Technical Program

The talk will focus on coded caching and the interesting salient features that arise when caching is applied in advanced PHY wireless scenarios. Focusing on recently revealed deep connections between caching and fundamental primitives of wireless networks, such as feedback, multiple antennas, and topology, we will seek to answer different questions that arise, and which include: -      When can super-small caches have a substantial impact? -      What is the relationship between caching and CSIT-type feedback? (it turns out that this is a deep relationship, of a synergistic as well as competing nature) -      How can caching data the night before, allow for the (paradoxical) ability to "buffer" CSI? -      How can non-linearity/non-separability (which can reside in wireless settings) be used to boost wireless caching? We will also briefly look at how advanced PHY techniques can be used to ameliorate some of the bottlenecks of coded caching, exploring for example how interference enhancement techniques can be used to ameliorate the dreaded "worst-user" effect of coded caching.

This talk is aimed to discuss the critical and essential importance of spatial models for accurate system-level analysis and optimization of emerging ultra-dense and heterogeneous cellular networks. With the aid of stochastic geometry and point process tools, new mathematical methodologies for system-level analysis and optimization will be illustrated. In addition, their application to emerging cellular network concepts will be discussed and validated with the aid of empirical data from publicly available databases.

SDN in conjunction with NFV offer a novel opportunity for the joint management of network, compute and storage resources. These technological advances can revolutionise the way we design and manage wireless systems, and eventually enable the operators to satisfy the increasingly stringent requirements of emerging services. We will discuss the latest developments in this area, while focusing on the particularly important application of mobile video delivery in 5G HetNets. A suite of network-aware video caching solutions will be presented, that can provably improve the users' satisfaction and also reduce the network operating expenditures. We will conclude with key open questions regarding the caching and delivery of (video) content in these SDN-enabled wireless systems. 

Panel with: o Prof. Petros Elia (Eurecom, France) o Dr. Anastasios Giovanidis (UPMC & CNRS, France) o Prof. George Iosifidis (Trinity College Dublin, Ireland) o Dr. Georgios Paschos (Huawei, France)

Stochastic Geometry allows one to statistically represent contention for spectrum in large wireless networks. It is based on the computation of spatial averages over a snapshot of the network state, which are not easy to interconnect with the time averages of queuing dynamics. The aim of this lecture is to survey ongoing research on the mathematical interconnection of Stochastic Geometry and Queuing Theory in this wireless networking context. The survey will cover progress on the analysis of a fundamental model, which is the spatial queuing process where user service rates are determined by the interference field. It will also revisit the classical problem of multi-hop relaying queues in mobile networks, for which a new mean-field analysis will be discussed.

Resource allocation in interference networks is a timeless and important challenge. In future generations of mobile networks, heterogeneous and conflicting performance criteria are introduced leading to multi-objective programming problems. When efficiency is optimized often fractional programming - a well established technique - can be applied. However, in interference networks, the fractional objective functions are in general not suitable for fractional programming. In this invited talk, the combination of fractional programming with monotonic optimization (to achieve global optimality with high complexity) and with sequential convex programming (to achieve local optimality with lower complexity) are proposed. Their applications are illustrated with distributed and centralized power control in 5G and beyond.

Performance analysis of wireless local area networks is often done by modeling the evolving interactions between nodes as coupled Markov chains. The evolution of the empirical distribution of nodes across the back-off states, in the limit of a large number of nodes and under a suitable scaling of the back-off parameters, is characterised by an ODE called the McKean-Vlasov equation. The classical fixed-point analysis applies when the ODE has a globally asymptotically stable equilibrium. In more general cases, when the ODE has several stable limit sets, one needs a finer analysis. The talk will provide an overview of the Freidlin-Wentzell theory and its applicability to the case of multiple stable limit sets. The talk will also highlight an interesting issue of short-term unfairness, and an approach to predict it, when the back-off parameters are not scaled.

