Detection of EDoS attacks in SDN-based Cloud Model using Deep Learning based SDPN Technique

Offering IT services to businesses and consumers via the cloud is increasingly seen as the most cost-effective model. But it is vulnerable to emerging flaws. In particular, a recently identified form of attack known as an economic-denial-of-sustainability (EDoS) takes advantage of the pay-per-use model to gradually increase the utilisation of the cloud's resources, forcing the customer to fork over further money. Therefore, in this learning, we present a novel method for detection and mitigating EDoS assaults in the SDN-based cloud computing setting utilising a Stack Deep Polynomial Network (SDPN). First, the SDPN learns precise representations of multivariate time sequence in order to imitate the typical patterns. Then, the reconstructed input data is compared to the original. Finally, the probabilities derived from the reconstruction process may serve as both a tool for detecting outliers and a source of insight into what those outliers could mean. Existing systems often utilise a hard threshold to examine the anomalies, which leads to growing mistake rates; in contrast, the suggested scheme introduces a threshold value utilising SDPN to lower error rates. When likened to other solutions and our past work, the findings of this extensive analysis indicate remarkable performance.