Goal-based Workflow Adaptation for Role-based Resources in the Internet of Things

Date

Nov 28, 2017

Time

2:00 PM - 3:00 PM

Speaker

Dipl.-Medieninf. Steffen Huber

Affiliation

Hochschule Karlsruhe

Language

en

Main Topic

Informatik

Other Topics

Informatik

Description

In recent years, the Internet of Things (IoT) has increasingly received attention from the Business Process Management (BPM) community. The integration of sensors and actuators into Process-Aware Information Systems (PAIS) enables the collection of real-time data about physical properties and the direct manipulation of real-world objects. In a broader sense, IoT-aware workflows provide means for context-aware workflow execution involving virtual and physical entities. However, IoT-aware workflow management imposes new requirements on workflow modeling and execution that are outside the scope of current modeling languages and workflow management systems. Things in the IoT may vanish, appear or stay unknown during workflow execution, which renders their allocation as workflow resources infeasible at design time. Besides, capabilities of Things are often intended to be available only in a particular real-world context at runtime, e.g., a service robot inside a smart home should only operate at full speed, if there are no residents in direct proximity. Such contextual restrictions for the dynamic exposure of resource capabilities are not considered by current approaches in IoT resource management that use services for exposing device functionalities. With this work, we aim at providing the modeling and runtime support for defining such restrictions on workflow resources at design time and enabling the dynamic and context-sensitive runtime allocation of Things as workflow resources. Therefore, we contribute an ontology for the modeling of Things, Roles, capabilities, physical entities, and their context-sensitive interrelations. A Thing can play a specific Role only under certain contextual restrictions defined by Semantic Web Rule Language (SWRL) rules. At runtime, the existing relations between the individuals of the ontology represent the current state of interactions between the physical and the cyber world. Through the dynamic activation and deactivation of Roles at runtime, the behavior of a Thing can be adapted to the current physical context. Also, we allow workflow modelers to define the explicit goal of a workflow activity either by using semantic queries or by specifying high-level goals from a Tropos goal model. From a runtime perspective, we contribute the Semantic Access Layer (SAL) middleware to enable the transformation of activity goals into semantic queries as well as their execution on the ontology for role-based Things. The SAL enables the discovery of fitting Things, their allocation as workflow resources, the invocation of referenced IoT services, and the continuous monitoring of the allocated Things as part of the ontology.

Last modified: Nov 28, 2017, 8:51:33 AM