Large-scale wireless networks are projected to dominate the information technology sector in the future, giving rise to a new set of research problems on scalability. The main goal of this project is to develop an essential understanding of the impact of large scales on the performance of wireless networks. In particular, the project examines how the finiteness of resources (memory, computational power, etc.) at individual nodes affects the overall network performance. The developed understanding is then used to design a set of algorithms that support efficient operation of large-scale networks of nodes with very limited resources. The algorithmic aspect is particularly important given that some of widely considered algorithms require excessive resources at individual nodes and, hence, are not scalable. However, the project demonstrates the existence of algorithms that require only negligible resources but achieve comparable performance. Furthermore, the study reveals that completely new protocols are needed to support operation of large-scales wireless networks.
In contrast to the majority of earlier studies that examined either large networks with unlimited node resources or small networks with limited node resources, the focus of this project is on relationship between the network size and node resources. Most of the considered problems are impractical to be addressed experimentally due to a considerable cost of building large-scale prototypes. Moreover, even simulating such systems is often very difficult because of computational limitations. Thus, a comprehensive research agenda is based on an analytical framework that overcomes the difficulties imposed by large scales.
