Smart ShuffleLink: A Spark-Driven Architecture for Cyber Attack Detection and Mitigation under Bigdata Environment
In big data context, cyber-attack detection (CAD) and mitigation process employs big data analytics to evaluate massive amounts of network data in real-time, using modern algorithms to spot abnormalities and predict malicious activity. This research develops a novel Spark-Driven Architecture for CAD and Mitigation under Bigdata Environment. Primarily, the Re-descending Function-based Improved Gaussian Function (RF-IGF) approach is deployed for processing the CAD big data. To handle big data, Spark framework is used in this study. In spark, the master node handles data partitioning, while, the slave nodes handle feature extraction and feature fusion. Features like Mutual information (MI), statistical features and raw data are extracted in this work. The extracted features are fused using a new improved SoftMax with improved Local Response Normalization based Cross attention module (ISILRN-CAM). The fused features are then fed to hybrid classifier that encompasses hybrid ShuffleNet and LinkNet models together termed as ShuffleLinkNet. After detection, the scores attained from individual ShuffleNet and LinkNet are subjected to enhanced Score Level Fusion (ESLF), which fuses the score and results in fine detected output. Once the attack is detected, entropy-based cyber-attack mitigation is carried out to mitigate the impact of attacks. From analysis, the accuracy results using ShuffleLinkNet with ESLF reached a higher value of 0.965.
Authors
- Mantri Pavan Kalyan
- Sadhu Radha Krishna
Publication Details
- Journal
- Discrete Mathematics Algorithms and Applications
- Published
- 2026-11-19
- DOI
- https://doi.org/10.1142/s1793830926501193
- Primary Topic
- Network Security and Intrusion Detection
- Type
- article
- Field-Weighted Citation Impact
- 0.00