Behavior Analysis of IP Hosts Based on the Temporal Social Relationship in a Managed Network
碩士 === 國立中正大學 === 資訊工程研究所 === 106 === Profiling network behavior of Internet hosts is important for network management and security monitoring. It is useful in understanding the behavior of hosts and detecting abnormal activities. In this work, we use bipartite graphs and one-mode projection to buil...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Others |
| Language: | en_US |
| Published: |
2018
|
| Online Access: | http://ndltd.ncl.edu.tw/handle/pbps94 |
| Summary: | 碩士 === 國立中正大學 === 資訊工程研究所 === 106 === Profiling network behavior of Internet hosts is important for network management and security monitoring. It is useful in understanding the behavior of hosts and detecting abnormal activities. In this work, we use bipartite graphs and one-mode projection to build a social relationship graph of hosts in a managed network and perform a clustering algorithm to group hosts that often accessed the same services for a long time. After that, we analyze the behavior of clusters based on the common services of hosts in the cluster. The experiment results demonstrate that our method can discover clusters of hosts that often accessed the same services for many days from a real network traffic of primary and secondary schools of a county in Taiwan Academic Network. We can summarize the behavior of hosts and detect outlier clusters with unusual behavior.
|
|---|