Utilizing Role-based Models for Distributed On-Demand Service Composition
In future smart computing environments, e.g., smart home, smart mobility, smart grid or pervasive health, autonomously designed subsystems are dynamically interconnected to smart service systems to develop their full potential. Other than existing approaches from the domain of Self-organizing Software Systems, where the application structure emerges from internal constraints or mechanisms, smart service systems require complex application models and a coordinating instance for their setup. In this paper, we utilize role-based application models, by means of a collaboration specification, for automated, on-demand service composition in decentralized infrastructures, i.e., infrastructures without a predefined central coordinator. We propose a two-phase development methodology in which a collaboration designer specifies the overall collaboration including its abstract functionality and several (other) developers complement it with its concrete performance. We verify our approach using a case study which rigorously follows the proposed development methodology and demonstrate that role-based application models provide sufficient information to guide the discovery and composition process for smart service systems.
Mon 3 Apr
|13:30 - 14:00|
|14:00 - 14:30|
|14:30 - 15:00|