Embedding MANETs in the Real World

We believe that research on ad hoc networking so far has focused on too broad a range of scenarios and optimization issues and we argue that limiting the settings to bare bones services and small networks will allow for quick MANET deployment. Especially if these solutions interoperate easily with the Internet and offer the possibility to grow the MANET networks from there. Starting at this end will also help us to identify and possibly enlarge the "ad hoc horizon", beyond which ad hoc networking is no longer useful. In this paper we report on our experiences from five years of real world ad hoc network experimentation. Our approach is pragmatic and includes several implementations and a testbed software package. We highlight issues that we have encountered and identify crucial problems and potential dangers that deserves special attention in order to make ad hoc networking usable.