Font Size: a A A

Research On IP Geolocation Based On Network Measurement

Posted on:2021-09-20Degree:MasterType:Thesis
Country:ChinaCandidate:Y R JinFull Text:PDF
GTID:2518306113951569Subject:Computer Science and Technology
Abstract/Summary:PDF Full Text Request
IP geolocation is a technology to determine the geographic location of a device by IP address,and is widely used in the fields like network security,network management,and targeted advertising.As network devices are increasingly connected to the Internet,the scale of the network continues to expand.Limited by the number of IPv4 addresses,the global network structure presents a complex multi-level topology.Due to the lack of platforms such as Planet Lab and Looking Glass that provide public landmark nodes,high-precision IP positioning technology still faces huge challenges.In addition,the existing IP positioning technology is usually low in reliability,poor in timeliness,or requires a large amount of known landmark node data.The positioning error has not been effectively dealt with yet,and the algorithm accuracy is impacted heavily by the quality of the network environment.As to solve the above problems,the main work of this article is as follows:1.This paper proposed a IP geolocation algorithm based on PoP partitions and weighted least squares algorithm.This algorithm can effectively reduce the cumulative error composed of detection error and iteration error,and it has better robustness to changing network environment,which greatly improves the success rate of the IP positioning algorithm based on PoP partition and the stability of the iteration result.2.This paper proposed a high precision IP geolocation algorithm based on large-scale POI landmark nodes and path similarity.At first,this paper crawled and denoised POI(Point of Interest)data on the online map service,cleaned data by technologies like domain name resolution,and tracked the routing paths from the detection node to each landmark node establishing a path fingerprint database.This paper designed a two-dimensional string subsequence kernel algorithm to calculate the similarity of routing paths and an outlier detection cluster algorithm to optimize the IP positioning results.This algorithm collected 2238 and 1429 available POI landmarks in Beijing and Shanghai,the experimental results show this algorithm significantly improves the accuracy of IP geolocation,achieving a median positioning error of 5.7km.Eventually,this article summarized the work of entire paper and pointed out the next research direction of related technologies.
Keywords/Search Tags:IP geolocation, network topology, path similarity, PoP partition
PDF Full Text Request
Related items