Final Report Summary - INFLOW (Information Flow in Opportunistic Wireless Networks)
The project INFLOW (Information Flow in Opportunistic Wireless networks) aims to develop the methodology for the modeling and analysis of optimal information flow in future wireless networks, thus leading to the design of new network protocols and algorithms. Building on recent breakthroughs in network information theory, as well as advances on distributed systems, the objective consists in using the wireless opportunistic and cooperative principles as guiding vehicles to provide new results that push the theoretical performance limits of mobile and wireless networks. The proposed interdisciplinary approach draws on advanced mathematical models, founded on information theory, stochastic analysis, the methodology and techniques developed for the analysis of algorithms, and statistical physics inspired methods in order to study the emerging behavior of large-scale networks.
The project is organized in three work-packages: WP1 (develop stochastic models of optimal information flow in wireless networks), WP2 (design of new optimization-driven algorithms and protocols), and WP3 (implementation and evaluation).
In WP1, we presented results on a stochastic geometry model, in which we investigate the fundamental limitations of wireless networks, and derive cut-set bounds on the information theoretic capacity of wireless networks. Moreover, we studied the information propagation speed in multi-lane vehicle-to-vehicle networks such as roads or highways, with a focus on the impact of time-varying radio ranges, which have not been studied before in this context. We assessed the existence of a vehicle density threshold under which information propagates at the fastest vehicle speed and above which information propagates dramatically faster.
In WP2, we adopted principles from statistical physics to optimally solve real-life multi-scale network resource management problems. We developed new Message Passing algorithms for optimal utilization of cognitive radio networks, as well as for Belief-Propagation assisted scheduling for input queued switches. We also participated in the preparation of a survey intended to advertise the potential of the statistical physics approach in solving networking and communications problems.
In WP3, we provided specific test-cases that show a complete roadmap on how the statistical physics framework can lead to efficient solutions in real-world applications, including energy monitoring in the NITOS testbed to test message passing Wifi association and we collaborated with engineers for a hardware implementation of a Belief Propagation-based multicast switch. We employed the innovative Energy consumption Monitoring Framework, developed by ITI-CERTH, for the validation of distributed message-passing algorithms for optimal access point association, to maximize throughput, minimize energy, and study the trade-off, showing that our approach significantly improves the performance compared to the standard approach. We performed simulations which we compare with real-life experiments in NITOS.
Towards achieving the objectives of INFLOW, as regards research results and career development, the fellow was integrated with the group of Communications and Networks of ITI-CERTH, working on team research activities. His effective integration has also been strengthened by participation in other EU research projects. Thus, the project has been instrumental in both developing leadership and managerial abilities in pursuing independent research, as well as establishing several working research collaborations with leading experts in Greece and other European institutions.
The project is organized in three work-packages: WP1 (develop stochastic models of optimal information flow in wireless networks), WP2 (design of new optimization-driven algorithms and protocols), and WP3 (implementation and evaluation).
In WP1, we presented results on a stochastic geometry model, in which we investigate the fundamental limitations of wireless networks, and derive cut-set bounds on the information theoretic capacity of wireless networks. Moreover, we studied the information propagation speed in multi-lane vehicle-to-vehicle networks such as roads or highways, with a focus on the impact of time-varying radio ranges, which have not been studied before in this context. We assessed the existence of a vehicle density threshold under which information propagates at the fastest vehicle speed and above which information propagates dramatically faster.
In WP2, we adopted principles from statistical physics to optimally solve real-life multi-scale network resource management problems. We developed new Message Passing algorithms for optimal utilization of cognitive radio networks, as well as for Belief-Propagation assisted scheduling for input queued switches. We also participated in the preparation of a survey intended to advertise the potential of the statistical physics approach in solving networking and communications problems.
In WP3, we provided specific test-cases that show a complete roadmap on how the statistical physics framework can lead to efficient solutions in real-world applications, including energy monitoring in the NITOS testbed to test message passing Wifi association and we collaborated with engineers for a hardware implementation of a Belief Propagation-based multicast switch. We employed the innovative Energy consumption Monitoring Framework, developed by ITI-CERTH, for the validation of distributed message-passing algorithms for optimal access point association, to maximize throughput, minimize energy, and study the trade-off, showing that our approach significantly improves the performance compared to the standard approach. We performed simulations which we compare with real-life experiments in NITOS.
Towards achieving the objectives of INFLOW, as regards research results and career development, the fellow was integrated with the group of Communications and Networks of ITI-CERTH, working on team research activities. His effective integration has also been strengthened by participation in other EU research projects. Thus, the project has been instrumental in both developing leadership and managerial abilities in pursuing independent research, as well as establishing several working research collaborations with leading experts in Greece and other European institutions.