The number of devices connected to the Internet of Things (IoT) is growing at an enormous rate globally. In the next generation networks, distributed fog computing deployments at the network edge can provide computing resources to the users, especially for latency-sensitive applications. Further, the heterogeneous needs of the fifth generation (5G) networks demand the virtualization of network functions, termed as network function virtualization (NFV). Therefore, an integrated NFV and fog computing resource allocation framework for IoT is of prime importance. Accordingly, in this paper, we model the interactions between the data service operators (DSOs) and the authorized data service subscribers (ADSSs) as an equilibrium problem with equilibrium constraints (EPEC), and utilize the alternating direction method of multipliers (ADMM) as a large-scale optimization tool to obtain solutions. This results in the optimization of resource pricing for the DSOs and the amount of resources to be purchased by the ADSSs. Moreover, we propose a many-to-many matching based model to allocate the fog node (FN) resources according to the VNF resource requirements of the ADSSs. Simulation results show the effectiveness of our proposed approach in achieving efficient resource allocation in NFV enabled IoT fog computing.

Wireless network virtualization is basically the abstraction, isolation, and sharing of wireless resources among different entities. Consequently, virtualization provides great flexibility and higher network efficiency, and enables easier migration to new technologies in wireless networks. Traditionally, a wireless network virtualization controller manages the virtual resources (including radio resources and infrastructure resources) known as slices which are available to the Service Providers (SPs). The SPs then allocate their purchased resources to serve their subscribed mobile users. Such a centralized allocation decouples the Quality-of-Service (QoS) management by the SPs from the virtual resource management by the controller. In this paper, we propose a matching based wireless network virtualization resource allocation mechanism: a distributed three-sided (3D) matching between radio resources, physical infrastructure and mobile users. The Restricted Three-sided Matching with Size and Cyclic preference model (R-TMSC) is implemented to obtain a stable solution. Simulation results show that our proposed spectrum-oriented and user-oriented algorithms outperform the traditional resource allocation schemes. The spectrum-oriented algorithm enhances the user throughput and the system performance, within a lesser run time. Furthermore, for an increasing number of users, the proposed algorithms serve more users than traditional methods.

Network Function Virtualization (NFV) is one of the key drivers of 5G systems, which involves the virtualization of the Evolved Packet Core (EPC) and the 5G Core. This entails deploying virtual instances of core network functions in Cloud Networks (CNs), resulting in a virtual EPC (vEPC)/5G Core network. NFV resource allocation is a popular research topic in 5G systems, and a distributed solution based on the interdependencies between all the important entities is very crucial. Accordingly, in this paper, we propose a three-sided matching based framework for virtual resource allocation in next-generation networks. We utilize the Restricted Three-sided Matching with Size and Cyclic preference (R-TMSC) problem to model the relationships between Tracking Areas (TAs) (Base Stations (BSs) organized together in groups), CNs and Virtual Network Function (VNF) instances in a vEPC/5G Core network. The simulation results clearly demonstrate the superior performance of the proposed framework in terms of the data rates provided by the CNs and user satisfaction, compared to a centralized random allocation approach.

The radio frequency spectrum crunch has triggered the harnessing of other sources of bandwidth, for which visible light is a promising candidate. Even though visible light communication (VLC) ensures high capacity, coverage is limited. This necessitates the integration of VLC and device-to-device (D2D) technologies into heterogeneous networks. In particular, mobile users which are accessible by the VLC transmitters can relay data to mobile users which are not, by means of D2D communication. However, due to the distributed behaviors of mobile users, determining optimal data transmission routes from VLC transmitters to end mobile devices is a major challenge. In this paper, we propose a reinforcement learning (RL)-based approach to determine multi-hop data transmission routes in an indoor VLC-D2D heterogeneous network. We obtain the rewards for the RL-based method dynamically, by formulating the interactions between the mobile users relaying the data as an equilibrium problem with equilibrium constraints and using alternating direction method of multipliers to solve it. The proposed technique can achieve optimal data transmission routes in a distributed manner. The simulation results demonstrate the effectiveness of the proposed approach, showing that transmission routes with low delays and high capacities can be achieved through the learning algorithm.

Fog Computing potentially plays a pivotal role in delivering real-time data services to users, where large number of Fog Nodes (FNs) are deployed by various Data Service Operators (DSOs) to provide efficient services to the Authorized Data Service Subscribers (ADSSs). As a result, there exist tradings between DSOs and ADSSs in providing and purchasing these services, respectively. There also exists competition among all DSOs for providing these services at the prices that can maximize their profits. Moreover, competition exists among all ADSSs for purchasing the required amount of resources at the lowest available prices, and thus minimizing their costs. In this paper, we model the aforementioned competitions in fog computing as an Equilibrium Problem with Equilibrium Constraints (EPEC). In the EPEC, the DSOs provide incentives to the ADSSs and balance the utilities between the DSOs and the ADSSs. At the same time, the ADSSs leverage the incentives provided by the DSOs to their advantage. As the size of a typical fog computing network is large, the Alternating Direction Method of Multipliers (ADMM) algorithm, that has been recognized as a key method in large scale optimization can be employed. Utilizing the fast convergence and decomposition properties of ADMM, we achieve optimum results. Simulation results show that with the proposed framework, optimization of the utility functions of DSOs and ADSSs can be achieved in real-time. It is also shown that compared to the profit in traditional cloud computing and data center services, the total maximum profit of the ADSSs is improved to a great extent in fog computing.