Authors: Sudhir Vishnubhatla
Abstract: Information Management Systems (IMS) have historically operated in centralized architectures where ingestion, storage, and retrieval workflows were executed in tightly controlled environments. However, the rapid growth of digital documents in regulated domains such as finance, healthcare, and public archives demands real-time processing, semantic enrichment, and compliance-aware access. The emergence of Edge AI deploying lightweight intelligence at the data source—combined with hyperscale cloud services now offers a hybrid path forward. This article synthesizes research from 2000–2024, spanning early distributed file systems, service-oriented architectures, edge intelligence frameworks, and cloud-native analytics. We propose a layered architecture for real-time document understanding in IMS that leverages edge devices for low-latency inference while relying on the cloud for scalability, orchestration, and governance. Three illustrative figures demonstrate the evolution from reference edge-cloud topologies to optimized deployment pipelines, culminating in end-to-end IMS analytics integration.
DOI: http://doi.org/10.5281/zenodo.17452281
