Software-oriented architecture (SOA) is a promising paradigm for efficiently leveraging the functionality of individual IoT devices to build IoT applications. However, deploying SOA for IoT data-gathering applications requires spatial context-awareness and the ability to aggregate similar available services, which presents a challenge. To address this challenge, this paper proposes a formulation for spatial context-aware service composition with a novel quantitative model for spatial context. We demonstrate that incorporating spatial context into service composition is an NP-Hard problem and model it as an integer linear program. We propose two heuristic approaches capable of producing near-optimal solutions in real-time. We implement a simulation of the composition problem to study the performance of our approaches. Our experimental results show that the proposed methods are scalable compared to the branch-and-cut algorithm. These results set a precedent against which future work on solutions to this problem can be compared.

Cooperative-Autonomous-Driving (CAD) systems stringently require that vehicle status information (e.g, speed, position, etc) be timely disseminated for safety reasons. CAD systems rely on real-time information to make critical decisions; hence, the paramount criticality of temporally valid information generation and dissemination. However, the timely information update messages’ delivery faces numerous challenges due to the highly alternating wireless signal propagation in vehicular environments as a result of, for instance, shadowing and blockage, which lead to the unavailability of reliable communication links between cooperating vehicles. Under such harsh conditions, Reconfigurable-Intelligent-Surfaces (RISs) have been proven to highly contribute in mitigating the propagation-induced impairments of the wireless environments and, hence, promoting more robust communication links, which, in turn, allow for maintaining the required freshness of information. In the above-context, this paper revolves around the minimization of the Age-of-Information (AoI) perceived by each of a CAD system’s destination node. The problem is formulated as an Integer-Linear-Program (ILP), which turns out to be quite complex. To work around this complexity, it is proposed herein to use decomposition based on the Lagrangian relaxation method, which largely facilitates the problem’s resolution following typical dynamic programming methodologies. Consequently a feasible solution is extracted using a relatively simple heuristic. An analytical framework is established to reveal insights into the proposed solution and gauge its merits through the establishment of a thorough simulation framework involving various scenarios aiming at verifying its correctness, validity and superiority as compared to other solutions derived using the state-of-the-art branch-and-cut method implemented by CPLEX.

Internet of Things (IoT) applications, such as augmented/virtual reality, tactile Internet, immersive gaming, etc., are currently experiencing an unprecedented growth in their demand. IoT devices are constrained by limited computation and power features and might experience excessive computational latency to support resource-intensive tasks. Multiaccess edge computing (MEC) appears to be a promising solution in this regard to expedite the computations of resource-intensive tasks by offloading them to the edge of the network. This article considers a scenario where a base station (BS) serves traffic streams from multiple IoT devices. The packets from each stream arrive at the BS (following a stochastic process) and then forwarded to their respective destinations after they are processed by the MEC node. The scheduling decisions are aimed to keep the information fresh at the destination. The information freshness is captured by Age of Information (AoI) metric. We aim to minimize the expected sum AoI for the MEC-assisted IoT network and provide mathematically traceable expressions for the AoI. First, an optimization problem is formulated to find the optimal scheduling policy in order to minimize the expected sum AoI. The optimization problem is an integer linear programming (LP) problem, which is generally difficult to solve. Hence, we provide a simpler formulation of the problem and derive a more traceable expression for the expected sum AoI. With this approach, the joint impact of stochastic arrivals, scheduling policy, and unreliable channel conditions on the AoI is assessed. We also propose low-complexity algorithms to obtain results for larger networks. Finally, through extensive simulations, we demonstrate the effectiveness of our proposed methods as compared to other existing strategies in terms of achievable AoI.

Fully automated vehicles deployed with high computational/perceptive capabilities will soon become a reality. Such capabilities enable the cooperation among vehicles and the realization of interacting autonomous driving systems. Edge computing has emerged to provide a plethora of computational services to reduce network latency. Applications at the edge that apply analytics on the sensory data are therefore indispensable for self-driving vehicles. We consider in this paper a network that interconnects vehicles to an edge server at a roadside unit. Each vehicle extracts multiple information by sampling multiple processes and sends them to the corresponding edge application. To make timely decisions, “fresh” information needs to be offloaded, processed, and delivered back to vehicles; in this context, we adopt a new metric called Age of Information (AoI) that has been lately used to measure the freshness of information. We seek to jointly schedule vehicles’ transmission of information and schedule information processing at the edge to minimize the AoI of all processes. We mathematically formulate the problem and prove its NP-Hardness. To overcome this hardness, we propose a logic-based Benders decomposition to divide the problem into a master and several subproblems. Then, we present an exact polynomial-time solution for the subproblems, a scalable heuristic for the master, and devise a valid yet efficient Benders cut. We implement the system simulation on the well-known traffic simulator SUMO and compare the decomposition with CPLEX branch-and-cut; Although the problem is highly intricate, our method finds a near-optimal solution (maximum deviation is 7% from optimal solution) with a speedup that reaches 95%. We study the system performance by varying different system parameters.

