Scientific Workflows in IoT Environments: A Data Placement Strategy Based on Heterogeneous Edge-Cloud Computing

Abstract

In Industry 4.0 and IoT environments, the heterogeneous edge-cloud computing paradigm can provide a more proper solution to deploy scientific workflows compared to cloud computing or other traditional distributed computing. Owing to the different sizes of scientific datasets and the privacy issue concerning some of these datasets, it is essential to find a data placement strategy that can minimize data transmission time. Some state-of-the-art data placement strategies combine edge computing and cloud computing to distribute scientific datasets. However, the dynamic distribution of newly generated datasets to appropriate datacenters and exiting the spent datasets are still a challenge during workflows execution. To address this challenge, this study not only constructs a data placement model that includes shared datasets within the individual and among multiple workflows across various geographical regions, but also proposes a data placement strategy (DYM-RL-DPS) based on algorithms of two stages. First, during the build-time stage of workflows, we use the discrete particle swarm optimization algorithm with differential evolution to pre-allocate initial datasets to proper datacenters. Then, we reformulate the dynamic datasets distribution problem as a Markov decision process and provide a reinforcement learning-based approach to learn the data placement strategy in the runtime stage of scientific workflows. Through using the heterogeneous edge-cloud computing architecture to simulate IoT environments, we designed comprehensive experiments to demonstrate the superiority of DYM-RL-DPS. The results of our strategy can effectively reduce the data transmission time as compared to other strategies.

Task Scheduling for Federated Learning in Edge Cloud Computing Environments by Using Adaptive-Greedy Dingo Optimization Algorithm and Binary Salp Swarm Algorithm

“…However, making the entire system bandwidth available to a mobile device to transmit data may lead to network congestion and increase the energy consumption of the mobile device. Du and Tang [17] constructed a data placement model that dynamically allocates newly generated datasets to appropriate data centers and removed exhausted datasets during workflow execution. Ra [18] proposed a greedy staged offloading algorithm to solve the problem of task offloading.…”

A Deep Learning Algorithm for Special Action Recognition of Football

“…Multistage, two-stage, and one-stage object identification techniques are available. Early examples of cross-strategic networks are R-CNN and SPPNet [31][32][33]. Each of the stages of a search may be taught on its own or in conjunction with the others.…”

Lidom: A Disease Risk Prediction Model Based on LightGBM Applied to Nursing Homes

“…For example, a health risk warning will be given when the detected diastolic blood pressure is more significant than 90 mmHg or the systolic blood pressure is greater than 140 mmHg, when the detected total cholesterol is greater than 6.2 mmol/L or triglyceride is greater than 2.3 mmol/L, when the detected fasting blood sugar is higher than 7.0 mmol/L, when the detected low-density lipoprotein is greater than 3.12 mmol/L, etc. However, these early warnings are single decision-making outputs based on a single indicator, not health risk predictions, after a comprehensive analysis of the obtained physical symptoms and physiological values [3]. Therefore, this single decision making lacks comprehensiveness in the early warning of health risks.…”