Back to Search
Start Over
AliasClassifier: A High-Performance Router Alias Classifier.
- Source :
- Electronics (2079-9292); May2024, Vol. 13 Issue 9, p1747, 25p
- Publication Year :
- 2024
-
Abstract
- The task of router alias resolution for IPv4 networks presents a formidable challenge in the realm of router-level topology inference. Despite the considerable potential exhibited by machine-learning-based alias-resolution methods for IPv4 networks, several constraints impede their effectiveness. These constraints include a low discovery rate of aliased IPs, a failure to account for router aggregation, and a dearth of valid features in current schemes. In this study, we introduce a novel alias resolver, AliasClassifier, which is based on the Random Forest model and the alias triangulation algorithm. This innovative model identifies the key six features from a set of four prevalent routing behaviors that are typically employed to distinguish aliased IPs from non-alienated IPs. Subsequently, the AliasClassifier aggregates aliased IP pairs into routers using an alias triangulation algorithm. Experimental results demonstrate that AliasClassifier excels in discovering aliased IPs in IPv4 networks, boasting a resolution accuracy as high as 94.8% and a recall rate of 40.4%. Its comprehensive performance significantly surpasses that of state-of-the-art alias resolvers such as TreeNET, MLAR, and APPLE. Furthermore, as a typical centralized alias parser, AliasClassifier's deployment cost is remarkably low. Consequently, AliasClassifier emerges as an ideal tool for router alias resolution in large-scale IPv4 networks. [ABSTRACT FROM AUTHOR]
- Subjects :
- IP networks
RANDOM forest algorithms
TRIANGULATION
TOPOLOGY
ALGORITHMS
Subjects
Details
- Language :
- English
- ISSN :
- 20799292
- Volume :
- 13
- Issue :
- 9
- Database :
- Complementary Index
- Journal :
- Electronics (2079-9292)
- Publication Type :
- Academic Journal
- Accession number :
- 177180185
- Full Text :
- https://doi.org/10.3390/electronics13091747