NeTS: Large: Collaborative Research: Foundations for Network Cooperation at Signal-scale

点击翻译按钮获取中文摘要

Current wireless network architectures are based on interference avoidance, which advocates eliminating simultaneous transmissions to avoid collisions at the receivers. However, this design principle is largely an artifact of design simplification. In contrast, if neighboring nodes pool their resources, and cooperate in their signal transmissions, the network could turn interference to its advantage for potentially many-fold increase in network capacity. This cooperative viewpoint necessitates revisiting networking research?s foundations, which are being addressed with a two-part strategy:1. Network-centric Cooperative Signal Design: Cooperative signaling injects "network" into signal design, thereby breaking conventional boundaries. Nodes have to understand how their transmissions will be perceived, decoded, suppressed, cancelled, enhanced or forwarded by other nodes. This fundamental shift in signal design (from conventional point-to-point PHYsical layer) is being addressed by developing capacity bounds, distributed codes and messaging protocols for scalable cooperation. 2. Signal-centric Cooperative Network Design: The converse to network-inspired signal design is ?signal-centric? network design. Network resource allocation and control have to be cognizant of signal-level interactions between groups of cooperating nodes,? breaking conventional design boundaries in network protocol design. This foundational change is leading to completely new problem formulations in scheduling, routing and protocol design to harness cooperative signal-scale gains.The project goals are nothing short of rewriting networking fundamentals. By questioning the basic design paradigms, we expect the project will impact research in multiple communities. Our experiment codes and measurements will be open-sourced as community asset. We will also establish a unique inter-university education program including joint advising and collaborative experiments.