Pervasive computing is on the horizon with ever shrinking device form factors and increasing computation power. Sensor devices such as those used to monitor temperature, pressure, humidity, flow, audio, video etc. allow acquisition of information regarding the physical environments, which can be utilized in actuation and visualization. Traditionally, sensing data is either stored locally or transported to a centralized location at the perimeter of the network. We argue that the information producers and consumers in a cyber physical system are likely to be distributed and with different capabilities. This necessitates a uniform abstraction that can hide device/platform dependent characteristics, intricacies of communication and networks, and utilize distributed heterogeneous storage among different devices. A distributed storage solution also has the benefit of fault tolerance in events of disasters, such as infrastructure failure or natural calamities, wherein remnants of information stored over the distributed storage system from the destroyed devices can be invaluable. In this paper, we propose a Distributed File System that is agnostic to platforms and operating systems of deeply embedded devices.
