| In certain scenarios such as disaster response, hastily formed ad hoc networks often have limited resources, such as low bandwidth and processing capability. A sudden increase in data transmission in such resource-constrained networks frequently results in congestion, causing increased delay in sending data.;The data transmitted in these ad hoc networks may be tied to critical events, which need to be delivered quickly so decision makers can address high priority issues as soon as they occur. Increased delays may cause slow response times and poor performance. Thus, the research problem we address is: How do we prioritize and ensure timely delivery of critical data in a congested ad hoc network..;In this thesis, we present Triage, a framework that enables real-time data transmission prioritization based on changes in sender context information. Existing quality of service mechanisms prioritize data transmission based on data content type. Our Triage framework additionally prioritizes data based on sender context as well as role, and enables relative importance of context values to be configured dynamically at run time.;We analyze the performance of Triage across several different types of network topologies using both a simulation environment and a sensor network testbed. We also demonstrate empirically how critical data transmission in our Triage framework outperforms baseline data transmission with higher throughput and lower packet delay, without significantly affecting fairness measurements. |
