Measurement Infrastructure Survey

• Public Measurement Infrastructure
• "Measuring the Immeasurable: Global Internet Measurement Infrastructure" PAM2001
• NIST site for XIWT/IPWT data

Measurement Data Users

PI Survey Responses	Dataset Used	Research Problem Addressed	Comments
M. Faloutsos & C. Faloutsos, UC Riverside	NLANR/UOregon BGP tables with traceroutes and pings	Interdomain topology modeling and evolution. and end-to-end delay behavior (routing bottlenecks)	completeness of data? need one-way delays and capacities (will check Surveyor)
Don Towsley, UMass	BGP tables	peering relationships	need topology generation
Walter Willinger, AT&T Labs	NLANR/UOregon BGP tables	Check model assumptions	mechanism not consistent with data
Don Towsley, UMass	Auckland WITS	TCP load; end2end loss	need ability to multicast scheduled measurements
Rolf Riedi, Rice	Auckland WITS	TCP load, multiplexing, correlation between traffic streams	need recent traces; multiscale statistics; traffic components; need more info on location of sources
Raghu Kacker, NIST	IEPM, XIWT PingER data
Bruce Siegell, Telcordia	Surveyor delay and traceroute	accurate one-way delay measurement needed for Felix project (ended Aug 2000)	too many routing changes; inconvenient tools; NDA required; need better tools
Nagi Rao, ORNL	homegrown	end2end delay for improving RTT by re-routing	only few daemons; use forced routing to send traffic through them.

NMS Measurement Issues Panel Participants

• Bruce Hajek, University of Illinois
• Walter Willinger, AT&T Labs-Research
• Don Towsley, University of Massachusetts
• Rolf Reidi, Rice University
• David Jaffe, Cisco Systems
• Michalis Faloutsos, UC Riverside
• Shivkumar Kalyanaraman, RPI
• kc claffy, CAIDA

Action Items

• kc: get more information on KEYNOTE and commercial companies providing Internet measurement services
• Chuck Brownstein: Access to XIWT/IPERF data?
• DONE. kc: email to Cisco specifying wishlist for inside router
• kc: Contact Nagi Rao re skitter collaboration or code sharing options