Abstract

As corporate digital services get more connected and essentially rely on cloud channels, high-grade security systems get more essential. To solve this problem research present a Pigeon search-driven extreme gradient boosting (PS-XGboost) approach to security improvement in cloud workflow orchestration. Initially, research gather a financial dataset. Convolutional Neural Network (CNN) which extract critical features from the dataset. This model is a combination of deep learning adaption and generalization skills with good broad classification abilities of PS-XGBoost. The fundamental characteristics that separate the regular as well as the abnormal behavior are identified from the huge dataset of regular workflow patterns and used to train the CNN. Therefore, the proposed model incorporates several advantages in contrast to conventional anomaly detection techniques. Beside the study performance measurements and parameters evaluation, which will help to assess how successful our proposed solutions are under the context of cloud security are assessed. Finally, the approach that research proposed and could give comprehensive view of the anomaly reasons, which could make security teams' job easier by reducing their search period of security flaws.