Dynamic Traffic Engineering for Cooperative Fog-Cloud Environment: Trade-off Analysis of Cost and Utilization Under Different Load Conditions

Abstract

Fog computing has become an attractive computing method for different IoT (Internet of Things) applications that require low latency and location awareness. It provides low latency by bringing computational power closer to the network edge, working as a complement to cloud computing. Despite its advantages, fog computing faces challenges due to the limited resources (CPU processing capacity, network bandwidth, memory, and power backup) of fog nodes. This work introduces a novel optimization model for a cooperative fog-cloud environment dealing with dynamic traffic. We analyze how different arrival rates impact bandwidth costs, link utilization, and server resource utilization. Our results show that fog resource utilization is greater than cloud resource utilization under varying traffic conditions, with blocking rates remaining within an acceptable range (0-15%). The key contributions include the formulation of an optimization model that optimizes resource allocation, addresses blocking factors in fog networks, and offers valuable insights for managing dynamic traffic in fog computing networks.

Keywords