Scalable Data Mining Architecture for Real-Time Big Data Analytics Using Distributed Frameworks
International Journal of Advanced Research in Science Communication and Technology · 12.07.2026 · 536
DOI: 10.48175/ijarsct-37658 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Advanced Research in Science Communication and Technology, online erschienen am 12.07.2026. Verfasst von Dr. Amit K. Mogal. Die Arbeit nennt 46 Literaturangaben.
Abstract (Originalsprache)
The rapid growth of high-velocity and heterogeneous data generated from sources such as IoT devices, social media, and enterprise systems has created significant challenges for traditional data processing approaches. Real-time big data analytics requires scalable architectures capable of processing continuous data streams with low latency and high throughput. This study proposes a scalable data mining architecture for real-time big data analytics using distributed frameworks, specifically Apache Kafka, Apache Flink, and Apache Spark. The architecture adopts a hybrid, layered design that integrates data ingestion, stream processing, distributed storage, and analytics components to enable efficient and reliable data processing. The proposed model leverages Apache Kafka for high-throughput data ingestion, Apache Flink for true stream processing with low latency, and Apache Spark for micro-batch processing and advanced analytics tasks. This hybrid approach addresses the limitations of single-framework systems by combining the strengths of multiple distributed technologies. The architecture is implemented in a distributed environment and evaluated using real-time streaming scenarios to assess performance across key metrics, including latency, throughput, scalability, and resource utilization. Experimental results demonstrate that the proposed architecture achieves significant improvements in real-time processing performance, particularly in reducing latency and enhancing throughput as data volume and cluster size increase. Apache Flink outperforms Spark Streaming in latency-sensitive tasks, while Spark provides flexibility for iterative and machine learning workloads. Statistical validation using ANOVA confirms that the observed performance differences are statistically significant, reinforcing the reliability of the results. The architecture also exhibits strong horizontal scalability and fault tolerance, making it suitable for large-scale, data-intensive applications. The findings highlight the effectiveness of hybrid distributed frameworks in addressing the challenges of real-time big data analytics and demonstrate the importance of integrating multiple processing paradigms within a unified architecture. The proposed model contributes to advancing scalable and efficient data mining systems and provides a foundation for future research in areas such as AI-driven data pipelines, edge computing, and energy-efficient distributed systems.
Zitieren
APA
Mogal, D. A. K. (2026). Scalable Data Mining Architecture for Real-Time Big Data Analytics Using Distributed Frameworks. International Journal of Advanced Research in Science Communication and Technology, 536. https://doi.org/10.48175/ijarsct-37658
Vancouver
Mogal DAK. Scalable Data Mining Architecture for Real-Time Big Data Analytics Using Distributed Frameworks. International Journal of Advanced Research in Science Communication and Technology. 2026:536. doi:10.48175/ijarsct-37658
BibTeX
@article{Mogal2026,
author = {Mogal, Dr. Amit K.},
title = {Scalable Data Mining Architecture for Real-Time Big Data Analytics Using Distributed Frameworks},
journal = {International Journal of Advanced Research in Science Communication and Technology},
year = {2026},
pages = {536},
doi = {10.48175/ijarsct-37658}
}
EUVIMED