INSPIRE-5Gplus-related Aalto paper presented at WCNC 2020

On 28 May, Chafika Benzaid from Aalto University presented an INSPIRE-5Gplus-related paper online at the IEEE Wireless Communications and Networking Conference (WCNC 2020). The paper, which she co-authored with her Aalto colleagues Mohammed Boukhalfa and Tarik Taleb, is titled “Robust Self-Protection Against Application-Layer (D)DoS Attacks in SDN Environment”. The paper was presented on the third day of WCNC 2020 in a session on “Security and privacy” (Session T3-S27).

In the paper, the authors propose a new framework for fully autonomous detection and mitigation of app-layer DDoS attacks, leveraging Deep Learning (DL) for attack detection and using SDN controller for mitigation policy enforcement. A key contribution of the work is to build a DL model that is resilient to attacks against Machine Learning (ML). The performance results show the effectiveness of the proposed framework in autonomously protecting against application-layer DDoS attacks even in the presence of adversarial attacks.

In the Q&A after the presentation, two questions came up regarding the features extraction process and the main novelty compared to existing work using DL and SDN. For the first question, Chafika Benzaid explained the extraction process and the number of features used by the DL model. For the second question, she emphasized the fact that the proposed solution is fully autonomous and considering the security of ML techniques, something that was overlooked in previous contributions.