In theinfrastructure of the Internet of Things (IoT), the attack along with anomaly detection (outlier detection) is an intensifying matter in the IoT domain. The intricacy and sternness of Distributed Denial of Service (DDoS) attacks are augmenting daily amidst every sort of attack. For spotting such attacks, divergent machine learning-related methods are proposed. But the prediction accurateness of the prevailing techniques is very minimal. Hence, in this project paper, ODLNN utilized IDS are proposed for DDoS attacks identification in the cloud IoT environs. The phase of training is first implemented, and then, in the proposed ODLNN, the testing phase is executed. The system has been trained by the suggested methodology with numerousprocedures (Preprocessing, HDFS, MapReduce, and Classification) for classifying the data, such as normal and also attacked data. Once the training is accomplished in the testing segment, the sensor values from divergentIoT sensor devices are verified and categorized as normal and also attacked data by equating the training outcomes. While the testing outcomes comprise the normal data, the data encryption utilizing MCSA-ECC is hoarded in the cloud, or else the data is hoarded as a log file of the cloud for being utilized in the upcoming times. Lastly, the IoT encrypted data is decrypted by the IoT applications in the cloud, and for further procedures, the real data is utilized. The proposed methodology’s results are assayed and analogized with prevailingtechniques for assuring those in the cloud, and the DDoS attack is perceived by the suggested ODLNN more proficiently.
Volume 12 | 02-Special Issue
Pages: 971-991
DOI: 10.5373/JARDCS/V12SP2/SP20201156