The explosion of data from IoT devices, on social platforms, or transactional systems has increased the demand on real time big data processing. This paper presents an extensive survey and improvement over the recent architectures, stream processing platforms and smart frameworks that have been proposed for processing high velocity data in a scalable and sustainable way. Priority in this space is given to systems which focus on tight high-efficiency scaling with low latency and high throughput. Based on recency research from the years 2020–2025 this chapter discusses new developments on tools such as Apache Kafka, Flink, Spark Streaming, and integration of explainable AI and edge computing for real-time analytics. A hybrid modular architecture is illustrated, which integrates distributed processing, machine learning and NoSQL storage to cope with varying requirements of the finance, healthcare and smart energy domains. It also reviews major challenges like system integration, resource constraints, and data quality and provides approaches to enhance efficiency with less cost to the environment and operations.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Energy-Efficient Architectures and AI-Driven Strategies for Real-Time Big Data Processing

  • Ikram Lefhal Lalaoui,
  • Essaid El Haji,
  • Mohamed Kounaidi

摘要

The explosion of data from IoT devices, on social platforms, or transactional systems has increased the demand on real time big data processing. This paper presents an extensive survey and improvement over the recent architectures, stream processing platforms and smart frameworks that have been proposed for processing high velocity data in a scalable and sustainable way. Priority in this space is given to systems which focus on tight high-efficiency scaling with low latency and high throughput. Based on recency research from the years 2020–2025 this chapter discusses new developments on tools such as Apache Kafka, Flink, Spark Streaming, and integration of explainable AI and edge computing for real-time analytics. A hybrid modular architecture is illustrated, which integrates distributed processing, machine learning and NoSQL storage to cope with varying requirements of the finance, healthcare and smart energy domains. It also reviews major challenges like system integration, resource constraints, and data quality and provides approaches to enhance efficiency with less cost to the environment and operations.