Evaluating the Joint Use of Alternate Internet Delivery Models in Deterministic Rural Networks

Abstract

Rural residents are often faced with the problem of expensive yet unreliable Internet connectivity. This coupled with the sparse population/business density, poses a detriment to the general development of such communities. In mitigating this disparity, some rural areas have adopted alternate Internet models to improve access to Internet services within their community. Such alternate models include communal access points, delay tolerant networks, and resource sharing networks. Thus, this work investigates the joint use of the aforementioned alternate Internet models for improving connectivity. Further, this work explores methods of routing time-dependent data in quasi-deterministic scenarios within these challenged rural environments. We formulate the routing problem as a maximum unsplittable multicommodity flow problem and propose an approximation algorithm. Using data representative of a rural Appalachian region, we evaluate our approximation algorithm and find that it performs close to optimal. We also propose a routing protocol based on our approximation algorithm. Using the same data, the results show that the proposed routing protocol performs comparably well with state-of-the-art routing protocols in delivering messages under time constraints. Additionally, the proposed routing protocol offers an overhead cost that is at least 3 times lower, on average, than other well-known routing protocols.