Measurement in the Internet
Paul Barford
University of Wisconsin - Madison
Spring, 2001
Why measure the Internet?

The size, growth, complexity and diversity of the Internet
make it impossible to understand, manage, protect or
provision without measurement.
 Prior measurement studies have taught us a great deal
– Self-similarity of packet traffic
– Ubiquity of power laws

Businesses use measurements to provision, manage and
operate
 Future developments will require more and/or different
measurements
What can be measured in the
Internet?

Structure
– Topology, routing, CDN’s, wireless, etc.

Traffic
– Transport, end-to-end performance, etc.

Users and Applications
– WWW, (x,my)Napster, Peer-to-Peer,Streaming, security, etc.

Failures
– In all areas

Nefarious behavior
– Pattern attacks, port scans
Where are measurements made in
the Internet?

For some measurements, this is obvious
– Web logs

The goal for other measurements is to be
“representative”
– Various “Internet weather reports”

Placement of measurement nodes is not a well
understood problem
– More is better??
How are Internet measurements
made?

Passive methods
– Application monitors (logs), packet monitors

Active methods
– Probes, application simulation

Surveys
 Significant infrastructure is always required
 All methods present difficulties
When must Internet
measurements be made?

Diurnal traffic cycle
 Time scales depend on “what” and “how”
 Passive measurements are typically continuous
– Can generate huge data sets
– Many people will not allow access to their logs

Active measurements are typically discrete
– Important characteristics can be missed
– Probes can be filtered and/or detected
Who is doing Internet
Measurements?

Businesses do a great deal of measurement
– What measurements are they taking and what do they
do with their data?

Instrumentation for measurement-based research
is relatively new
– Developments over the past 12 years have been slow
– 10’s of current studies (see CAIDA and SLAC pages
for lists of these)

Most studies are not coordinated and relatively
narrowly focused
In the past…

Bellcore Ethernet packet traces – Leland et al.
– High time resolution LAN data collected over 4 year
period beginning in 1989
– Thorough analysis showed self-similar properties

NPD study – Paxson
– Characterized routing and packet behavior in wide area
– First installation of large measurement infrastructure
Internet topology: Skitter Project

CAIDA
 Internet topology measurement infrastructures
– Traceroute studies focused on router/link discovery
– Skitter: 17 sources, 54,000 destinations world wide
– Data from Skitter is publically available

www.caida.org
Skitter example – BGP paths
Surveyor

Advanced Network Systems
 Infrastructure for measuring one-way Internet
latency, loss and routes between hosts
– 61 sites world wide
– GPS enabled
– Closed platform (until recently)

www.advanced.org
Surveyor example - delay
Wide Area Web Measurements
(WAWM)

Barford and Crovella (Wisconsin & Boston Univ.)
 Application level measurement infrastructure
– 11 clients (national and international)
– GPS enabled

Combined passive and active measurements
 Application of critical path analysis to TCP
transactions
 www.cs.wisc.edu/~pb
Example of WAWM results
File transfer delay for 500KB file between Denver and Boston
National Internet Measurement
Infrastructure (NIMI)

Paxson (ACIRI), Adams and Mathis (PSC)
 Secure management platform for wide area
measurements
– Designed for general probe installation

Distributed client infrastructure
– Principally academic sites (currently 41 world wide)

A number of projects are running on NIMI
 www.ncne.nlanr.net
Short Term Challenges of
Internet Measurement

Developing methods for gathering more precise
data
– Current tools are frequently quite poor

Developing methods for analyzing measurement
data more thoroughly
– What we know about how things work is limited

Developing methods for understanding causes
and effects across multiple domains
The future – Global Internet
Measurement (GIM)

Ubiquitous measurement capability
– Embedded into the design of the Internet
– Emphasis on extensible API’s for measurement
– Analysis capability must be built in



Measurement quality and soundness
Data formats that enable aggregation that reflects higher level behaviors
Many difficulties
– Management, security, privacy, heterogeneity, deployment, etc.

See Performance Evaluation vol. 864
www.elsevier.com/locate/peva
Conclusion

Measurements are necessary for understanding Internet
structure and behavior
– Rule #1: Expect surprises!

Internet measurements are difficult
– Rule #2: Expect the data to be “dirty”

Measurements are necessary for populating and
validating any reasonable Internet models
– Rule #3: Garbage-in-garbage out

Current measurement infrastructures can provide a great
deal of data but fall short of the GIM goal