Microservices-based IoT Application Placement within Heterogeneous and Resource Constrained Fog Computing Environments

Abstract

Fog computing paradigm has created innovation opportunities within Internet of Things (IoT) domain by extending cloud services to the edge of the network. Due to the distributed, heterogeneous and resource constrained nature of the Fog computing nodes, Fog applications need to be developed as a collection of interdependent, lightweight modules. Since this concept aligns with the goals of microservices architecture, efficient placement of microservices-based IoT applications within Fog environments has the potential to fully leverage capabilities of Fog devices. In this paper, we propose a decentralized microservices-based IoT application placement policy for heterogeneous and resource constrained Fog environments. The proposed policy utilizes the independently deployable and scalable nature of microservices to place them as close as possible to the data source to minimize latency and network usage. Moreover, it aims to handle service discovery and load balancing related challenges of the microservices architecture. We implement and evaluate our policy using iFogSim simulated Fog environment. Results of the simulations show around 85% improvement in latency and network usage for the proposed microservice placement policy when compared with Cloud-only placement approach and around 40% improvement over an alternative Fog application placement method known as Edge-ward placement policy. Moreover, the decentralized placement approach proposed in this paper demonstrates significant reduction in microservice placement delay over centralized placement.