Big Data has come a long way from its conceptual stage to becoming one of the primary pillars of current business and governmental information processing technologies. Modern enterprises face a challenge, because the existing data processing technologies are not capable of handling the growing volumes of data generated by interconnected devices and application transactions at a constantly accelerating pace. This research presents a comprehensive technical and pragmatic investigation into big data technologies and approaches, concentrating specifically on the particular issue of the inability of traditional data architectures to support analytical processes and decision making. The paper discusses in detail different data architecture components including Hadoop, Apache Spark, Kafka, and cloud based data lakes. By means of comparison between traditional and novel data architectures, as well as using sophisticated mathematical models and architectural diagrams, this research demonstrates that implementation of Big Data technologies results in decreased latencies in analytics orders of magnitude, up to 40-60% lower costs for infrastructure, and predictive capabilities not available in traditional batch processing architectures.
Pradhan et al. (Thu,) studied this question.