학술논문

Quantifying regularity of the Internet Interdomain Routing based on Border Gateway Protocol (BGP) data bases
Document Type
Conference
Source
2020 International Conference on Electrical, Communication, and Computer Engineering (ICECCE) Electrical, Communication, and Computer Engineering (ICECCE), 2020 International Conference on. :1-5 Jun, 2020
Subject
Communication, Networking and Broadcast Technologies
Components, Circuits, Devices and Systems
Computing and Processing
Fields, Waves and Electromagnetics
General Topics for Engineers
Power, Energy and Industry Applications
Robotics and Control Systems
Signal Processing and Analysis
Time series analysis
Fractals
Entropy
Fluctuations
Internet
Complexity theory
Process control
border gateway protocol
time series
dynamics
complexity
regularity
Language
Abstract
The time series of border gateway protocol (BGP) updates from 4 autonomous systems (AS) are analyzed. Methods of modern data analysis, such as multifractal detrended fluctuation analysis (MF-DFA) and refined composite multiscale entropy (RCMSE) are used. Analysis is carried out on the entire BGP updates data sets, as well as on their shorter, most regular parts. It is found that the process of BGP updates variability, though reveals general similarity in the dynamics of behavior, all the same is very specific for each of the considered data sets. It is shown that the variability of NTT and Tinet BGP updates is the closest one to monofractal behavior, though the same data sets strictly differ by their extent of complexity. The variability of BGP updates for ATT is the most multifractal one compared to data sets from other AS-es. It is characterized by a lesser extent of complexity compared to BGP updates from Tinet, NTT and IJJ As-es, according to entropy measure calculations. Further analysis, carried out on shorter parts of BGP time series, additionally shows essential differences for different AS-es. Namely, Tinet is characterized by the highest extent of multifractality and a lesser entropy measure, i.e. is the most ordered one compared to BGP updates from other AS-es. A conclusion is made that that BGP updates time series from different AS-es are characterised by a different extent of regularity and multifractality, even if observation is carried out for the same time period. Moreover, for the first time is shown that the dynamical features of variability for different AS-es may strictly differ even for shorter periods when the BGP updates process is most regular over the entire observation period.