کنفرانس بین المللی فناوری اطلاعات و دانش

صفحه اصلی / شانزدهمین کنفرانس بین المللی فناوری اطلاعات و دانش

An Attention-Enhanced Hybrid Deep Learning Framework for Detecting Denial-of-Wallet Attacks in Serverless Platforms

نویسندگان :

Mohammad Mehmandoost¹ HadiShahriar Shahhoseini²

1- دانشگاه علم و صنعت ایران 2- دانشگاه علم و صنعت ایران

کلمات کلیدی :

Denial-of-Wallet،serverless computing،cybersecurity،deep learning

چکیده :

Denial-of-Wallet (DoW) attacks pose a significant threat to serverless computing frameworks, as they exploit auto-scaling capabilities to inflate operational costs. This paper presents a hybrid deep learning framework, for effective detection of Denial-of-Wallet attacks. Our architecture couples Convolutional Neural Networks for feature extraction with Long Short-Term Memory networks and Gated Recurrent Units to learn temporal dependencies and a Bahdanau attention mechanism to enhance feature relevance. To address class imbalance and improve feature consistency, data are preprocessed using the Synthetic Minority Oversampling Technique and Standard Scaler. Experimental results demonstrate our model achieving a detection rate of 98.3\% with one of the quickest inference times compared with similar methods while concurrently maintaining a lightweight architecture with fewer parameters. These features make our model specially suitable for real-time detection of Denial-of-Wallet attacks in serverless environments.