A Low-Latency ,Energy Efficient Cache-Enabled Cloudlet Selection Framework for Fog Networks

Abstract

Exponential industrial development in all fields will result in an enormous number of smart device deployment.Deployment of internet of thing (IoT) applications is challengingbecauseoftheirstringentultra-reliable,low-latencycommunication(URLLC), and different quality-of-service (QoS) requirements. This scalability and low-latency requirement of IoT nodes results in the inefficiency of cloud computing. An extension of cloud computing named fog computing has emerged to satisfy user’s needs. It will act as a key enabling technology to support the future 5th Generation (5G) IoT applications. Inthisthesis,fognetworksarepresentedasapromisingsolutiontotherequirements ofIoTapplications. However,thepresenceofanenormousnumberofsmartdevicesin fog networks, affect the performance of the network. To maintain the performance of the network, it is a need to find optimal cloudlet selection strategies. A mixed-integer non-linear programming (MINLP) problem that has an objective of latency minimization under cache size, association, and optimal power allocation constraints is formulated. An exhaustive search on proposed cache-enabled (CE) fog architecture cannot be applied because of the problem’s combinatorial and NP-hard nature. The increase in the number of IoT and fog nodes increases the search space for finding the solution, hence an Outer Approximation Algorithm (OAA) is proposed. Though cloudlet nodes in fog networks have relatively lower power consumption, dense IoT application deployments may result in a high aggregate of energy consumption. This brings the motivation to modify the problem formulation for energy efficiency. A joint node association and energy efficiency (JNAEE) maximization problem is formulated. The proposed problem is a non-linear concave programming problem. A mesh adaptive direct search algorithm (MADS) compared with a more complex OAA algorithm is used to study the network’s performance. The simulations are done using the OPTI toolbox in MATLAB.