Modeling data network sessions
Seminar Room 1, Newton Institute
AbstractA session is a higher order entity resulting from amalgamating packets, connections, or groups of connections according to specified but not unique rules. For example, using various rules, the flow of packets past a sensor can be amalgamated into higher level entities called sessions using a threshold rule based on gaps between packet arrivals. We rapidly review some probability modeling based on sessions before turning to statistical analysis. Statistical analysis of these sessions based on packets is complex: session duration (D) and size (S) are jointly heavy tailed but average transmission rate (R=S/D) is sometimes not heavy tailed and arrival times of sessions is not Poisson. By segmenting sessions using a peak rate covariate, we find conditional on a peak rate decile, within this decile segment session initiations can be modeled as Poisson. For modeling the distribution of (D,S,R), the conditional extreme value (CEV) model may be a useful variant. (Joint work at various times with Jan Heffernan, Bikramjit Das, Luis Lopez-Oliveros, T. Mikosch, Bernardo D'Auria, H. Rootzen, A. Stegemen.)
If it doesn't, something may have gone wrong with our embedded player.
We'll get it fixed as soon as possible.