Detection and Mitigation of Induced Low Rate TCP-Targeted Denial of Service Attack

Abstract

Low rate TCP-targeted denial of service attack is a cleverly crafted attack in which an attacker exploits congestion avoidance algorithm and uniformity of min RTO in Transmission Control Protocol (TCP). Attacker congest the network for a brief period of time then keep quiet for some time. This phenomenon is repeated after min RTO time. This attack causes degradation of service and denial of service to those TCP flows which satisfies certain condition. Attacker Launches this attack by exploiting the technique of optimistic acknowledgement which is used for sending of acknowledgement before data has been received. By this technique attacker induces server to perform the attack. Ever since the discovery of this attack, lot of solution, detection scheme have been proposed, each having their own merits and demerits. Mostly these schemes are complex and not scalable. In this paper a novel scheme has been proposed which reduces random bytes from a random TCP segment to verify the authenticity of those optimistic acknowledgement. As attacker does not know segment size whenever he sends optimistic acknowledgement it is dropped. Thus the attack can be mitigated using this technique.