@Article{cmc.2026.076825,
AUTHOR = {Mohammed Saad Javeed, MD AL Rafi, Arifa Akter Eva, Muhammad Firoz Mridha, Qiangfu Zhao, Jungpil Shin},
TITLE = {A Unified API-Driven IPAM Framework with LSTM-Based Anomaly Detection for Hybrid Cloud Environments},
JOURNAL = {Computers, Materials \& Continua},
VOLUME = {},
YEAR = {},
NUMBER = {},
PAGES = {{pages}},
URL = {http://www.techscience.com/cmc/online/detail/26760},
ISSN = {1546-2226},
ABSTRACT = {Hybrid and multi-cloud infrastructures make IP address management (IPAM) difficult, especially when IP and Domain Name System (DNS) records must stay consistent across on-premises networks and cloud platforms. Traditional IPAM tools often lack deep automation and cross-platform visibility, which leads to DNS drift, IP conflicts, and configuration errors. This paper proposes a unified, Application Programming Interface (API)-driven IPAM framework that integrates Infoblox Network Identity Operating System (NIOS) with Amazon Web Services (AWS) Route53 and Azure DNS using Infrastructure-as-Code and CI/CD pipelines. We generate an IPAM event log from Infoblox API simulations and fuse it with the UNSW-NB15 cybersecurity dataset to train a deep Long Short-Term Memory (LSTM)-based anomaly detection model. On this fused dataset, the propose model achieves an accuracy of 0.912 and an F1-score of 0.900, outperforming Random Forest, Logistic Regression, one-dimensional convolutional neural network (1D-CNN), and Gated Recurrent Unit (GRU) baselines. The automation pipeline also reduces end-to-end provisioning latency to 1010 ms and improves DNS consistency to 99.1% across Infoblox, AWS, and Azure. The framework provides scalable, auditable, and policy-driven IPAM while reducing manual work and configuration drift in hybrid cloud environments.},
DOI = {10.32604/cmc.2026.076825}
}