A Novel Collaborative Cognitive Dynamic Network Architecture

📝 Abstract

Increasing mobile data demands in current cellular networks and proliferation of advanced handheld devices have given rise to a new generation of dynamic network architectures (DNAs). In a DNA, users share their connectivities and act as access points providing Internet connections for others without additional network infrastructure cost. A large number of users and their dynamic connections make DNA highly adaptive to variations in the network and suitable for low cost ubiquitous Internet connectivity. In this article, we propose a novel collaborative cognitive dynamic network architecture (CDNA) which incorporates cognitive capabilities to exploit underutilized spectrum in a more flexible and intelligent way. The design principles of CDNA are perfectly aligned to the functionality requirements of future 5G wireless networks such as energy and spectrum efficiency, scalability, dynamic reconfigurability, support for multi-hop communications, infrastructure sharing, and multi-operator cooperation. A case study with a new resource allocation problem enabled by CDNA is conducted using matching theory with pricing to illustrate the potential benefits of CDNA for users and operators, tackle user associations for data and spectrum trading with low complexity, and enable self-organizing capabilities. Finally, possible challenges and future research directions are given.

💡 Analysis

Increasing mobile data demands in current cellular networks and proliferation of advanced handheld devices have given rise to a new generation of dynamic network architectures (DNAs). In a DNA, users share their connectivities and act as access points providing Internet connections for others without additional network infrastructure cost. A large number of users and their dynamic connections make DNA highly adaptive to variations in the network and suitable for low cost ubiquitous Internet connectivity. In this article, we propose a novel collaborative cognitive dynamic network architecture (CDNA) which incorporates cognitive capabilities to exploit underutilized spectrum in a more flexible and intelligent way. The design principles of CDNA are perfectly aligned to the functionality requirements of future 5G wireless networks such as energy and spectrum efficiency, scalability, dynamic reconfigurability, support for multi-hop communications, infrastructure sharing, and multi-operator cooperation. A case study with a new resource allocation problem enabled by CDNA is conducted using matching theory with pricing to illustrate the potential benefits of CDNA for users and operators, tackle user associations for data and spectrum trading with low complexity, and enable self-organizing capabilities. Finally, possible challenges and future research directions are given.

📄 Content

A Novel Collaborative Cognitive Dynamic Network Architecture Beatriz Lorenzo, F. Javier Gonzalez-Castano, Yuguang Fang

Abstract—Increasing mobile data demands in current cellular networks and proliferation of advanced handheld devices have given rise to a new generation of dynamic network architectures (DNAs). In a DNA, users share their connectivities and act as access points providing Internet connections for others without additional network infrastructure cost. A large number of users and their dynamic connections make DNA highly adaptive to variations in the network and suitable for low cost ubiquitous Internet connectivity. In this article, we propose a novel collaborative cognitive dynamic network architecture (CDNA) which incorporates cognitive capabilities to exploit underutilized spectrum in a more flexible and intelligent way. The design principles of CDNA are perfectly aligned to the functionality requirements of future 5G wireless networks such as energy and spectrum efficiency, scalability, dynamic reconfigurability, support for multi-hop communications, infrastructure sharing, and multi-operator cooperation. A case study with a new resource allocation problem enabled by CDNA is conducted using matching theory with pricing to illustrate the potential benefits of CDNA for users and operators, tackle user associations for data and spectrum trading with low complexity, and enable self-organizing capabilities. Finally, possible challenges and future research directions are given. Keywords—Future network architecture, cognitive radio, network reconfigurability, matching theory, data and spectrum trading, QoS.

I. INTRODUCTION The rapid growth of wireless devices and services exacerbates the problem of spectrum scarcity and poses potential challenges for mobile operators, especially in terms of quality of service provisioning. According to Cisco [1], mobile traffic is expected to grow up to 1000 times by 2020, overwhelming the cellular infrastructure. The ever-increasing mobile applications such as mobile social networks, online gaming and high-definition video streaming may further accelerate this process.
To accommodate the explosive growth in mobile traffic and devices, a new paradigm shift is needed in the design of 5G network architectures, moving from infrastructure-centric networks with exclusive spectrum ownership towards dynamic user-centric approaches incorporating cognitive capabilities. In this context, a new generation of dynamic network architectures (DNAs) is emerging where users share their connections and act as access points for others without additional infrastructure cost. A framework for topology optimization in a DNA is developed in [2] by considering users’ quality of service (QoS) requirements on access point selection. Flexible reconfigurability to adapt to traffic dynamics is achieved by using a genetic algorithm. Perez-Romero et al. [3] propose power-efficient resource allocation schemes for a DNA and show power reductions of approximately 40% with respect to the conventional cellular approach. Moreover, several business models and incentive mechanisms have been proposed and shown the potential of DNA to generate profits for users and operators [4], [5]. Considering that a large portion of spectrum is underutilized temporally and spatially, integrating cognitive radio capabilities into the DNA for spectrum harvesting [6] will facilitate access to additional unused spectrum to meet the growing spectrum demand and mitigate interference among adjacent DNAs. Furthermore, the harvested spectrum may have different propagation or penetration properties, which can be exploited to support diverse applications with various QoS requirements.

Fig. 1. Cognitive network architectures

Novel architectures for cognitive networks based on D2D [7], small cells [8] and multi-hop communications [6], [9] have been proposed to further increase spectrum efficiency. These architectures are illustrated in Fig. 1. D2D spectrum sharing intends to offload traffic from the cellular infrastructure when source and destination are close to each other. Interference minimization is a key issue when multiple D2D pairs share the same resources. In [7] the spectrum sharing problem between a set of D2D pairs and multiple co-located cellular networks is studied. Each D2D link can either access a sub-band occupied by a cellular subscriber or obtain an empty sub-band for its exclusive use. The problem is formulated as a Bayesian non-transferable utility overlapping coalition formation game.
Cognitive capabilities have been utilized to reduce intercell interference among the macrocell and small cells to deploy spectrum efficient cognitive heterogeneous networks [8]. Although the introduction of small access points can solve the capacity demand

This content is AI-processed based on ArXiv data.