In this paper we make a comparable study of the various types of Reflector Denial of Service attacks popularly known as DRDoS attacks. We discuss their cause, effects, defense mechanisms proposed so far, the effectiveness of these defense mechanisms and their future relevance. We have also shown how reflection attacks are a potential threat to the cloud which is one of the most popular and highly evolving arenas in the Internet.

/pdf/the-innocent-perpetrators-reflectors-and-reflection-attacks-3w3folgspn.pdf

The Innocent Perpetrators: Reflectors and Reflection Attacks

Various embodiments described herein are directed to optimizing cloud computing infrastructures functionality based on an abuse prevention and remediation platform. A tenant profile may have a tenant confidence score for a tenant, the tenant confidence score being an indicator of the reputation of the tenant usage of cloud computing resources. Based on the confidence score of the tenant, one or more policies for the tenant may be identified limiting access to cloud computing resources. If the virtual internet protocol address (VIP) of the tenant is determined to be tainted, the VIP may be quarantined in a tainted VIP pool, the quarantining excluding the VIP from being selected for use until the VIP is clean. A cleanup routine may be executed, the cleanup routine communicating remedial actions for the tainted VIP. Upon completion of the cleanup routine, the VIP may be restored to a clean VIP pool.

Automatically preventing and remediating network abuse

The latest trend in the field of computing is the migration of organizations and offloading the tasks to cloud. The security concerns hinder the widespread acceptance of cloud. Of various, the DDoS in cloud is found to be the most dangerous. Various approaches are there to defend DDoS in cloud, but have lots of pitfalls. This paper proposes a new reputation-based framework for mitigating the DDoS in cloud by classifying the users into three categories as well-reputed, reputed and ill-reputed based on credits. The fact that attack is fired by malicious programs installed by the attackers in the compromised systems and they exhibit similar characteristics used for discriminating the DDoS traffic from flash crowds. Credits of clients who show signs of similarity are decremented. This reduces the computational and storage overhead. This proposed method is expected to take the edge off DDoS in a cloud environment and ensures full security to cloud resources. CloudSim simulation results also proved that the deployment of this approach improved the resource utilization with reduced cost.

/pdf/credit-based-methodology-to-detect-and-discriminate-ddos-1m32rndqok.pdf

Mogankumar P.C

Papers

Credit Based Methodology to Detect and Discriminate DDOS Attack From Flash Crowd in A Cloud Computing Environment