Management of Distributed Denial of Service Attack in Cloud Computing Environment

Abstract

Cloud Computing is a recent technology, it provides a simple and unambiguous taxonomy of three service models available to cloud consumers: software as a service (SaaS), platform as a service (PaaS), and infrastructure as a service (IaaS). There are several security issues with the delivery model of cloud. Our work is to dealing with management of Distributed Denial of Service attack on SaaS model of cloud computing environment. If DDoS attack is capable enough to violate the Service Level Agreement (SLA) on availability it can cause huge financial claim and it will affect the reputation of industries in a market. So our basic aim is to design a management model that will avoid the SLA violation on availability due to a DDoS attack. Our model works in three stages (1) Detection of DDoS attack (2) Avoidance of DDoS attack and (3) prevention of DDoS attack. Feedforward Neural Network method for detection of DDoS attack. Sigmoid function is used as Neural modal for obtaining the desire output. The Supervised learning model adjusts the connection weight and bias value of ANN model. Using predefined datasets to train the ANN model. For the Avoidance of DDoS attack data center dynamically allocate the resources on virtual machines. A new virtual machine will be clone based on the image file of the original. Replicate the resources on a virtual machine in order to avoid the SLA violation (Availability issues). Message Authentication Code (MAC) is used for prevention of DDoS. The message Authentication code increases the overhead on the network. The design goal is to decrease the overhead of MAC on the network so we are using Router Packet Filtering method that reduces that MAC overhead on packet over the network. This lower overhead increases the speed of authentication and reduces the amount of dynamically allocated resources that will prevent the violation of the SLA on cloud computing.