In this paper, we analyze the performance of a wireless caching system with heterogeneous traffic and relaying capabilities satisfying secrecy constraints for one of two receiving users. In this setup, the second user has no secrecy requirements and receives cacheable content either from the relay helper or the core network through a wireless base station. The wireless relay helper can assist both users since it is equipped with finite storage that is split into cacheable and non-cacheable storage. Concurrently, a passive eavesdropper tries to overhear transmissions to the user with secrecy requirements. Consequently, we examine how this relay’s storage split and the eavesdropper affect the performance of the average throughput and delay of the system as the transmission powers, the relay’s transmission probability, and the relay’s cache size vary. © 2023 IEEE.

Virtual Private LAN Service (VPLS) is a VPN technology that connects remote client sites with provider networks in a transparent manner. Session key-based HIPLS (S-HIPLS) is a VPLS architecture based on the Host Identity Protocol (HIP) that provides a secure VPLS architecture using a Key Distribution Center (KDC) to implement security mechanisms such as authentication, encryption etc. It exhibits limited scalability though. Using multiple distributed KDCs would offer numerous advantages including reduced workload per KDC, distributed key storage, and improved scalability, while simultaneously eliminating the single point of failure of S-HIPLS. It would also come with the need for optimally placing KDCs in the provider network. In this work, we formulate the KDC placement (KDCP) problem for a secure VPLS network as an Integer Linear Programming (ILP) problem. The latter is NP-hard, thereby suggesting a high computational cost for obtaining exact solutions especially for large deployments. Therefore, we motivate the use of a primal-dual algorithm to efficiently produce near-optimal solutions. Extensive evaluations on large-scale network topologies, such as the random Internet graph, demonstrate our method’s time-efficiency as well as its improved scalability and usefulness compared to both HIPLS and S-HIPLS.

We study a discrete-time wireless network that serves both cacheable and non-cacheable traffic with assistance of a relay node with storage capabilities for both types of traffic. We investigate how allocating the storage capacity to cacheable and non-cacheable traffic affects the network throughput. Our numerical results provide useful insights by varying not only the allocation of cacheable to non-cacheable storage but also the rate by which non-cacheable content is transmitted, the rate by which cacheable content is requested, as well as different popularity distributions of the cached files.

The Internet of Things (IoT) should be able to react with minimal human intervention and contribute to the Artificial Intelligence (AI) era requiring real-time and scalable operation under heterogeneous network infrastructures. This thesis investigates how cooperation and allocation of resources can contribute to the evolution of future wireless networks supporting the IoT.

First, we examine how to allocate resources to IoT services which run on devices equipped with multiple network interfaces. The resources are heterogeneous and not interchangeable, and their allocation to a service can be split among different interfaces. We formulate an optimization model for this allocation problem, prove its complexity, and derive two heuristic algorithms to approximate the solution in large instances of the problem.

The concept of virtualization is promising towards addressing the heterogeneity of IoT resources by providing an abstraction layer between software and hardware. Network function virtualization (NFV) decouples traditional network operations such a routing from proprietary hardware platforms and implements them as software entities known as virtualized network functions (VNFs). In the second paper, we study how VNF demands can be allocated to Virtual Machines (VMs) by considering the completion-time tolerance of the VNFs. We prove that the problem is NP-complete and devise a subgradient optimization algorithm to provide near-optimal solutions. Our numerical results demonstrate the effectiveness of our algorithm compared to two benchmark algorithms.

Furthermore, we explore the potential of using intermediate nodes, the so-called relays, in IoT networks. In the third paper, we study a multi-user random-access network with a relay node assisting users in transmitting their packets to a destination node. We provide analytical expressions for the performance of the relay's queue and the system throughput. We optimize the relay’s operation parameters to maximize the network-wide throughput while maintaining the relay's queue stability. A stable queue at relay guarantees finite delay for the packets. Furthermore, we study the effect of the wireless links' signal-to-interference-plusnoise ratio (SINR) threshold and the self-interference (SI) cancellation on the per-user and network-wide throughput.

Additionally, caching at the network edge has recently emerged as an encouraging solution to offload cellular traffic and improve several performance metrics of the network such as throughput, delay and energy efficiency. In the fourth paper, we study a wireless network that serves two types of traffic: cacheable and non-cacheable traffic. In the considered system, a wireless user with cache storage requests cacheable content from a data center connected with a wireless base station. The user can be assisted by a pair of wireless helpers that exchange non-cacheable content as well. We devise the system throughput and the delay experienced by the user and provide numerical results that demonstrate how they are affected by the non-cacheable packet arrivals, the availability of caching helpers, the parameters of the caches, and the request rate of the user.

