OPTIMIZED NEURAL NETWORK USING CUCKOO SEARCH BASED INTRUSION DETECTION AND PREVENTION SYSTEM IN CLOUD COMPUTING ENVIRONMENT

Nowadays, Cloud Computing is the preferred choice of IT organizations as it provides dynamic and pay-per-use based services to its clients. The main problem of the cloud computing is security and privacy of data which occurs because of its open and distributed structural design. Thus, it becomes essential to design an intrusion detection system (IDS) to provide security to the system. IDS help to protect cloud ecosystem from the malicious activities of the attacker. In this research work, IDS is designed to detect malicious node by using an optimization technique along with the concept of AI (Artificial Intelligence).The Cuckoo Search (CS) algorithm is used as an optimization technique and it is a meta-heuristic approach which is inspired by the behavior of birds and CS algorithm operates on the healthiness function. After the feature optimization, different types of feature are categorized based on the node's nature into two types namely normal and attackers. On the basis of extracted features, train the proposed system using ANN (artificial neural network) as classifier to classify the attackers which affect the networks in cloud environment. ANN is a multiclass classifier which is used to solve multi-class problems and due to this reason ANN is used in proposed work with CS optimization algorithm. In this research work, ANN is used to distinguish between attacker nodes and genuine nodes based on their optimized feature sets. Thus, instead of passing data to the attacker node, the node passes the data to the genuine node and hence, the system is protected. By using the above motioned concept in cloud computing the possibility of results improvement become high because only genuine node involving in the data transmission process. To know the performance of the system, the QoS (Quality of service) parameters such as PDR (Packet delivery ratio), energy consumption rate and total delay with and without prevention algorithm are measured. The development of proposed ARS is done in the MATLAB 2016a software with the help of various toolboxes like data acquisition, artificial intelligence, optimization and curve fitting.