Ad-hoc and MESH networks have been a topic for research in the last few years. It is now starting to be mature and we do see implementations in the various areas of interest. One important research area in MANETs is routing. There is no doubt that routing in MANETs is a challenging task, and has therefore received tremendous amount of attention from researchers all over the world. This has led to the development of many different routing protocols for the MANET, approaching the common challenges in different ways. These routing protocols are normally categorized as either re-active, pro-active or hybrid protocols. One of the pro-active routing protocols commonly used in a MANET is the Optimized LinkState Routing Protocol (OLSR).
OLSR is a routing protocol originally designed for mobile ad hoc networks, and has been standardized and implemented. However, the current standard needs to be enhanced in order to meet requirements for multi-homing, metric retrieval, metric-based routing and path calculation in a network with global Internet connectivity.
The work presented in this thesis is targeted at making specific improvement to OLSR, based on the existing UniK OLSRd implementation. These enhancements include a metric-based ramework, a new radio-aware airtime metric and finally a multi-homing solution with loadbalancing. The foundation for the two last mentioned mprovements is the metric-based framework. The requirements for these improvements were initiated during the EU-project referenced in the list of publications.
The enhancements that are designed, implemented and tested in this thesis, provide OLSR with improved capability to carry generic metric information in control messages. In addition they also make it possible to retrieve and use the airtime cross-layer metric for path calculation.
Finally, the tests performed in a Linux based test environment consisting of several laptops with the OLSR enhancements implemented in this thesis, show that there is an improvement in path selection when the radio aware airtime metric is used. Instead of always electing the path with the lowest hop count, a better path with lower airtime cost which considers both errors, speed and transmission data rates will be selected. As a result, better networkutilization with less path congestion has been achieved.