Abstract:Tactile Internet is an emerging concept that focuses on supporting high-fidelity, ultra-responsive, and widely available human-to-machine interactions. To reduce the transmission latency and alleviate Internet congestion, fog computing has been advocated as an important component of the Tactile Internet. In this paper, we focus on an energy-efficient design of fog computing networks that support low-latency Tactile Internet applications. We investigate two performance metrics: Service response time of end-users and power usage efficiency of fog nodes. We quantify the fundamental tradeoff between these two metrics and then extend our analysis to fog computing networks involving cooperation between fog nodes. We introduce a novel cooperative fog computing concept, referred to as offload forwarding, in which a set of fog nodes with different computing and energy resources can cooperate with each other. The objective of this cooperation is to balance the workload processed by different fog nodes, further reduce the service response time, and improve the efficiency of power usage. We develop a distributed optimization framework based on dual decomposition to achieve the optimal tradeoff. Our framework does not require fog nodes to disclose their private information nor conduct back-and-forth negotiations with each other. Two distributed optimization algorithms are proposed. One is based on the subgradient method with dual decomposition and the other is based on distributed alternating direction method of multipliers via variable splitting. We prove that both algorithms can achieve the optimal workload allocation that minimizes the response time under the given power efficiency constraints of fog nodes. Finally, to evaluate the performance of our proposed concept, we simulate a possible implementation of a city-wide self-driving bus system supported by fog computing in the city of Dublin. The fog computing network topology is set based on a real cellular network infrastructure involving 200 base stations deployed by a major cellular operator in Ireland. Numerical results show that our proposed framework can balance the power usage efficiency among fog nodes and reduce the service latency for users by around 50% in urban scenarios.

Heterogeneous Online Learning for “Thing-Adaptive” Fog Computing in IoT

Distributed Online Learning of Fog Computing under Nonuniform Device Cardinality

An Online Fog Computing Task Offloading Algorithm Based on Robust Evolutionary Optimization

Learning and Management for Internet of Things: Accounting for Adaptivity and Scalability

Optimizing Resources Allocation for Fog Computing-Based Internet of Things Networks.

Green Offloading in Fog-Assisted IoT Systems: An Online Perspective Integrating Learning and Control

Distributed Online Optimization of Fog Computing for Internet of Things under Finite Device Buffers.

Exploring Renewable-Adaptive Computation Offloading for Hierarchical QoS Optimization in Fog Computing

AdaptiveFog: A Modelling and Optimization Framework for Fog Computing in Intelligent Transportation Systems

Task Scheduling in Fog Enabled Internet of Things for Smart Cities

Cognitive Balance for Fog Computing Resource in Internet of Things: An Edge Learning Approach

An Online Learning Approach to Computation Offloading in Dynamic Fog Networks

Adaptive Fog Configuration for the Industrial Internet of Things

Reliable Fog-Based Crowdsourcing: A Temporal–Spatial Task Allocation Approach

Adaptive Offloading for Time-Critical Tasks in Heterogeneous Internet of Vehicles

Green Fog Planning for Optimal Internet-of-Thing Task Scheduling

Optimizing Internet of Things Fog Computing: Through Lyapunov-Based Long Short-Term Memory Particle Swarm Optimization Algorithm for Energy Consumption Optimization

Computing resource allocation in three-tier IoT fog networks: A joint optimization approach combining Stackelberg game and matching

SIoTFog: Byzantine-resilient IoT Fog Networking.

Distributed Optimization for Energy-Efficient Fog Computing in the Tactile Internet

Online Learning Offloading Framework for Heterogeneous Mobile Edge Computing System