In recent years, significant progress has been made on determining the fundamental limits of cache-aided networks. This talk will present some of these new information-theoretic results. The main focus will be on new coding opportunities that arise in cache-aided networks. As we will see, such opportunities arise in particular in heterogenous networks where some receivers are inherently weaker than others or when the communication needs to be kept secret.

The proliferation of mobile internet access poses new challenges to wireless service providers as the capacity growth of their networks cannot cope with the rate of increase of mobile wireless traffic. Alternate means are considered to deal with the  excessive traffic demand, that exploit  the proliferation of wireless networks in unlicensed parts of the spectrum as well as of handheld devices with multiple radio interfaces. Traffic off-loading from the cellular network to a wifi access point is possible for mobile users with wireless interfaces for both networks. We will present schemes that motivate operators, access point owners and users to cooperate in order to maximize use of available capacity in the different networks; the schemes are  based on double auction mechanisms. In an alternate approach, a mobile user may gain internet access when another user with  cellular internet connection  is willing to relay its traffic received through a direct link between the users. We will present incentives mechanisms that facilitate the creation of such User Provided Networks in a way that all participants gain in terms of access capacity as well as energy consumption. Finally will present a design and implementation of a novel cloud-controlled UPN that employs software defined networking support on mobile terminals, to dynamically apply data forwarding policies with adaptive flow-control.

In this talk, we explain how the new idea of agent action encoding can be used to create coordination in distributed wireless networks. In particular, the concept of power modulation and a particular scheme to implement it are presented; the scheme beats classical distributed power allocation schemes such as the iterative water-filling algorithm while relying on the same knowledge. More generally, the concept of coded power control consists in embedding coordination information (e.g., about the channel state) in a sequence of transmit power levels. It is shown how information theory can be exploited to determine the limiting performance of coded power control. The framework developed applies in fact to the general problem of characterizing the limiting performance of a network with multiple agents who have partial information.

With centralized processing, cooperative radio, real-time cloud computing, and clean infrastructure, Cloud Radio Access Network (CRAN) is a "future proof" solution to sustain the mobile data explosion in future wireless networks. The technology holds great potential in enhancing future wireless systems with necessary capability to accommodate unprecedented traffic demand. However, cloud wireless systems inevitably encounter scalability issues in terms of computational and implementation complexities. This talk discusses the challenges and recent developments in technologies that potentially address the scalability issues of CRANs. In particular, I will focus on a randomized Gaussian message passing algorithm to achieve perfect scalability and convergence in the PHY layer of CRANs. 

Most stochastic geometry-based analyses of wireless networks focus on spatial (or ensemble) averages, such as the success probability or rate of the typical link. However, in a realization of the point process describing the network, no link is typical. In most networks, it is desirable that the per-link performances are concentrated around their mean (the performance of the typical link). Hence an important question is whether the performances of most links are close to that of the typical link or whether there is a wide disparity. We present a mathematical framework based on the so-called meta distribution that permits an analysis of the per-link performance distribution, rather than just its average. The framework is applied to Poisson bipolar and cellular networks (with and without power control), and D2D communications. We show how the deviations from the typical link performance can be minimized. Lastly, we introduce and analyze the spatial outage capacity, which is the maximum density of links in a network that satisfy an outage constraint.

Wireless communication networks composed of devices that can harvest energy from nature will lead to the green future of wireless, as energy harvesting offers the possibility of perpetual network operation without adverse effects on the environment. By developing effective and robust communication techniques to be used under energy harvesting conditions, some of the communication devices in a heterogeneous network can even be taken off the grid. Energy harvesting brings new considerations to system level design of wireless communication networks, leading to new insights. These include randomness and intermittency of available energy, as well as additional system issues to be concerned about such as energy storage capacity and processing complexity. Additionally, one can now envision such devices engaging in energy cooperation by powering one another to improve overall network performance. The goal of this talk is to furnish the audience with fundamental design principles of energy harvesting and energy cooperating wireless communication networks which is an emerging research area.