Context Awareness: An Experiment with Hoarding

Computer mobility allows people to use computers in varied and changing environments. This variability forces applications to adapt thus requiring awareness of the computational and physical environment (e.g. information about power management, network connections, synchronization opportunities, storage, computation, location-based services, etc.). An important application for mobility is hoarding, i.e. automatic file replication between devices. To be accurate and not obstructive to the user, the hoarding mechanism requires both context awareness (e.g. amount of usable storage) and estimation of future environment conditions (e.g. network connection, tasks to be performed by the user in the near future, etc.). However, making applications context-aware is hindered by the complexity of dealing with the large variety of different modules, sensors and service platforms, i.e. there is no middleware supporting such applications and their development in a uniform and integrated way. This paper presents the architecture for an environment awareness system (EAS) and how it applies to hoarding. EAS is a middleware component that acts as an intermediary between applications and all mechanisms that assess the surrounding environment. It lets applications query and combine environment properties in a standardized way. Crucial for the success of automatic file hoarding is the EAS’s capability of supporting environment prediction based on simple reasoning and pattern detection. Thus, applications may advise users accordingly or even make decisions on their behalf.