Enhancing the performance of ad hoc networking by lower layer design
Jarmo Prokkola · 2008
The research of early ad hoc-like networks, namely multi-hop packet radio networks, was mainly concentrated on lower layers (below network layer). Interestingly, the research of modern ad hoc networks has been mainly restricted to routing protocols. This is understandable, since routing is very challenging in such dynamic networks, but the drawback is that the lower layer models used in the studies are often highly simplified, giving inaccurate or even incorrect results. In addition, modern ad hoc network solutions are usually suboptimal because lower layers, not designed especially for ad hoc networking, are used. Thus, ad hoc networking performance, in general, can be notably enhanced by considering also the design of lower layers. The simple deployment and robustness make wireless ad hoc networks attractive for several applications (e.g., military, public authority, peer-to-peer civilian, and sensor networking), but the performance of the current solutions is typically not adequate. The focus of this work is on the effects of lower layer functionalities on the performance of ad hoc networks, while also taking into account the effects of upper layers (e.g., the effect of application traffic models). A CDMA (Code Division Multiple Access) based dual channel flat ad hoc network solution, incorporating cross-layering between all three lowest layers, is proposed and analyzed. The main element of this is the Bi-Code Channel Access (BCCA) method, in which a common code channel is used for broadcast information (e.g., route discovery), while a receiverspecific code channel is used for all directed transmissions. In addition, a new MAC (Medium Access Control) solution designed for BCCA is presented. Moreover, a novel network layer spreading code distribution (NSCD) method is presented. The advantage of these methods is that they are designed especially to be used in ad hoc networks. With an extensive set of case studies, it is shown that the presented methods outperform the typically used ad hoc network solutions (based on IEEE 802.11) in different kind of scenarios, environments, modeling concepts, and with different parameters. Detailed simulations are carried out in order to analyze the effects of different features at the lower layers, finding also interesting phenomena and dependencies between different layers. It is also shown that close attention should be paid to lower layer modeling even though the overall network performance would be in focus. In addition, various interesting features and behavior models regarding ad hoc networking are brought up.