1

Measuring the Digital Divide with PingER

Prepared by Les Cottrell, SLAC, for the Round Table: Developing Countries Access to Scientific

Knowledge, October 23-24, 2003, ICTP Trieste, Italy

www.slac.stanford.edu/grp/scs/net/talk03/ictp-oct03.ppt

Partially funded by DOE/MICS Field Work Proposal on Internet End-to-end Performance Monitoring (IEPM), also

supported by IUPAP

2

Outline• How PingER makes the measurements

• Typical usage & benefits

• Set the stage for what values mean

• Application to Digital Divide

• How’s the world doing today, where is it going?

• Comparisons with development indices

• Summary

3

Methodology

• Use ubiquitous ping

• Each 30 minutes from monitoring site to target : – 1 ping to prime caches– by default send10x100Byte pkts– 10x1000Byte pkts

• Record loss & RTT, (+ reorders, duplicates)

• Derive throughput, jitter, unreachability …

4

Architecture

• Hierarchical vs. full mesh

WWWWWW

ArchiveArchive

MonitoringMonitoringMonitoringMonitoring MonitoringMonitoring

RemoteRemote

RemoteRemoteRemoteRemote

RemoteRemote

FNAL

Reports & Data

CacheMonitoringMonitoring

SLAC Ping

HTTP

ArchiveArchive

1 monitor hostremote host pair

5

Countries Monitored

6

Recent additions• Added remote hosts in Albania, Macedonia,

Serbia/Montenegro, Belarus, Turkey, Armenia, Mexico, Cuba, Azerbaijan, Tajikistan, Turkmenistan, Kyrgyzstan, Philippines & Namibia

• Contacts– Working with contacts in Vietnam, and Tunisia – Working with Iran site to set up monitor host

• Increased hosts monitored from CERN to give better European view – Now monitoring 60 countries from CERN– ~80 from SLAC

7

Countries Monitored

Used to monitorOnly 1 host

Need > 1 host to reduce anomalies

Country

Hosts Country

Hosts Country

Hosts Country Hosts

Albania 1 Estonia 1 Latvia 1 Slovakia 2

Argentina 6 Finland 1 Lithuania 1 Slovenia 1

Armenia 2 France 11 Macedonia 2 S Africa 3

Australia 4 Georgia 2 Malaysia 3 Spain 6

Austria 2 Germany 13 Mexico 5 Sweden 4

Azerbaijan 2 Ghana 1 Moldova 2 Switzerland 8

Bangladesh 1 Greece 1 Mongolia 1 Taiwan 1

Belarus 2 Guatemala 2 Netherlands 12 Tajikstan 1

Belgium 3 Hungary 5 New-Zealand 4 Thailand 0

Brazil 21 Iceland 3 Nigeria 1 Turkey 2

Bulgaria 1 India 10 Norway 2 Turkmenistan 1

Canada 11 Indonesia 0 Pakistan 1 Uganda 1

Chile 4 Iran 4 Peru 1 Ukraine 2

China 6 Ireland 2 Philippines 1 UK 36

Colombia 4 Israel 5 Poland 4 US 208

Costa-Rica 1 Italy 13 Portugal 2 Uruguay 3

Croatia 5 Japan 11 Romania 1 Uzbekistan 2

Cuba 2 Jordan 1 Russia 12 Venezuela 2

Czech-Rep 3 Kazakhstan 2 Saudi Arabia 2 Vietnam 0

Denmark 1 Korea 2Serbia & Montenegro 2  

Egypt 1 Kyrghzstan 1 Singapore 1  

~ 80 countries

480 sites

800 hosts

3600 pairs

8

PingER Benefits

• Provides quantitative historical (> 8yrs) and near real-time information– Aggregate by regions, affiliations etc.– How bad is performance to various

regions, rank countries?– Trends: who is catching up, falling behind,

is progress being made?– Compare vs. economic, development

indicators etc.• Use for trouble shooting setting

expectations, identify needed upgrades, choosing a provider, presenting to policy makers, funding bodies

Monitoring site vs. Remote sites screen shot

• Aimed at: end-user (net-admin & sophisticated user), planners • Measures analyzes & reports round-trip times, losses, availability,

throughput ...– Uses ubiquitous ping, no special host, or software to install/configure at remote

sites– Low impact on network < 100bits/s, important for many DD sites– Covers ~80 countries (99% of Internet connected population)

