期刊名称：International Journal of Computer Trends and Technology
电子版ISSN：2231-2803
出版年度：2016
卷号：41
期号：1
页码：10-15
DOI：10.14445/22312803/IJCTT-V41P102
出版社：Seventh Sense Research Group
摘要：In the tremendous growth of internet world, networking communications play an important role. Network communication is one of the sharing of information between server and clients. But in today fast technology, the number of clients has increased and consequently the server is unable to send the response to all the legitimate clients in time. It may also happen due to the attack of intruders. So the prevention of this kind of attacks is the important aspect throughout the network communication. Specifically, unsupervised data mining clustering techniques allow to effectively distinguishing the normal traffic from malicious traffic in a good accuracy. In this paper, a bird view for a set of probabilistic packet marking methodologies has been discussed to prevent the DDOS attacks using clustering algorithm during peak hours of server. These various methodologies are useful to find the IP address of clients and find the intruders among them depending upon the client’s behavior. And also we envision DDoS attack starts when network traffic is more than our default threshold. In this type of packet marking protocol, packets are marked based on predefined probability.
