AT&T Home | AT&T Labs, Inc.
Search Research

HOME

WHO WE ARE
About Us
Innovators
Our People
In the Community
Locations
Career Opportunities

WHAT WE DO
Research Areas
Projects
Software Tools
Licensing Portfolio
Patents

COLLABORATION
Universities
Industry

NEWS & INFORMATION
Tech View
Media Library
AT&T Labs News
Publications

Nick Duffield

 

NICHOLAS G DUFFIELD
email: duffield@research.att.com
180 PARK AVE - BUILDING 103
FLORHAM PARK, NJ, 07932




Publications

A complete list of publications is here. Partial indices in DBLP, MathSciNet,

What I do

I am a Distinguished Member of Technical Staff in the Internet and Network Systems Research Laboratory at AT&T Labs-Research in Florham Park, New Jersey. I am an AT&T Fellow and an IEEE Fellow, and a member of the Editorial Board of the IEEE/ACM Transactions on Networking.

I work on measurement, inference and analysis of traffic and performance in the internet, including:

  • Traffic Sampling: I am an inventor of the Smart Sampling methods that underpin AT&T's Traffic Analysis Service, and of Trajectory Sampling, which will enable new classes of measurement applications. I helped start the Packet Sampling (PSAMP) working group in the Internet Engineering Task Force, and was its first chair.
  • Network Tomography: correlating network measurements to infer what can't be measured diriectly. The main applications are inference of traffic matrices (a.k.a. Tomo-Gravity) and of network performance.
  • Active Measurement: (i) tailoring probe streams to reveal the dynamics of network congestion, and (ii) tailoring probing paths in order to to focus on the performance of customer traffic paths.
  • Attack Detection: network measurement infrastructures generate enormous volumes of data.  Adaptive aggregation techniques reduce data volumes while retaining detail for heavy hitters (PANIC). Routine stress testing of traffic aims to help identify misbehaving traffic. Detection sensitivity can be adapted to the ambient intensity of attacks in order to minimize the cost of errors in identifying attacks.

My Background

I receved a B.A. in Natural Sciences (in 1982) and the Certificate of Advanced Study in Mathematics (in 1983) from the University of Cambridge, U.K, and a Ph.D. in Mathematical Physics (in 1987) from the University of London, U.K. I held post-doctoral and faculty positions in Dublin (and here, and here) and Heidelberg before moving to AT&T in 1995. I worked in mathematical physics until the early 90's, then found applications of some of its methods to queueing theory, which got me started in networking...


My Projects

-> Harbinger : Agile Network Monitoring
-> Tomo-gravity : Traffic Matrix Inference
-> Trajectory Sampling (TS)
-> Adaptive Defense : Optimal Use of Available Sensors, Actuators
-> Stress Testing: Inferring Traffic Legitimacy Through Selective Impairment
-> PANIC : Scalable Heavy Hitter Detection
-> Smart Sampling : Accurate Sampling of Flow Records
-> Network Performance Tomography
-> Badabing: Probing the Structure of Network Congestion

My Patents

-> 7536455 Optimal combination of sampled measurements, 05-19-2009
-> 7508769 Consistent sampling for network traffic measurement, 03-24-2009
-> 7437385 Methods and apparatus for detection of hierarchical heavy hitters, 10-14-2008
-> 7299283 Apparatus for size-dependent sampling for managing a data network, 11-20-2007
-> 7293086 Traffic matrix estimation method and apparatus, 11-06-2007
-> 7197048 Virtual private network, 03-27-2007
-> 7080136 Method and apparatus for size-dependent sampling for managing a data network, 07-18-2006
-> 6912232 Virtual Private Network, 06-28-2005
-> 6873600 Consistent sampling for network traffic measurement, 03-29-2005
-> 6452933 Fair queuing system with adaptive bandwidth redistribution, 09-17-2002
-> 6310857 Method and apparatus for smoothing and multiplexing video data flows, 10-30-2001