9

What do values mean?• Loss – most important single metric

– < 1% good, (throughput ~ 1/sqrt(loss)) – > 1% voice over IP slightly annoying– >4-6%

• non-native language speakers unable to communicate, • video conferencing becomes irritating, • hard to edit files remotely (characters on screen lag keyboarding) • Significant degradation of TCP performance

– >10-12% TCP sessions time out, FTP fails, mail still works (keeps retrying)

• RTT (Round Trip Time)– Remote instrument control needs low RTT– >500ms significant interactive voice problems, typing/echo

problems

10

Throughput - meaning• Derived throughput ~ 1460 Bytes /

(RTT * sqrt(loss))• 56 - 64 kbps home modem limit,

Internet backbone 1980• 128 kbps ISDN• Home DSL / cable modems low

limits 200-500kbps• Internet backbone late 80’s 1.54

Mbps• Internet backbone early 90’s

45Mbps• Internet backbone today 10Gbits,

big sites 155Mbps – 1Gbps• Testbeds/cutting edge end to end

TCP throughput (Land Speed Record, Guinness Records Book)– 2001 1Gbps– 2003 2.4Gbps/s (February) =>

5.6Gbps (October)

• Example of the increases in capacity of a well connected site in US

11

Usage Examples

Median Packet Loss Seen From nbi.dk

0

5

10

15

20

25

30

35

40

45

50

11/1

/98

11/8

/98

11/1

5/98

11/2

2/98

11/2

9/98

12/6

/98

12/1

3/98

12/2

0/98

12/2

7/98

1/3/

99

1/10

/99

1/17

/99

1/24

/99

% 1

00 B

yte

Pac

ket

Lo

ss D

uri

ng

Day

.

Ten-155 became Ten-155 became operational on operational on December 11.December 11.

Smurf Filtersmurf Filtersinstalled oninstalled onNORDUnet’sNORDUnet’sUS connection.US connection.

To North America

To Western Europe

Packet Loss between DESY and FNAL in February and March 2000.

0

2

4

6

8

10

12

1 3 5 7 9 11 13 15 17 19 21 23 25 27 1 3 5 7 9 11 13 15 17 19

Day of the Month

Da

ily

Pa

ck

et

Lo

ss

(%

)

DFN closes Perryman POP and looses direct peering with ESnet

Peering re-established via Dante at 60 Hudson

February March

Peering problems

Upgrades & ping filtering

12

Usage Examples

• Selecting ISPs for DSL/Cable services for home users– Monitor accessibility of routers etc. from site– Long term and changes

• Trouble shooting– Identifying problem reported is probably network related– Identify when it started and if still happening or fixed– Look for patterns:

• Step functions• Periodic behavior, e.g. due to congestion• Multiple sites with simultaneous problems, e.g. common problem link/router …

– Provide quantitative information to ISPs

Identify need to upgrade and effects

• BW increase by factor 300• Multiple sites track• Xmas & summer holiday

13

Rate LimitingRTT Loss

2 hosts at same site see sudden step-like increase in loss from < 1% to 20-30% at similar time

boromir.nask.waw.pl

gollum.nask.pl

www.pol34.pl Another host in Poland sees no problems, i.e. helps to have another nearby host

RT

T Los

s

Similar effects for Greek (uoa.gr), Bulgarian (acad.bg), Kazakhstan (president.kz), Moldovan (asm.md) and Turkish (metud.edu.tr) sites

If no step function or nearby host may not notice, so also compare synack vs ping

Can ping routers along path to see where onset occurs

At any given time, about 5% of monitored hosts are doing this, most in developing countries. Recently (August 2003) seen an increase in ping rate limiting

RTT Loss

boromir.nask.waw.pl

gollum.nask.waw.pl

www.pol34.pl

RT

T Los

s

14

Digital Divide Regions• Design regions

– to match well known world regions and – to have similar connectivity within region

• Then order by derived throughput– Derived throughput ~ MSS [=1460Bytes] / (RTT*sqrt(loss))

• Want to show general behavior & variability (outliers)• Developed:

– U. S.+Canada, Japan+Taiwan+Singapore+Korea, Australia+NZ, Europe (excl. SE Europe, Russia)

• Developing (Digital Divide):– Africa, S. America, C. America, C. Asia, China, S. Asia, Caucasus, M.

