Sibren Isaacman

C-LINK

The modern internet serves as a hub of information and a vital communication pathway throughout the

developed world. In North America, for instance, over 73% of the population has access (either at home, work, or in public areas) to the internet’s resources and including only people over the age of 18 raises this further, to 80% [4]. This is in stark contrast to the world average of 20%, with some areas of the world seeing internet penetration rates as low as 5% [3]. Closing this “digital divide” will become ever more crucial in allowing make themselves competitive on the world stage [1]. Bringing technology to developing regions has even been named one of the “Millennium Development Goals” by the United Nations [2]. We have developed C-LINK, a system for bringing

internet connectivity to developing regions.

 

Despite the latency involved in receiving data, there are applications that can be built on top of delay

tolerant networks (DTNs). In previous work [5], we have shown that using a combination of collaborative caching and predictive prefetching, we can increase the usability of computer networks in remote areas. In collaborative caching, computers within a village are able to access webpages stored on other connected machines. The second technique, predictive prefetching, delivers webpages not yet requested to a client or village on the assumption that they will be requested in the near future. Both of these techniques, even with naïve implementations, result in lower miss rates of fetched pages compared to traditional caching strategies with no prefetching. Our realization of this techniques is our C-LINK system, named for the components of the system (City Fetch Engine, Load Manager,

Interface, Notifier, and Kioskdaemon).

C-LINK will provide the secondary school students and teachers of Cinco Pinos, Nicaragua with an

internet connection. Basic computer classes offered at the local library with the assistance of a local NGO, Comisión de Desarrollo Rural, and Peace Corps volunteers, who teach at the local high school, will ensure that students and teachers make use of the connection with an eye towards educational applications in the future. However, this deployment would also allow the collection of real-world data. By deploying C-LINK in Cinco Pinos, we will be able to obtain feedback from the intended users via informal interviews, direct observation, and semi-structured surveys. In addition to data such as hit rates, latencies and qualitative data on user experiences, a deployment in Cinco Pinos will allow us to collect usage traces that accurately reflect the data in which users in developing regions are interested. guarantees.

 

Furture development of the C-LINK system will include explorations of support for interactivity, security issues. Further, near-future research goals inclue examining how correlated user mobility can be leveraged to provide access to mobile users even when the users are out of range of the kiosk.

 

REFERENCES

[1] L. Osin. Computers in education in developing countries:Why and how? Education and Tech. Series, 3(1), 1998.

[2] United Nations. Millennium Development Goals. 2008. http://www.un.org/millenniumgoals/.

[3] Miniwatts Marketing Group. World Internet Usage Statistics and World Population Stats, May 2008.

http://www.internetworldstats.com/stats.htm.

[4]United States Census Bureau. Table 1127. Internet Access and Usage and Online Service Usage: 2006.

http://www.census.gov/compendia/statab/tables/08s1127.pdf

[5] S. Isaacman and M. Martonosi. Potential for collaborative caching in largely-disconnected villages. Proc. of the 2008 ACM workshop on Wireless networks and systems for developing regions, September 2008.