26: Internet Measurement Results
"To find out what happens to a system when you interfere with it you have to interfere with it (not just passively observe it)." - George Box"What was observed by us in the third place is the nature or matter of the Milky Way itself, which, with the aid of the spyglass, may be observed so well that all the disputes that for so many generations have vexed philosophers are destroyed by visible certainty, and we are liberated from wordy arguments." - Galileo Galilei.
Lecture outline: Study two research works which will act as a case-study of the method of empirical research
Preamble: Rise of internet measurement based research
1. Main internet measurement research results
Traffic network monitoring:
Internet traffic findings reported by Carey Williamson [2001]
Internet traffic invariants reported by Floyd and Paxson.
Network/ traffic interactions: deterministic component (transmission and propagation delays) and stochastic component (queueing delay).
Internet traffic is self-similar on many orders of magnitude (microseconds till minutes)
Internet traffic has diurnal and weekly trends (scales of hours to days to months)
Long tailed distributions in Internet workloads
Measurement of worm propagation
Application properties:
HTTP traffic properties
DNS traffic results
BGP traffic results
Internet infrastructure:
Router/ AS topology characterization
Network tomography
Traffic Matrix
2. Sound measurement techniques
‘Strategies of sound Internet measurement’ from a paper by Vern Paxson
Whenever possible, calibrate what you measure.
Primary reference for this lecture:
1. “Link-level measurements from an 802.11 b mesh network” by Aguayo et al. [link].
2. “Empirically derived analytic models of wide-area TCP connections” by Vern Paxson [link].
Secondary references for this lecture:
1. “Wide area traffic: the failure of Poisson modeling” by Vern Paxson [link].
2. “Why we don't know how to simulate the Internet” by Vern Paxson [link].