Finally, in the last paper, we consider a time-slotted wireless system that serves both cacheable and non-cacheable traffic with the assistance of a relay node. The latter has storage capabilities to serve both types of traffic. We investigate how allocating the storage capacity to cacheable and non-cacheable traffic affects the system throughput. Our numerical results provide useful insights into the system throughput e.g., that it is not necessarily beneficial to increase the storage capacity for the non-cacheable traffic to realize better throughput at the non-cacheable destination node.

We consider a cost minimization problem for High Volume Servers (HVS) equipped with Virtual Machines (VMs) to serve Virtual Network Functions (VNF) demands for resources. Given a scheduling period, a central scheduler decides which VM to use for each VNF demand. Each VM can be activated or disabled with different costs. Each VNF has a delay-weighted pricing function to indicate its completion time tolerance. We prove the NP-completeness of the problem and develop an algorithm based on Lagrangian relaxation and subgradient optimization to deal with this computational complexity. Finally, our numerical results demonstrate our algorithm’s effectiveness compared to two benchmarks.

The Internet of Things (IoT) promises that "anything that can be connected, will be connected". It comprises of Information and Communication Technologies that interconnect billions of physical and visual things with some "basic" intelligence. The emerging IoT services will be able to react with minimal human intervention and further contribute to the big data era that requires real-time, ultrareliable, ubiquitous, scalable, and heterogeneous operation.

This thesis is the result of our investigations on problems dealing with the evolution of such technologies. First, we explore the potential of using relay i.e., intermediate, nodes that assist users to transmit their packets in a a cellular network. Paper I provides insights into how adapting the cooperation of the relay's receiver and transmitter optimizes the network-wide throughput while the relay's queue stability is guaranteed.

The next part of the thesis copes with the resource allocation of services on IoT devices equipped with multiple network interfaces. The resources are heterogeneous and can be split among dierent interfaces. Additionally, they are not interchangeable. In paper II, we develop optimization models for this resource allocation problem, prove the complexity of the models, and derive results that give intuition into the problems. Moreover, we propose algorithms that approximate the optimal solution and show under which circumstances this is possible.

Finally, in paper III, we present a resource allocation problem specically for smart cities services. In comparison to the previous problem denition, resources are of one type but the IoT network device can oer capacities that vary over time. Furthermore, services have a tolerance regarding their preferred scheduling, namely, their allocation over time. We parametrize each service with a pricing function to indicate its tolerance to be served at the beginning of the scheduling window. We prove that the problem is computationally hard and provide numerical results to gain insight into how different pricing weight functions impact the allocations' distribution within the scheduling window.

The emerging vision of smart cities necessitates the use of Internet of Things (IoT) network devices to implement sustainable solutions that will improve the operations of urban areas. A massive amount of smart cities services may demand allocation of computational resources, such as processing power or storage, that IoT devices offer. Within this context, we present an IoT network device comprising interfaces with one specific computational resource available. The efficient utilization of available IoT resources would improve the Quality of Service (QoS) of the IoT network that serves the smart city. All resource allocations must be completed within a given scheduling window and every service is parametrized by a pricing weight function to indicate its tolerance to be served at the beginning of the scheduling window. We propose a mathematical optimization formulation to minimize the total cost of allocating all demands within the scheduling window considering the tolerance level of each service at the same time. Moreover, we prove that the problem is computationally hard and we provide numerical results to gain insight into the impact of different pricing weight functions on the allocations’ distribution within the scheduling window.

In this work, we investigate the operation of energy efficient relay nodes assisting the transmission of packets from a number of users to a destination node. We study the impact of switching randomly a fraction of the relays on and off to the aggregate throughput, the average queue size and delay per packet of systems with relays transmitting in either Full-or Half-Duplex mode and under different channel transmissions characteristics. Furthermore, we prove analytically and illustrate via simulation means how these network metrics are affected.

In the Internet of Things (IoT), devices and gateways may be equipped with multiple, heterogeneous network interfaces which should be utilized by a large number of services. In this work, we model the problem of assigning services' resource demands to a device's heterogeneous interfaces and give a Mixed Integer Linear Program (MILP) formulation for it. For meaningful instance sizes the MILP model gives optimal solutions to the presented computationally-hard problem. We provide insightful results discussing the properties of the results on the properties of the derived solutions with respect to the splitting of services to different interfaces.