East, SE Europe, Russia

Israel has much better connectivity than neighbors in Mid East so distorts Mid East results, move to Europe?!Greece is part of Europe, should it be part of S. E. Europe, choice varies with time…

15

Region Map

• Also have affinity groups, e.g. AMPATH, Silk Road, CMS, XIWT and can select multiple groups

16

Current State – Aug ‘03 (throughput)

• Within region performance better– E.g. Ca|EDU|GOV-NA, Hu-SE Eu, Eu-Eu, Jp-E Asia, Au-Au, Ru-Ru|

Baltics• Africa, Caucasus, Central & S. Asia all bad

Bad < 200kbits/s < DSL Poor > 200 < 500kbits/s

Acceptable > 500kbits/s, < 1000kbits/s

Good > 1000kbits/s

17

Region Countries #

AfricaGhana, Namibia, Nigeria, Namibia, S. Africa, Uganda 6

C AsiaKazakhstan, Kyrghzstan, Mongolia, Tajikistan, Turkmenistan, Uzbekistan 9

S AsiaBangladesh, India, Indonesia, Malaysia, Pakistan, Thailand, (Vietnam) 16

M East Egypt, Iran, Israel, Jordan, Saudi Arabia, Turkey 10

Caucasus Armenia, Azerbaijan, Georgia 5

S AmericaArgentina, Brazil, Columbia, Chile, Peru, Uruguay, Venezuela 13

China China including Hong Kong 5

Russia Russia 5

C America Costa Rica, Cuba, Guatemala, Mexico 4

SE Europe(Albania), Bulgaria, Croatia, Macedonia, Moldova, Romania, Serbia/Montenegro, Slovenia 13

18

Trends

Africa shown for onlyUganda seen from SLAC,since adding new countries with very different throughputs distorts result

S.E. Europe, Russia: catching upLatin Am., Mid East, China: keeping upIndia, Africa: falling behind

Derived throughput~MSS/(RTT*sqrt(loss))

19

Russia

• E.g. Upgrade to KEK-BINP link from 128kbps to 512kbps, May ’02: improved from few % loss to ~0.1% loss

• Russian losses improved by factor 5 in last 2 years, due to multiple upgrades

20

Loss Comparisons with Development (UNDP)

Positive correlation with Human Development or GDP

21

Derived throughput~MSS [=1460Bytes] / (RTT*sqrt(loss))

Europe

Netherlands

Turkey

Belgium

23

Network Readiness

• NRI from Center for International Development, Harvard U. http://www.cid.harvard.edu/cr/pdf/gitrr2002_ch02.pdf

• Using derived throughput ~ MSS[=1460B] / (RTT * sqrt(loss))– Fit to exponential is better

Internet for all focusA

&R

focus

NRI Top 14Finland 5.92US 5.79Singapore 5.74Sweden 5.58Iceland 5.51Canada 5.44UK 5.35Denmark 5.33Taiwan 5.31Germany 5.29Netherlands 5.28Israel 5.22Switzerland 5.18Korea 5.10

24

Challenges• Effort:

– Usually negligible for remote hosts– Monitoring host: < 1 day to install and configure, occasional updates to

remote host tables and problem response – Archive host: 20% FTE, code stable, could do with upgrade, contact

monitoring sites whose data is inaccessible– Analysis: your decision, usually for long term details download & use

Excel– Trouble-shooting:

• usually re-active, user reports, then look at PingER data• have played with automating alerts, data will/is available via web services

• Ping blocking– Complete block easy to ID, then contact site to try and by-pass, can be

frustrating– Partial blocks trickier, compare with synack or TCP/ping

• Derived throughputs poor for well connected sites (<0.1% loss)• Funding to sustain efforts

25

• 35+ monitoring sites in 15 countries– Plan to add ICTP Trieste if funded – Other projects used toolkit, e.g. XIWT, PPCNG/EDG …

• SLAC with help from FNAL• Digital Divide collaboration (MOU) with ICTP, Trieste

– eJDS– looking for a EU grant for eJDS and PingER

• Need funding for coming year:– Working with DoE, NSF, Pew Charitable Foundation …– Tasks:

