MUMBAI, India, May 29 -- Intellectual Property India has published a patent application (202631048284 A) filed by Manjit Kumar Nayak; Debasis Gountia; Kousik Dehury; and Odisha University Of Technology And Research, Bhubaneswar, Orissa, on April 16, for 'method and system for ddos attack mitigation in iot network using hybrid deep learning.'
Inventor(s) include Manjit Kumar Nayak; Debasis Gountia; Kousik Dehury; Ranjita Champati; Aruna Tripathy; and Meera Viswavandya.
The application for the patent was published on May 29, under issue no. 22/2026.
According to the abstract released by the Intellectual Property India: "Disclosed herein is a method and system for Distributed Denial-of-Service (DDoS) mitigation in IoT network, which employs an IoT device layer (100), an edge device layer (200), a blockchain layer (300), a deep learning model (400) deployed in the edge device layer (200), and an Inter Planetary File System (IPFS) (500). The deep learning model (400) is built on a hybrid architecture comprising a 3D tensor input layer (402), a convolution neural network (CNN) block (404), a Long Short-Term Memory (LSTM) block (406), a transformer block (408), and a classification head (410). The edge device layer (200) acquires a plurality of selective features being indicative network traffic pattern data from the IoT device layer (100) to the edge device layer (200). The deep learning model (400) is configured to: pass the selective features simultaneously from the 3D tensor input layer (402) into the CNN block (404), the LSTM block (406), and the transformer block (408); concatenate feature vector outputs therefrom; and classify the concatenated feature vectors by the classification head (410) to a normal class or a DDoS class, each class being tagged with a corresponding attack complexity score (ACS) in a range of 0 to 5 as determined based on a classification accuracy, a scaling parameter, and a normalization factor. The classification results with the network traffic data are offloaded in the IPFS (500) that generates a content identifier (CID) for each offloading event. Smart contracts are executed on the blockchain layer (300) to perform IoT device registration, anchoring of the CID, data retrieval and verification, and blacklisting of the DDoS class associated IoT devices (102) as compromised or malicious devices in the network; and the compromised/malicious IoT devices (102) are terminated out from the IoT network traffic with a notification to the normal class associated IoT devices (104), while the corresponding ACS crosses a threshold value."
Disclaimer: Curated by HT Syndication.
