Designing mobile applications for challenged network environments necessitates new abstractions that target deployment architects, nondevelopers who are charged with adapting an ensemble of off-theshelf software to a deployment context. Data transfer is integral to mobile application design and deployments have inherent and contextual requirements that determine what data should be transferred and when. In this paper, we investigate building mobile applications in challenged network environments by focusing on abstractions to support disconnected environments and areas of sparse heterogeneous connectivity. We explore and characterize various methods of transmitting data using: existing synchronization tools, peer-to-peer communication, and sparse networks. We also introduce a new software tool called ODK Submit to help streamline application customization to challenged network environments.

No abstract

No abstract

Cellular networks are often the first telecommunications infrastructure in developing regions. By studying cellular net-work traffic, researchers gain insight into how technologies can be used to access services critical to further development. In this work, we approach a cellular traffic dataset provided by Orange in Cote d'Ivoire with the goal of identifying distinctions between urban and rural use of cellular infrastructure. We report on a number of interesting differences between urban and rural usage of cellular infrastructure. For instance, 70% of calls that originate in rural areas occur within the vicinity of the same antenna, whereas the same is true for only 23% of calls with urban origin. We are compelled to conclude that development efforts for rural areas might be implemented differently from development efforts in urban areas based on divergent use of current cellular infrastructure.

Tribal areas continue to be some of the most under-serviced places in the U.S. with respect to broadband coverage. Despite tribal and FCC interests in addressing this issue, there has never been a network-based characterization on how existing broadband services in these are areas are used. We present one of the first of such characterizations by analyzing traffic from the Tribal Digital Village (TDV) network, a tribal operated network that provides initial Internet services to 576 facilities in rural San Diego county. Our study reveals that the most requested web application in the network is Instagram, which has not yet been characterized in a community setting. Overall, we identify a high locality of interest with respect to content and social connectivity: 27% of the 150,368 unique media objects circulated in TDV Instagram traffic were produced by TDV users; local content creators have 8.2× more engagement with their media on the TDV network than non-local content creators; 26.5% of the 7.9 million media downloads were downloaded by multiple TDV network users; and on average, 42.6% of a user's 1-hop social neighborhood is comprised of users from the same reservation.