Skip to Content
[CAIDA - Center for Applied Internet Data Analysis logo]
Center for Applied Internet Data Analysis > publications : bib : networking : entries : broido03radon.xml
Bibliography Details
A. Broido, R. King, E. Nemeth, and k. claffy, "Radon Spectroscopy of Inter-Packet Delay", in IEEE High-Speeding Networking Workshop (HSN), March 2003.
Radon Spectroscopy of Inter-Packet Delay
Authors: A. Broido
R. King
E. Nemeth
k. claffy
Published: IEEE High-Speeding Networking Workshop (HSN), 2003
Entry Date: 2003-01-30

We demonstrate the feasibility of Internet spectroscopy techniques for analysis of rate limiting, packet interarrival delay and passive bitrate estimation of cell- or slot-based broadband connections. Working with highly diverse packet trace data, we find that delay's quantization in micro- and millisecond range is ubiquitous in today's Internet and that different providers have strong preferences for specific delay quanta in their infrastructures.


We use packet traces collected at an OC-48 link of an US Tier 1 provider in January and March 2002. The traces are captured by a Linux monitor equipped by Dag 4 cards from Endace, with CDMA-syncronized timestamps.


We describe an experiment in which we study interarrival times of packets sent from a home DSL connection to a university machine. The experiment confirms that DSL is using ATM as a layer 2 encapsulation. The use of ATM cells results in the quantization of the observed interarrival delay.


We formulate bitrate estimation for slot- or cell-based links as a statistical inverse problem, which we solve by entropy minimization for the Radon-transformed marginals of size-delay (packet size vs. packet interarrival time) distributions.

We find that typical delay quanta for DSL are 2.65 ms (for 160 Kbps upstream bitrate) and 3.31 ms (for 128 Kbsp bitrate). For cable infrastructures, we find that delay quanta is most often equal to 2, 3, and 6 ms.

We also show that in the whole volume of traffic traversing a high-speed link, the packet interarrival times are close to exponentially distributed. This property makes Poisson traffic models an attractive choice for the combination of speed and multiplexing present in our data.


A version of this paper appears as: A.Broido, R.King, E.Nemeth, kc claffy "Radon Spectroscopy of Packet Delay", Processings of 18-th International Teletraffic Congress (ITC18), Berlin, Sept. 2003, Elsevier

Stephen Donnelly. "High precision timing in passive measurements of data networks." Ph.D. Thesis, University of Waikato, New Zealans, June 2002.

J.Micheel, S.Donnelly, I.Graham. "Precision timestamping of network packets." Proceedings of IMW, Nov.2001.

  Last Modified: Tue Jul-28-2020 14:29:16 UTC
  Page URL: