Abstract

Computing for Development as a research area focuses on the design and implementation of new information and communication technologies for social and economic development. Conventional computing solutions are often inappropriate for these environments due to several contextual factors – cost, lack of power, lack of bandwidth, literacy, language, user interface and other socio- cultural factors. In this talk, I will describe the broad set of research challenges in this space and then focus on two specific research challenges: (a) Low-cost connectivity solutions for the next billion. (b) Enhancing Web access in emerging regions under poor network connectivity.

To address the first challenge, I will describe our experiences with developing Wireless Rural Extensions (WiRE) and Hermes, two contrasting solutions to the connectivity problem. WiRE is a clean-slate, solar-powered rural connectivity architecture that aims to change the existing cellular connectivity model and provide focused, reliable high bandwidth connectivity to rural regions with no dependence on the power grid. Hermes, on the other hand, aims to provide a new data connectivity on top of the existing cellular voice connectivity channels which are extremely unpredictable and hard to model.

To address the Web access challenge, I will specifically describe a new Web architecture stack for developing regions which supports a low-bandwidth transport layer, a contextual Web caching layer, a content adaptation layer and an intermittency-aware Web application layer. Given the poor connectivity conditions, complexity of Web pages and high levels of network sharing, traditional networking solutions completely break down in these environments and we need to fundamentally rethink how to enhance Web access in these settings across all layers of the protocol stack.