• (0.5 FTE) ongoing maintain data collection, explain needs, reopen connections, open firewall blocks, find replacement hosts, make limited special analyses, prepare & make presentations, respond to questions

• (+ 0.5 FTE) extend the code for new environment (more countries, more data collections), fix known non-critical bugs, improve visualization, automate some of reports generated by hand today, find new country site contacts, add route histories and visualization, automate alarms, detect rate limiting earlier, update web site for better navigation, add more DD monitoring sites/countries, improve code portability, understand regions better

• Also looking for small grants for helpers in developing countries• ICFA: show importance to policy makers, funding agencies, identify

sympathetic contacts at agencies, get support

Collaborations & Funding

26

Futures• More work on understanding regions

• Better/quicker detection of rate limiting

• Extend deployment, in particular Africa:

• Demonstration in coordination with WSIS in Geneva Dec 2003

27

Summary• Valuable light-weight tool for end-to-end performance• Good for trouble-shooting, planning, setting

expectations• World wide coverage• Performance from U.S. is improving all over• Performance to developed countries are orders of

magnitude better than to developing countries• Poorer regions 5-10 years behind• Poorest regions Africa, Caucasus, Central & S. Asia• Some regions are:

– catching up (SE Europe, Russia), – keeping up (Latin America, Mid East, China), – falling further behind (e.g. India, Africa)

28

More Information• PingER:

– www-iepm.slac.stanford.edu/pinger/

• MonaLisa– monalisa.cacr.caltech.edu/

• GGF/NMWG– www-didc.lbl.gov/NMWG/

• ICFA/SCIC Network Monitoring report, Jan03– www.slac.stanford.edu/xorg/icfa/icfa-net-paper-dec02

• Monitoring the Digital Divide, CHEP03 paper– arxiv.org/ftp/physics/papers/0305/0305016.pdf

• Human Development Index– www.undp.org/hdr2003/pdf/hdr03_backmatter_2.pdf

• Network Readiness Index– www.weforum.org/site/homepublic.nsf/Content/Initiatives+subhome

29

Africa – Getting Started• Recommendation on Monitoring from Open Round Table,

Trieste 2002– Devote resources to monitor in real time the connectivity of research &

educational institutions in developing countries

• MOU signed between SLAC & eJDS/ICTP – Use & extend SLAC PingER/project for monitoring eJDS “network” of

participants

• Find contacts at potential sites– Dec ’02 send emails to eJDS participants, ICFA/SCIC representatives

• Explain value to participants, needs for hosts to monitor, contacts

– Contact contacts, explain again

• When get host, check it is pingable• Could be multi-month process• ~75% contacts provided successful hosts

– 25% pings blocked and unable to resolve, emails exchange petered out

• Still need contacts for many countries

30

Africa Oct ‘03• Hosts in: Ife-Ife/Nigeria, Accra/Ghana,

Kampala/Uganda, Windhoek/Namibia, UCT/ZA, Johannesburg/ZA, Mosselbay/ZA

• Carriers:– GH uses UUNET/Satworks, NA uses

UUNET/xantic, NG uses TELIANET/NewSkies, UG uses Level(3)/globalconnex

– ZA varies from site to site: UUNET/ALTERNET, C&W Telecom S. Africa, CAIS telcom S. Africa

• UG, NA, NG, GH use satellites (> 600ms)• ZA uses landlines

31

Africa RTT• Monitored from N.

America & Europe– Depends on remote site

(not monitoring site)– Satellite for all except S.

Africa– Ghana problems

32

West Africa Loss• Ghana very poor performance

– Sudden increase in losses on August 18th– Not rate limiting according to synack– Sometimes get down to a few %– Route ESnet-UUNET/ALTER.NET– Losses appear on

last 2 hops in Ghana

• Nigeria better– Route via

TELIANET/newskies

33

Africa Derived Throughput• S. Africa (UCT) best, followed by Uganda, Nigeria and Ghana• Throughput to Nigeria site == home DSL/cable• Throughput to Ghana site == modem dialup• Uganda site == SLAC late 1980s

Derived throughput ~MSS[=1460Bytes]RTT * sqrt(loss)

34

Summary 1/2• Beware limited number of countries monitored and

even then limited number of sites in each country. Need more contacts: – Countries with no sites: e.g. Algeria, Cameroon, Kenya,

