Perseos - Pervasive self-adaptive software systems
This project targets software systems composed of many heterogeneous devices, which communicate each other through distributed and wireless interfaces. Such systems reveal emergent behaviors that may greatly affect system quality and reliability.Researchers have for some time now recognized the benefits of self-adaptive systems, particularly as the size and complexity increase beyond the ability of humans to respond manually in a coherent and timely manner to changes in environment, failures of components, and attacks from intruders.In this project we study the problem of coordinating heterogeneous self-adaptive mechanisms to solve conflicts and avoid chaos. We define a conceptual framework for self-adaptive complex software systems, and propose a coherent and efficient approach, to resolve conflicts among heterogeneous self-adaptive mechanisms. The conceptual framework provides the functionality to coordinate self-adaptive mechanisms that operate at different levels, within heterogeneous devices and with distinct roles and goals.We define models and meta-models as the conceptual foundation of cross-layer adaptation mechanisms. In our solution, we propose suitable models to represent the different adaptation activities, and meta-models to govern the interactions between adaptation activities. We define analysis techniques for meta-models and models to detect inconsistencies. We study a unified management protocol that will embody an appropriate common management interface for adaptable systems. Such a protocol allows basic interoperability between systems and controllers.