To obtain access to full text of journal and articles you must register!
- Article name
- Detection and selfsimilarity evaluation of network traffic stationary fragments
- Authors
- Peryshkin S. V, , Sergey.Peryshkin@gmail.com , sp@ntlab.ru, Society with Limited liability "Laboratory of Network Technologies", Moscow, Russia
- Keywords
- network traffic analysis / network behavior analysis / IT infrastructure behavior analysis / stationarity evaluation / Hurst exponent
- Year
- 2010 Issue 3 Pages 42 - 49
- Code EDN
- Code DOI
- Abstract
- Article is concerned with the problem of IT infrastructure network traffic analysis with the purpose of network objects behavior regular patterns detection. One of the most relevant tasks is development of complex network behavior analysis methods which do not have strict restrictions of applicability. An approach is proposed for extraction of traffic characteristics value ranges and its' combinations allowing sectioning of overall network traffic into persistent fragments, grouping of fragments achieved and stationarity pre-estimation of resulting streams. For the evaluation of stationarity grade the Hurst exponent was used, which was calculated by several different methods, and all the results was compared. Experiments taken over the real traffic of university network have proven the workability of such approach. Self-similarity degree appeared to be high for a number of network traffic characteristics combinations. Among the advantages of the approach: demonstrativeness increase of behavior profiles model, the ability to localize anomalies in early stages of traffic analysis, improvement of accuracy and productivity of network objects behavior analysis.
- Text
- To obtain access to full text of journal and articles you must register!
- Buy
