The Research Of Intrusion Detection Based On Dynamic Adaptive Samplingunder IPv6Environment

With the increase of network data transfer rates, intrusion detection systems are under increasingly heavy burden. At the same time,, it is at the stage of transition from IPv4to IPv6, how to detect the intrusion and attack more effectively in the next generation of Internet for huge amount of data is the key point of this thesis.Specific work of this thesis is as follows:1. Design suitable strategy in dealing with network data under IPv6environment. According to the extracted information to distribute network data to different intrusion detection unit by using splitting method with lower conflict rate. And define the initial sampling rate of data which has different attributes by design proper black list policy. And then to the cache for network packets through the buffer pool.2. Verify the effectiveness of method to perform intrusion detection by using sampling method under IPv6high speed environment, and make some improvement for the sampling methods. The distribution of intrusion packets in the overall sample space can be divided into two different categories. One is to meet random uniform distribution, and the other is to meet the distribution with different random probability. Pure random sampling, system sampling and merge sampling were used to perform intrusion detection for the former situation. The dynamic adaptive sampling was used to perform intrusion detection for the latter, and the sampling parameters ware changed to verify the impact on the detection results. Finally, this thesis verifies the effectiveness of sampling intrusion detection method by analyze the experimental results.