| Extended Abstract
Known for their flexibility and dynamism, Mobile Ad hoc NETworks (MANETs) are being deployed in a variety of environments – ranging from on-demand networks in response to sudden events (e.g., a crisis,) to pre-planned strategic networks (e.g., military networks). However, the very same features (flexibility, dynamism) that render MANETs powerful also render the design and analyses of MANETs a challenge. As an example, the fact that MANETs have the flexibility to respond to network fluctuations and re-organize themselves accordingly via MANET routing, renders both the design of appropriate MANET routing protocols as well as maintenance (via network management) of MANETs, a challenge. To further challenge the matter, since resources (in terms of bandwidth) in MANETs are scarce, the MANET routing protocol not only needs to be adaptive to the uncertainties (random mobility, link changes) in the network but should also not contribute much to the bandwidth overheads.
In fact, we observe that an important component of MANETs is the presence of a stable routing algorithm that assures connectivity and reachability amidst the stochastic behavior of the network. To this end, a large amount of research has been devoted to designing MANET routing protocols that can adapt to dynamics and unpredictability that plague MANET. This has resulted in a myriad of protocols (e.g., TORA, DSR, AODV, OLSR, …). From this list, OLSR and its variants, seems to have been chosen for a number of military networks (e.g., FCS, WIN-T, TMOS). While a tremendous amount of effort has gone into specifying the OLSR protocol requirements, surprisingly very little has been done until very recently on analyzing the impact of OLSR parameters on the performance of ad-hoc networks. More specifically, since MANETs are sensitive to bandwidth usage, very little study on the impact of the various protocol messages (e.g., Hello messaging, TC messaging) and timers, exist to-date.
This work therefore has aimed to fill in the above gap by conducting a detailed analysis of OLSR, and in addition, to provide a link between the outcome of the analysis and adaptive network management systems, that work together to achieve high MANET efficiency. More specifically, we apply the vastly researched body of discrete event simulation (DES) techniques to analyze and study the behavior of OLSR and link the outcome of the analysis with an adaptive policy-based network management paradigm |