Realising management and composition of Self-Managed Cells in pervasive healthcare

Research in pervasive and autonomic computing focuses on supporting services for pervasive applications, but often ignores how such applications can be realised through the federation of autonomous entities. In this paper we propose a methodology for designing collaborations between autonomous components, using the Self-Managed Cell (SMC) framework. We focus on the structural, task-allocation and communication aspects of management interactions between SMCs. We propose a catalogue of architectural styles for SMC interactions, and a model for combining architectural styles in patterns of interactions that can be enforced by different SMCs in large collaborations. This allows us to specify the management of large-scale systems by composing management functions using architectural styles as building block abstractions. A scenario for a health monitoring application involving a number of SMCs is used throughout the paper to illustrate how complex structures can be thus built.