Data Dissemination in Unified Dynamic Wireless Networks

TL;DR

This paper introduces efficient algorithms for broadcast problems in dynamic wireless networks, unifying various models and demonstrating the importance of carrier sensing for improved performance and robustness.

Contribution

It presents a general communication model encompassing multiple wireless network types and provides optimal algorithms for broadcast tasks in dynamic settings.

Findings

Improved bounds for broadcast in static networks.

Optimal local broadcasting algorithm in SINR model.

Carrier sensing significantly enhances network communication efficiency.

Abstract

We give efficient algorithms for the fundamental problems of Broadcast and Local Broadcast in dynamic wireless networks. We propose a general model of communication which captures and includes both fading models (like SINR) and graph-based models (such as quasi unit disc graphs, bounded-independence graphs, and protocol model). The only requirement is that the nodes can be embedded in a bounded growth quasi-metric, which is the weakest condition known to ensure distributed operability. Both the nodes and the links of the network are dynamic: nodes can come and go, while the signal strength on links can go up or down. The results improve some of the known bounds even in the static setting, including an optimal algorithm for local broadcasting in the SINR model, which is additionally uniform (independent of network size). An essential component is a procedure for balancing contention, which has potentially wide applicability. The results illustrate the importance of carrier sensing, a stock feature of wireless nodes today, which we encapsulate in primitives to better explore its uses and usefulness.