Recent developments in computing devices such as miniaturization, increased autonomy and capacity of communications open new perspectives for applications. The ubiquitous or pervasive computing uses these new equipments to offer new types of applications. However, the realisation of software running on these environments raises a new set of challenges, such as building applications that can adapt themselves to their environment. This PhD thesis focuses on the autonomic computing field and offers an infrastructure for the execution of such applications.
This infrastructure proposes the use of a hierarchical service-oriented components runtime. We propose to enhance this execution platform with touchpoints to monitor and reconfigure applications during their execution. We also propose to add a mechanism to incorporate external events in the autonomic reasoning and a hierarchical architecture of autonomic managers. This architecture features the following properties: responsiveness, scalability and support simultaneous execution of multiple applications. We apply our proposal in the context of home automation and propose an implementation of our framework called Auto-Home. The implementation uses the open-source APACHE iPOJO framework. We have extended the iPOJO framework by adding the previously mentioned architectural elements of our infrastructure. The approach taken in this thesis was validated in the European project ANSO. The presentation will also mentionned my current work which deals with the construction of an Autonomic Services Selection framework. This framework aims to be used in the Auto-Home platform.The presentation will be concluded by drawing the perspectives of these works in the context of the Phoenix team and more generally in the LABRI laboratory.
Johann Bourcier, Laboratoire LIG, Université Joseph Fournier, Grenoble.
Talk
April 28, 2009 - 11:30 - 075, LaBRI
