Especially in logistics, process instances often interact with their realworld environment during execution. This is challenging due to the fact that events from this environment are often heterogeneous, lack process instance information, and their import and visualisation in traditional process engines is not sufficiently supported. To address these challenges, we implemented GET Controller and UNICORN, two systems that together enable event-driven process execution and monitoring. Their application is shown for a logistics scenario.
Due to the separation of concerns principle [
Copyright c 2015 for this paper by its authors. Copying permitted for private and academic
purposes.
events relevant for logistics from various heterogeneous sources, and (ii) GET
Controller, which executes and monitors logistics process instances thereby considering events
provided by UNICORN. The interaction between these components is event-based, and
their communication follows the publish-subscribe paradigm. To this extent, process
models are encoded in a dialect of BPMN [
UNICORN5 is the system handling heterogeneous events in different formats. It is
implemented in Java 7 and can be accessed through a web service (SOAP) as well as
through a web-based user interface (UI) written using the Apache Wicket framework.
Events can be published to UNICORN in XML, Excel, CSV, and JSON formats. As a
prerequisite for handling events, the corresponding event types must be defined, either
via the UI or via the web service importing its XSD. In UNICORN, Esper [
To address logistics-specific requirements, UNICORN stores routes and its
waypoints (locations a vehicle has to pass), which are relevant in logistics-specific event
processing. In this way, we are able to determine the distances between the current
position of vehicles or events and their waypoints, to estimate remaining travel times, and
to filter traffic incidents in the vicinity of routes. Thus, UNICORN provides high-level,
transportation-related event notifications, such as expired deadlines, delays, or missing
connections to its subscribers [
The logistics-specific process execution and monitoring is done in GET Controller6. For the implementation, the Activiti7 engine has been adopted. Activiti is a BPMNbased open-source process engine, encoded in Java. Given a BPMN process model as input, it enacts the corresponding process instances. The embedded task handler (i) allows granted users to execute tasks in the UI, and (ii) monitors the status of tasks in execution (e.g., executing or completed). The task handler of GET Controller is illustrated in Figure 1. On the upper-left side, all running instances are shown in a list. In the lower-left corner, the events relevant for the currently selected process instance (case 54) are shown: a diverted flight and a strike in Frankfurt. Both are also visualized on the map in the middle of the screen. On the right hand side, we provide the tasks of the process including their current status. Their execution may be triggered by external events but can also be performed by the user, which is then done through the offered buttons or forms in the lower-right corner.
Following the requirements identified in the GET Service project for integrating
process and event processing [
features, GET Controller receives process-execution-relevant events published by UNICORN. They can be external events as well as events signalling the task execution from process clients, e.g., a driver loading goods. Generally, the subscription to events is derived from the annotations in the deployed process model. We distinguish between event subscriptions for (i) a task, (ii) a group of tasks, or (iii) the whole process. Furthermore, these annotations determine for how long subscriptions are valid, i.e. (i) from the enabling to the completion of a specified task, (ii) as long as at least one task in the given group is running, and (iii) along the whole process execution repectively. 3
GET Controller interacts with UNICORN as shown in Figure 2. The communication
comprises five steps:
1. GET Controller deploys a process model which is then used for registering
subscriptions to UNICORN. A route can be provided in addition to localise events
relevant along this route for the user. For instance, to determine the routes of trucks
or to pick up airfreight at Frankfurt airport.
2. The execution of activities can be done via GET Controller (as shown in the panel
on the upper-right corner of the screenshot in Figure 1). However, in case of
logistics scenarios, GET Controller often derives progress of activities from events
received by UNICORN. For instance, rules in UNICORN can be used to correlate
the movement of a truck via GPS position events and to evaluate whether it would
finally reach its destination as specified (based on information contained in the
annotated process model or the corresponding route). The same rules can be defined
for flight events or any other movement and activity.
3. Each time an event is published to UNICORN (either by GET Controller or other
event sources), its match to a subscription is evaluated. For example, users can
subscribe to strike warnings at Frankfurt airport for their transportation process.
This has to be done using the annotation mechanism for process models described
in [
For our demonstration, we used data stemming from a real-world diversion of an
aeroplane on the route from Nice to Frankfurt, landing in Brussels. We extracted the
data out of a set of examined 6,495 European flight tracks (1,765,172 events) gathered
in July 2013. Flights diverted in 49 cases (0.7%). These diversions were
automatically detected by the approach presented in [
In summary, we are able to execute and monitor single-modal and multi-modal transportations as sketched through the example above. The progress of the flight and trucks as well as all happenings at the landing airport, such as an announced strike, are monitored. Furthermore, once the prediction of a flight diversion is made, UNICORN automatically forwards the corresponding event notification to GET Controller. Acknowledgement. The research leading to these results is part of the GET Service project and has received funding from the European Commission under the 7th Framework Programme (FP7) for Research and Technological Development under grant agreement 2012-318275.