Mobility-Aware IoT Application Placement in the Cloud – Edge Continuum

Abstract

The Edge computing extension of the Cloud services towards the network boundaries raises important placement challenges for IoT applications running in a heterogeneous environment with limited computing capacities. Unfortunately, existing works only partially address this challenge by optimizing a single or aggregate objective (e.g., response time), and not considering the edge devices' mobility and resource constraints. To address this gap, we propose a novel mobility-aware multi-objective IoT application placement (mMAPO) method in the Cloud -Edge Continuum that optimizes completion time, energy consumption, and economic cost as conflicting objectives. mMAPO utilizes a Markov model for predictive analysis of the Edge device mobility and constrains the optimization to devices that do not frequently move through the network. We evaluate the quality of the mMAPO placements using simulation and real-world experimentation on two IoT applications. Compared to related work, mMAPO reduces the economic cost by 28% and decreases the completion time by 80% while maintaining a stable energy consumption.

Scheduling IoT Applications in Edge and Fog Computing Environments: A Taxonomy and Future Directions

“…They provide low-latency and high-bandwidth access to decision engines for IoT devices. Majority of works such as [36,60,81,91] deployed the decision engines in Edge/Fog Layer. available or when IoT applications are insensitive to higher startup time.…”

“…Simulators keep the advantages of analytical tools while improving the credibility of evaluations by simulating the dynamics of resources, applications, and environments. The iFogSim [40,71] is among the most popular simulators for Fog computing [37,60,68,117]. Besides, several researchers have used Cloudsim [15] such as [23,133] or SimPy such as [28,84] for the simulation.…”

“…Based on the nature and popularity of these metrics in the literature, we categorize them into 1) time (e.g., deadline, response time, execution time, makespan) [22,23,66,81], 2) energy (e.g., battery percentage, saved energy) [12,18,51,51], 3) monetary cost (e.g., service cost, switching cost) [8,82,84,89], and 4) other metrics (e.g., number of interrupted tasks, resource utilization, throughput, deadline miss ratio) [13,14,35,46]. Also, we consider 5) decision overhead as an important evaluation metric to study the overhead of proposals (often in terms of time and energy), used in some works such as [36,60,91,130].…”

Scheduling IoT Applications in Edge and Fog Computing Environments: A Taxonomy and Future Directions

“…3) monetary cost (e.g., service cost, switching cost) [9,90,94,100], and 4) other metrics (e.g., number of interrupted tasks, resource utilization, throughput, deadline miss ratio) [14,16,38,49]. Also, we consider 5) decision overhead as an important evaluation metric to study the overhead of proposals (often in terms of time and energy), used in some works such as [39,64,102,142].…”

“…They provide low-latency and high-bandwidth access to decision engines for IoT devices. Majority of works such as[39,64,89,102] deployed the decision engines in Edge/Fog Layer.…”