Edge vs. Cloud: Empirical Insights into Data-Driven Condition Monitoring

Big Data and Cognitive Computing, Journal Year: 2025, Volume and Issue: 9(5), P. 121 - 121

Published: May 8, 2025

This study evaluates edge and cloud computing paradigms in the context of data-driven condition monitoring rotating electrical machines. Two well-known platforms, Raspberry Pi Amazon Web Services Elastic Compute Cloud, are used to compare contrast these two terms different metrics associated with their application suitability. The tested induction machine fault diagnosis models developed using popular algorithms, namely support vector machines, k-nearest neighbours, decision trees. findings reveal that while platform offers superior computational memory resources, making it more suitable for complex learning tasks, also incurs higher costs latency. On other hand, excels real-time processing reduces network data burden, but its resources found be a limitation certain tasks. provides both quantitative qualitative insights into trade-offs involved selecting most approach applications. Although scope empirical is primarily limited factors such as efficiency, scalability, resource utilisation, particularly specific models, this paper broader discussion future research directions key issues, including latency, variability, energy consumption.

Language: Английский

---

Towards Improving YARN performance for Frugal Heterogeneous SBC-based Edge Clusters

Published: May 3, 2024

Efficient resource allocation is crucial in clusters with frugal Single-Board Computers (SBCs) possessing limited computational resources. These are increasingly being deployed edge computing environments resource-constrained settings where energy efficiency and cost-effectiveness paramount. A major challenge Hadoop YARN scheduling load-balancing, as nodes within the cluster can become overwhelmed, resulting degraded performance frequent occurrences of out-of-memory errors, ultimately leading to job failures. In this study, we introduce an Adaptive Multi-criteria Selection for Resource Allocation (AMS-ERA) Frugal Heterogeneous Clusters. Our criterion considers CPU, memory disk requirements jobs aligns available resources optimal allocation. To validate our approach, deploy a heterogeneous SBC-based consisting 11 SBC conduct several experiments evaluate using wordcount terasort benchmark various workload settings. The results compared Hadoop-Fair, FOG IDaPS strategies. demonstrate significant improvement proposed AMS-ERA, reducing execution time by 27.2%, 17.4% 7.6% respectively benchmarks.

Language: Английский

---

Adaptive Multi-Criteria Selection for Efficient Resource Allocation in Frugal Heterogeneous Hadoop Clusters

Electronics, Journal Year: 2024, Volume and Issue: 13(10), P. 1836 - 1836

Published: May 9, 2024

Efficient resource allocation is crucial in clusters with frugal Single-Board Computers (SBCs) possessing limited computational resources. These are increasingly being deployed edge computing environments resource-constrained settings where energy efficiency and cost-effectiveness paramount. A major challenge Hadoop scheduling load balancing, as nodes within the cluster can become overwhelmed, resulting degraded performance frequent occurrences of out-of-memory errors, ultimately leading to job failures. In this study, we introduce an Adaptive Multi-criteria Selection for Resource Allocation (AMS-ERA) Frugal Heterogeneous Clusters. Our criterion considers CPU, memory, disk requirements jobs aligns available resources optimal allocation. To validate our approach, deploy a heterogeneous SBC-based consisting 11 SBC conduct several experiments evaluate using wordcount terasort benchmark various workload settings. The results compared Hadoop-Fair, FOG, IDaPS strategies. demonstrate significant improvement proposed AMS-ERA, reducing execution time by 27.2%, 17.4%, 7.6%, respectively, benchmarks.

Language: Английский

---