Telecommunication networks are currently realizing more-huge-than-ever data demands from subscribers all over the world. Due to the ongoing pandemic, nearly all businesses have adapted working models with remote operations. People engaged with major industries, e.g., academia, health and municipalities are utilizing online platforms to carryout their routine tasks. This indeed shifts the attention from one-to-one (unicast) communication to one-to-many (multicast) and many-to-many (multi-source multi-destination) communications. Network operators are facing increased pressure to provide quick responses in order to satisfy the bandwidth hungry and time sensitive user demands. This can only be done by enhancing deployability as well as manageability of the services. Network Function Virtualization (NFV) provides a transformation of traditional proprietary network designs to a more agile and software based environment in order to achieve flexible deployments, reduced setup costs and less-time-to-market for the new services which is very much needed in the current scenarios. Previous studies on NFV-enabled multicast problem either proposed Integer Linear Program (ILP) models, that are pretty unscalable, or heuristic-based techniques that do not guarantee good quality of the solutions obtained. In this article, we propose an NFV multicast resource optimization model exploiting the use of multiple sources and considering the end-to-end delay and bandwidth requirements. Herein, we propose a novel Dantzig-Wolfe (DW) decomposition model that tackles the complexity of the problem by breaking it down into a master problem and several pricing problems. We compare the DW approach with the ILP and heuristic methods and demonstrate that our approach achieves near to optimal solution (in comparison to heuristic based methods) much faster than ILP. We also study the dynamic admission of NFV-enabled multicast requests by solving the problem in an online manner using the batch processing of requests. We then evaluate the performance of the proposed algorithms through extensive simulations and demonstrate that proposed algorithms are promising and outperform existing solutions.

The Vehicle-as-a-Resource is an emerging concept that allows the exploitation of the vehicles' computational resources for the purpose of executing tasks offloaded by passengers, vehicles, or even an Internet-of-Things devices. This article revolves around a scenario where a roadside unit located at the edge of a hierarchical multitier edge computing subnetwork resorts to the utilization of idle vehicles computational resources through a fog-enabled substructure yielding a cost-effective computational task offloading solution. In this context, scheduling the offload of these tasks to the appropriate vehicles is a challenging problem that is subject to the interaction of major role-playing parameters. Among these parameters are the variability of vehicles availability and their computational power, the individual tasks' weighted priorities and their deadlines, the tasks required computational power as well as the required data to upload/download. This article proposes an infrastructure-assisted task scheduling scheme where the roadside unit receives computational tasks from different sources and schedules these tasks over a computationally capable vehicle residing within the roadside unit's range. The aim is to maximize the weighted number of admitted tasks while considering the constraints mentioned above. Compared to other works, this article broaches a more realistic scenario by considering a more accurate computational task and system model. Our system considers both the latency and throughput of task accomplishments by maximizing the weighted number of admitted tasks while at the same time respecting the tasks accompanied deadlines. Both radio and computational resources are part of the optimization problem. After proving the NP-hardness of the scheduling problem, we formulated the problem as a mixed-integer linear program. A Dantzig-Wolfe decomposition algorithm is proposed which yields to a master program solvable by the Barrier algorithm and subproblems solved optimally with a polynomial-time dynamic programming approach. Thorough numerical analysis and simulations are conducted in order to verify and assert the validity, correctness, and effectiveness of our approach compared to branch and bound and greedy algorithms.

We study the computation capability of the BP cycles counting algorithm in random networks by estimating the required convergence error to find the maximum number of points in the cycles distribution which helps to interpolate accurately the cycles counts for different sizes. The algorithm shows that the convergence error required to get all possible points is almost constant through all possible connection probability values. The number of points possible can not reach the maximum even with high convergence error values.

We calculate the cyclic entropy of a real virtual friendship network to have an insight on the degree of its robustness. Upon counting the number of cycles of different sizes in the network, a probability distribution function is resulted. An actual friendship network is found to have cyclic entropy bounded between random and small-world networks models. It has dual properties. Small world networks indicate the existence of critical network sizes: 150 and 700 at which the cyclic entropy is minimum. Scale-free networks have the highest cyclic entropy among all other complex network models regardless of the size of the network.

We modify the statistical mechanical based Belief Propagation (BP) algorithm to compute cycles in random networks using a phenomenological Gaussian distribution of cycles. The modified BP algorithm tested over any random network improves cycles computational time. CPU time is reduced up to 60% compared to the original BP algorithm.