Malawi, Mozambique, Senegal, Tanzania, Tunisia, Zambia– Countries with only one (possibly anomalous) site: Uganda,

Ghana, Nigeria, Namibia• Africa poorest performance region

– Factor 30 behind Europe in performance today– ZA and Uganda better performers– “Best” Uganda site, 8 years behind Europe– Ghana bad, Nigeria poor, similar to connectivity to homes in

Developed nations• Overall Africa not catching up• Ghana, Nigeria falling behind

35

Summary 2/2• Little uniformity in routes, many carriers

– Unlike S. America & Caucasus where AMPATH & Virtual Silk Highway have improved performance

• Hopefully Africa ONE project will help– Undersea fiber to link countries of Africa together

and to one another

36

Extra Slides

37

VisualizationKeep it simple, enable user to do their

own by making data available • Tables

– Time series (www-iepm.slac.stanford.edu/cgi-

wrap/pingtable.pl): • select metric (loss, RTT etc.), time ticks,

packet size, aggregations from/to, etc.

• Color code numbers, provide sort, drill down to graphs, download data (TSV), statistical summaries

– Monitoring site vs. Remote sites (www-

iepm.slac.stanford.edu/cgi-wrap/table.pl):• Select metric, region aggregations

• Drill down to time series, download data

• Graphs– Select source(s)/destination(s), metric,

time window, SQL selects, graph type

38

Publish information

#!/usr/bin/perl use SOAP::Lite; my $characteristic = SOAP::Lite -> service(‘http://www-iepm.slac.stanford.edu/tools/soap/wsdl/profile_06.wsdl') -> pathDelayOneWay("tt81.ripe.net:tt28.ripe.net”); print $characteristic->{NetworkTestTool}->{toolName},"\n"; print $characteristic->{NetworkPathDelayStatistics}->{value},"\n";

• www.slac.stanford.edu/cgi-wrap/pingtable.pl => tabular reports, also download data

• Data accessible from MonaLisa• Implementing web services

access prototype– Includes: PingER, IEPM-BE, RIPE-tt,

I2 E2Epi OWAMP– Use GGF/NMWG schema/profile,

e.g.• path.delay.roundTrip

39

Rate Limiting Moldova

cni.md

lises.asm.md

RTT Loss

Moldova Bulgaria

40

41

Loss Comparisons with Development (UNDP)

Weaker with education & literacy

Positive correlation with Human Development or GDP

42

Human Development Index (HDI ) RankSource: UN

0

20

40

60

80

100

120

Uni

ted

Sta

tes

Cze

chN

ethe

rland

sF

inla

ndH

unga

ryA

ustr

iaS

pain

Sw

itzer

land

Slo

vaki

aD

enm

ark

Italy

Cro

atia

Irel

and

Pol

and

Icel

and

Gre

ece

Bel

gium

Tur

key

Slo

veni

aIr

an, I

slam

icLi

thua

nia

Est

onia

Rom

ania

Geo

rgia

Ukr

aine

Mol

dova

,A

lban

ia

HD

I

GDP per capitaSource: UN

0

5000

10000

15000

20000

25000

30000

35000

40000

Un

ited

Sta

tes

Cze

ch

Ne

the

rla

nd

s

Fin

lan

d

Hu

ng

ary

Au

stri

a

Sp

ain

Sw

itze

rla

nd

Slo

vaki

a

De

nm

ark

Ita

ly

Cro

atia

Ire

lan

d

Po

lan

dIc

ela

nd

Gre

ece

Be

lgiu

m

Tu

rke

y

Slo

ven

ia

Ira

n,

Isla

mic

Lith

ua

nia

Est

on

ia

Ro

ma

nia

Ge

org

ia

Ukr

ain

e

Mo

ldo

va,

Alb

an

ia

GD

P p

er C

apit

a (P

PP

US

$)

NREN Core Network Size (Mbps-km)

10M

1M

100K

10K

1K

100

2000

2001

Leading

Advanced

In transition

Source: From slide prepared by Harvey Newman, presented by David Source: From slide prepared by Harvey Newman, presented by David Williams ICFA/SCIC talk on Serenate report. Data from the TERENA Williams ICFA/SCIC talk on Serenate report. Data from the TERENA CompendiumCompendium

Lagging

Derived throughput~MSS/(RTT*sqrt(loss))

Europe

Netherlands

Turkey

Belgium