MIMOSA: context-aware adaptation for ubiquitous web access

The ubiquitous computing scenario is characterizedby heterogeneity of devices used to access services,and by frequent changes in the user’s context. Hence,adaptation according to the user’s context and the useddevices is necessary to allow mobile users to efficientlyexploit Internet-based services. In this paper, we present adistributed framework, named MIMOSA, that couples amiddleware for context-awareness with an intermediarybasedarchitecture for content adaptation. MIMOSA providesan effective and efficient solution for the adaptationof Internet services on the basis of a comprehensive notionof context, by means of techniques for aggregating contextdata from distributed sources, deriving complex contextualsituations from raw sensor data, evaluating adaptationpolicies, and solving possible conflicts. The middlewareallows programmers to modularly build complex adaptiveservices starting from simple ones, and includes tools forassisting the user in declaring her preferences, as well asmechanisms for detecting incorrect system behaviors dueto a wrong choice of adaptation policies. The effectivenessand efficiency of MIMOSA are shown through the developmentof a prototype adaptive service, and by extensiveexperimental evaluations.