At the beginning of the project, we conceptualized the resilience mechanisms without a particular modelling language in mind and aimed for an approach that would be virtually independent of a particular language, as long as some kind of task annotations are supported. In our practical implementation, we used BPMN 2.0, since the filling line demonstrator already made use of Camunda.
The CEP engine is completely agnostic of the model: It just monitors the MQTT messages, detects deviations from the expected patterns and passes the ID of the faulty agent to the self-healing service. Within the RESPOND project, the CEP engine was realized by Software AG and their Cumulocity platform.
Fascinating attempt to increase resilience.
Do you use a particular standard or description language to model the process? How is the interface between the model (guarantees, assumptions,…) and the CEP engine solved?
Thank you for your questions!
At the beginning of the project, we conceptualized the resilience mechanisms without a particular modelling language in mind and aimed for an approach that would be virtually independent of a particular language, as long as some kind of task annotations are supported. In our practical implementation, we used BPMN 2.0, since the filling line demonstrator already made use of Camunda.
The CEP engine is completely agnostic of the model: It just monitors the MQTT messages, detects deviations from the expected patterns and passes the ID of the faulty agent to the self-healing service. Within the RESPOND project, the CEP engine was realized by Software AG and their Cumulocity platform.
Thanks for your clarification! I am looking forward to reading your paper!