An Investigative Study of the INFANT-Haggle

Recently there has been a lot of interest in research on performance-challenged networks, which are typically characterized by intermittent connectivity, high mobility, and frequent network partitioning. Attempts to overcome the difficulties in this environment include opportunistic networking and the associated applications that support it. Routing and routing algorithms has been the focus of most research. A number of simulation projects on opportunistic networks routing algorithms have been produced using data from real mobility traces. These simulations are thought to be more realistic than simulations based on generic random mobility models. One of these simulation projects is the Haggle project, which is an implementation of the Pocket Switched Network for mobile computing devices. In this thesis, we study the first implementation of the Haggle architecture, referred to as INFANT-Haggle. We present our results of profiling INFANT-Haggle, whose main objectives were to increase understanding of INFANT-Haggle, find out the performance bottleneck, and attempt to port the engine on Nokia 770 using SQLite as a database engine. The tools used were the Eclipse Test & Performance Tools Platform for profiling, the Scratchbox cross compiler, and the Maemo SDK platform for porting on Nokia 770. We discuss the methods used and present our findings