Analytic Distribution Model for Local Side Round Trip Time

Round trip time of local side network, so call Local Round Trip Time (LRTT), plays important role in network tomography and network management. It has been widely used, however none of the previous work focused on the statistical distribution of LRTT. We propose a model of LRTT’s distribution where LRTT is divided into 4 parts, 3 observable parts of them were analyzed through the mechanism of TCP protocol and given in analytic form. The unobservable part is also retrievable after solving the inverse problem. Furthermore, in order to decompose the principle parts of LRTT’s distribution, a simplification algorithm for the most common host and protocols was induced. The experimental result statistically validate the model and simplifying algorithm by adopting collected data in real world.

[1]  Jin Cao,et al.  Modeling time correlation in passive network loss tomography , 2011, 2011 IEEE/IFIP 41st International Conference on Dependable Systems & Networks (DSN).

[2]  Kevin Jeffay,et al.  Variability in TCP round-trip times , 2003, IMC '03.

[3]  Paul Barford,et al.  Network radar: tomography from round trip time measurements , 2004, IMC '04.

[4]  Donald F. Towsley,et al.  Inferring TCP connection characteristics through passive measurements , 2004, IEEE INFOCOM 2004.

[5]  Kang Li,et al.  New Methods for Passive Estimation of TCP Round-Trip Times , 2005, PAM.

[6]  Hao Jiang,et al.  Passive estimation of TCP round-trip times , 2002, CCRV.

[7]  Wei Zhang,et al.  Modeling End-to-End Delay Using Pareto Distribution , 2007, Second International Conference on Internet Monitoring and Protection (ICIMP 2007).

[8]  Ata Shingo,et al.  Using Mixed Distribution for Modeling End-to-End Delay Characteristics , 2005 .

[9]  Dario Rossi,et al.  Passive analysis of TCP anomalies , 2008, Comput. Networks.

[10]  Paul Barford,et al.  On the Performance of Round Trip Time Network Tomography , 2006, 2006 IEEE International Conference on Communications.