NeTS: Small: Large-Scale Opportunistic Data Crowdsourcing and Dissemination in Device-to-Device (D2D) Networks

点击翻译按钮获取中文摘要

The vast majority of today's wireless communication systems operate in the microwave spectrum below 3 GHz, which is experiencing severe shortage and has become a crowded resource. To meet the 1000x growth challenge in mobile broadband traffic, the millimeter wave (mmWave) band, operating at frequencies between 20 and 300 GHz, has been identified for next-generation (5G) cellular systems. While the use of mmWave band addresses the pressing needs of more wireless spectrum, it brings a new set of unique technical challenges such as severe path loss and undesired coverage holes. To this end, Device-to-Device (D2D) networks are proposed to employ short-range wireless links to establish opportunistic connections between mobile users. In this project, the researchers will explore a diversity of application-oriented problems in D2D, culminating in the formulation of both new fundamental theories and advanced technologies that contribute to the development of next-generation mobile communication systems. This project will effectively stimulate multi-disciplinary collaboration across a broad spectrum of fields, including anthropology, communications, computer science, economics, public health, demography, and sociology. It will also effectively enrich courses by implementation and experimental activities, providing students with hands-on experience.The proposed research includes two research thrusts to design, implement, evaluate, and prototype new protocols and algorithms, in support of efficient data gathering and dissemination in D2D. First, a class of applications involve large-scale data gathering from mobile devices. Although crowdsourcing has been discussed in recent years, the marriage of crowdsourcing and D2D creates new, interesting research problems, due to the unique non-deterministic network paradigm. The researchers will investigate several dimensions in support of D2D-based crowdsourcing, including a competition-based participant recruitment scheme for delay-sensitive applications and an effective quest algorithm to deliver crowdsourcing requests. Second, efficient data dissemination is indispensable in many D2D applications. In contrast to the prior work that focuses on classical multicasting from a source to a given set of receivers, the researchers propose to investigate a unique and interesting problem where the receivers are not explicitly known. In such settings, a natural approach is to distribute data at some depositories, that further deliver the content to interested data consumers upon requests. Under this framework, the researchers will devise algorithms to choose optimal depositories for maximizing the total profit and develop new incentive schemes to enable efficient dissemination. Complementing these research thrusts is an experimental prototyping and validation track, with various design choices and alternatives experimentally studied, evaluated and refined.