Professor, Department of Computer Science, UC Santa Barbara; Associate Director, Center for Information Technology and Society; Media Arts and Technology Program; Technology Management Program; and Computer Engineering Program, UC Santa Barbara
Research Sample: Excerpt from Global Internet Symposium in 2004
The excerpt below is part of a paper published at the Global Internet Symposium in 2004. While I do not necessarily think it does a complete job of explaining the idea, there is at least one major take-away. First, computer scientists are starting to pay attention to both the impact of society on their work and the impact of their work on society. From the Transliteracies perspective, how then can technology affect and be affected by the application of reading?
We are evolving into a society with nearly constant access to the Internet and information. Currently access at the fringes of the network is provided using ad hoc networking and novel devices like PDAs and advanced cell phones. Originally conceptualized for military networks, ad hoc networking allows a connection to the Internet even when not in range of a wireless access point. This is achieved by connecting through other users who relay transmissions to those who are connected.
However, deploying a wireless ad hoc network in a commercial arena raises a number of issues. Intermittent service along with a wide range of node capabilities may result in a frequently changing environment. Since users are not yet accustomed to such unpredictable behavior, emphasis is given on providing the most efficient connectivity possible. Much of the work to date attempts to address this problem by leveraging the resources of other nodes, such as cooperative nearby users. However it is often neglected that in a commercial setting this may not always be the case since operational costs prohibit users from participating in any scheme without appropriate compensation. Such issues have a significant impact on how existing Internet applications are ported to ad hoc networks and on the kinds of new applications that might be created.
The goal of this paper is then to develop applications and services that are uniquely suited to ad hoc environments. “Coupons” is a scheme that expands on the basic idea of information dissemination in ad hoc networks by investigating incentives for collaboration. Cooperation is stimulated by adopting a basic pyramid scheme where users are awarded credits as they pass a received piece of information, or a “coupon”, to other people. Using a simple flooding mechanism, as one user forwards the coupon to another, a pyramid is built with the users residing at the top receiving more benefit than those at the bottom.
We are particularly interested in examining the network costs that are usually associated with flooding schemes. Although flooding can be effective in terms of distributing data to a large set of nodes, it is known to suffer from several problems such as high message redundancy.