In this  framework design paper, we  designed to address the challenges of DDoS
attacks. It is very complicated to discern which detection approach should be
followed for a circumstantial dilemma. Signature-based detection approach can
disclose only known attacks and results in high detection accuracy with the low
false notifications. But the attacker can quickly adjust the attack signatures
or perform attacks with small variations. Therefore, the attacks remain
unidentified by this approach. Nowadays, anomaly-based detection approach has
been widely used for the detection of Net-DDoS as well as App-DDoS attacks. The
key challenges for this approach are online analysis, manipulating a huge
amount of data and the increasing false signal ratio due the presence of
uncertainty in data. Supervised and semi-supervised techniques are fancied for
controlling the huge amount of data but unsupervised techniques are adopted for
catching unfamiliar attacks. Nevertheless, such schemes do not fit for the
real-time detection. Therefore, implementing a mixed approach of supervised and
unsupervised techniques that can recognize both unknown and known DDoS attacks
in the real-time environment is a challenging task. From this review paper, we
have concluded that the researchers have stated different defense mechanisms
against the DDoS attacks. But due to lack of benchmarks against which the
performance of defense tools may be compared, the best solutions for defending
against such attacks are improbable.

a future research direction, we strongly believe that a perfect comprehensive
realtime defense framework could be the best and effective approach to battle
DDoS attacks. Building a defense mechanism as close as to the attack source
with an evitable participation of various service providers offering a source
address validation and filtering features, we hope to find sooner in the near

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!

order now