To communicate organizationally important data among working units in a federated cloud environment, secure mutual authentication is a crucial necessity. The administration of data while preserving its usefulness and security system is a worry for the cloud owner. Cloud data sharing is becoming more and more popular as a viable way to give people simple access to data. supported by positive advances in cloud technology. Yet, as more businesses and customers store their data on cloud servers, it poses a growing threat to user privacy and data security. People are inclined to encrypt information before sending it to the cloud in order to ensure its confidentiality. Yet the data processing has gotten challenging due to broad encryption methods. In order This paper provides a novel mutual authentication mechanism that combines deep learning-based recognition and control of security breaches to enable secure communication between participating entities. Convolutional Neural Networks (CNN) for online threat detection with RSA Cryptography and Schnarr's signature scheme-based public-key encryption technologies. A network intrusion dataset from the Canadian Institute for UNSW-NB15 is used to measure performance, and the AVISPA security analysis tool confirms its effectiveness in comparison to other methods in terms of security features and communication costs.