The paper "Securing Mobile Edge Computing Using Hybrid Deep Learning Method" by Olusola Adeniyi, Ali Safaa Sadiq, Prashant Pillai, Mohammad Aljaidi, and Omprakash Kaiwartya addresses the security challenges posed by Distributed Denial of Service (DDoS) attacks in Mobile Edge Computing (MEC) environments. The authors highlight the limitations of shallow Machine Learning (ML) models in effectively detecting and mitigating DDoS attacks, which are prevalent in MEC due to its low latency and high bandwidth capabilities. To overcome these limitations, they propose a novel hybrid Autoencoder-Multi-Layer Perceptron (AE-MLP) model for intrusion detection. The AE-MLP model leverages the feature extraction capabilities of autoencoders to capture intricate patterns and anomalies in network traffic data. The extracted features are then fed into a Multi-Layer Perceptron (MLP) network for further analysis and classification of potential threats. This integration of AE and MLP enhances the model's accuracy and robustness in identifying DDoS attacks while minimizing false positives. Extensive experiments using the NF-UQ-NIDS-V2 dataset, which contains a wide range of DDoS attacks, demonstrate that the proposed hybrid AE-MLP model achieves a high accuracy of 99.98%. The model outperforms several similar techniques, including shallow ML models like Random Forest and Support Vector Machine (SVM), which achieve accuracies of 99.59% and 68.70%, respectively. The paper also discusses the advantages and drawbacks of the AE-MLP hybrid approach. The hybrid model excels in anomaly detection and can learn patterns from large datasets, making it more adaptable to evolving attack techniques. However, the computational intensity of the hybrid model, with a training time of 4.96 seconds, is a significant drawback. To mitigate this, the authors suggest cloud-edge collaboration, where training is conducted in the cloud and the trained model is deployed to the edge for prediction. In conclusion, the proposed hybrid AE-MLP model provides an effective solution for DDoS attack detection in MEC environments, offering real-time decision-making and optimized latency and bandwidth utilization. Future work could focus on improving the efficiency of the hybrid model and exploring more sophisticated attack detection techniques.The paper "Securing Mobile Edge Computing Using Hybrid Deep Learning Method" by Olusola Adeniyi, Ali Safaa Sadiq, Prashant Pillai, Mohammad Aljaidi, and Omprakash Kaiwartya addresses the security challenges posed by Distributed Denial of Service (DDoS) attacks in Mobile Edge Computing (MEC) environments. The authors highlight the limitations of shallow Machine Learning (ML) models in effectively detecting and mitigating DDoS attacks, which are prevalent in MEC due to its low latency and high bandwidth capabilities. To overcome these limitations, they propose a novel hybrid Autoencoder-Multi-Layer Perceptron (AE-MLP) model for intrusion detection. The AE-MLP model leverages the feature extraction capabilities of autoencoders to capture intricate patterns and anomalies in network traffic data. The extracted features are then fed into a Multi-Layer Perceptron (MLP) network for further analysis and classification of potential threats. This integration of AE and MLP enhances the model's accuracy and robustness in identifying DDoS attacks while minimizing false positives. Extensive experiments using the NF-UQ-NIDS-V2 dataset, which contains a wide range of DDoS attacks, demonstrate that the proposed hybrid AE-MLP model achieves a high accuracy of 99.98%. The model outperforms several similar techniques, including shallow ML models like Random Forest and Support Vector Machine (SVM), which achieve accuracies of 99.59% and 68.70%, respectively. The paper also discusses the advantages and drawbacks of the AE-MLP hybrid approach. The hybrid model excels in anomaly detection and can learn patterns from large datasets, making it more adaptable to evolving attack techniques. However, the computational intensity of the hybrid model, with a training time of 4.96 seconds, is a significant drawback. To mitigate this, the authors suggest cloud-edge collaboration, where training is conducted in the cloud and the trained model is deployed to the edge for prediction. In conclusion, the proposed hybrid AE-MLP model provides an effective solution for DDoS attack detection in MEC environments, offering real-time decision-making and optimized latency and bandwidth utilization. Future work could focus on improving the efficiency of the hybrid model and exploring more sophisticated attack detection techniques.