Deep Learning with an LSTM-Based Defence Mechanism for DDoS Attacks in WSNs
P. Manju Bala, S. Usharani, Tamilarasan Ananth Kumar, Rajendrane Rajmohan, P. Praveen Kumar · 2021
A wireless sensor network (WSN) contains a set of sensing points used to track and store the external environment status and coordinate the collected information at a centralised point. WSNs involve lightweight, self-governing devices that are focused on batteries that are implemented in a distributed way to track external or environmental conditions. Deep learning with the Distributed Denial of Service (DDoS) security mechanism based on long short-term memory (LSTM) has suggested detecting and isolating attacks in the information-forwarding process. Deep learning with the LSTM-based system to achieving energy consumption and optimized load balancing at the WSN fusion centre is described in this chapter. To facilitate several practical applications, the nodes and access points are linked to the distributed routers. Because of open data, the WSN is facing a security challenge. In this scenario, external users can be checked by ensuring authentication is required. Several lightweight authentication mechanisms in practical applications have been developed to achieve safe communication. However, WSNs are extremely vulnerable to DDoS attacks because during data routing, there is no synchronisation between nodes. The original algorithm for effective detection of DDoS attacks, for example, jamming, homing, flooding and jamming, is also mentioned in this chapter, as are test results that can isolate adversaries reliably and are more robust to DDoS attacks.