Let me grab your App: Preliminary proof-of-concept design of opportunistic content augmentation

With the rate of mobile traffic growth eventually outpacing 4G technology upgrades, it appears crucial to explore new means to offload cellular networks. Exploitation of device-to-device opportunistic short range connectivity for offloading purposes is an intriguing and widely investigated direction, but might ultimately play a negligible practical role: in front of the huge overall content universe a user may select to access, it is indeed unlikely that a specific user request is found in a possibly small number of neighboring mobile devices. Focusing on a concrete use-case scenario, namely Apps download from Google's Android market, we discuss a preliminary design and implementation of CarpeDroid, a local coordination approach devised to proactively augment the network response to a user lookup (in our use-case, coming from the App Market server) with download opportunities for content locally stored in neighboring devices and likely to fall within the users' interest. Our approach aims at acting as a simple recommending system, locally operated on the end user device, and not requiring any interaction with centralized servers. Efficiency and “better than nothing” privacy protection is achieved through the exchange of Bloom filters devised to summarize content opportunistically available in neighboring devices.