8 July 2020 in IEEE Transactions on Network and Service Management
With network slicing, physical networks are partitioned into multiple virtual networks tailored to serve different types of service with their specific requirements. In order to optimize the utilization of network resources for delay-critical applications, we propose a new multi-domain network virtualization framework based on a novel multipath multihop delay model. This framework encompasses a novel hierarchical orchestration mechanism for mapping network slices onto physical resources and a mechanism for dynamic slice resizing. The main idea is to locally redefine the delay requirements on each network domain depending on the conditions in the rest of the network. Delays larger than threshold (debt) are allowed in certain domains if there is a possibility to compensate such excessive delays in other segments of the network that can transmit the messages with less latency (credit). This tradeoff or delay threshold redefinition on different segments of the route is referred to as network latency equalization. For performance comparison, minimum cost routing with latency constraints is used as a baseline. We show that our approach enables significantly better utilization of the network resources measured in the number of slices with the same latency requirements that can be accommodated in the network.
14 May 2024 in Intelligent Systems with Applications
4 October 2022 in IEEE Transactions on Vehicular Technology
Vehicle-centric edge computing provides a promising approach to reducing the network operational costs of the future latency-limited applications (e.g. connected-autonomous vehicles, virtual reality) by migrating the required services toward the vehicle-provided resources. However, uncertainties from unknown traffic and availability of these resources as well as diversity of vehicle platforms can result in significant performance degradation of such service migration mechanisms. In this paper, we propose vehicle-centric networking concept to provide low latency services by designing time compression mechanism for computing programs so that the program execution latency can be significantly reduced. In addition, by jointly tailoring data and control planes we implement a highly responsive and cost-efficient edge service. By leveraging network functions virtualization (NFV) and software defined networking (SDN) technologies we provide a suitable environment for at-edge programmability to customize network functionalities for low latency services in multi-platooning systems for vehicular networks. We develop a responsive edge system by implementing control plane over pre-established access points (e.g., road-side units), allocating redundant resources, and utilizing an adaptive service migration mechanism in order to avoid any degradation of the quality of edge services. We design a robust double Knapsack (RDK) framework, where a spare knapsack of redundant resources is allocated to each plane to cope with the uncertainties of vehicle-provided resources. By combining with VNFs, SDN technology has been used to design a centralized management framework to control and monitor the edge computing in multi-platooning systems.
4 October 2022 in IEEE Transactions on Vehicular Technology
In this paper, we propose an application customized network slicing mechanism to cope with dynamic traffic in future networks and various service requirements for network users, namely, context-aware network slicing (CANS). The resource allocation is carried out in frequency (f), space (s) and time (t) domains (referred to as three domains (3D) resource allocation) while the customization targets at low latency and high reliability, characteristic for IoT applications. We first formulate CANS as a two-step matching game with unknown preferences for access and back-haul networks. To dynamically resize mobile users' network slices (NSs), we propose application customized CANS++ method based on a coalition game to minimize the hand-off risk in the network. We then develop two-step stable matching (TSSM) and $\epsilon$ -Pareto stability techniques to find the solutions to CANS and CANS++, respectively. To improve energy and cost efficiency, an artificial intelligence-based Dynamic Network Architecture (DNA) with ON/OFF scheduling (AI-DOS) is proposed to dynamically maximize NS utilization. Finally, we carry out extensive numerical analysis and demonstrate that context-aware 3D network slicing and resizing schemes can improve at least three times the users' utilities. These utilities are enhanced up to $180\%$ by minimizing their hand-off risk. Furthermore, NS shrinking technique can reduce the number of active nodes (controlled in the s-domain) by factor five while the average utility of macro base stations, controlled in the t-domain, is improved between $70\%$ and $500\%$ .