Comparison of Some Internet Active E2E Measurement Infrastructures

There are many Internet measurement infrastructures in existence today (January 2004). This pages attempts to provide a simple comparison of some representative public domain infrastructures that perform active end-to-end monitoring.
  PingER AMP Surveyor RIPE TTM BW NDT Script route Planet Lab NIMI PiPES NTAF NWS MonA LISAd
Dedicated hardware No Yes Yes Yes No Yes No Yes No No No No No
Typical link speed Mbps 10/100 10/100 10/100 10/100 1000 All N/A 100 10/100   1000 10/100 10/100
Active Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Passive No No No No No Yes No Yes9 Yes Yesf Yes No No
End-to-end Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Hierarchical Yes Yes Yes Yes Yes Yes Yes No Yes Yes
Full mesh No Yes No No Maybe Yes No Yes Yes No
Production today Yes Yes No Yes Yes Yes Yes No Yes Yes Yes Yes
Measure sites 35 120 10 4 200e 200 35 145 6 40
Responder sites 500 130 40 Any 200 145 40
Regular Yes Yes Yes Yes Yes No No Yes No1 Yes Yes Yes Yes
On demand No Yes     No Yes Yes Yesa Yes Yes No Yes No
Resource limits N/A N/A N/A N/A No Time Yes No Yes No No No
Authentication No No No No ssh Java   sshb Strong AES key No No  
Authorization No No No No Low slices Strong No No No
Initial effort3 V. low8 Need hdw Need GPS Need GPS Acct & pwd V. Low7   Need hdwc   Need GPS Low Low Medium
Sandbox5 No No No No Yes central Yes central Yes scripts Yes Yes scripts No No No Yes central
Net load v. low v. low low low high high short low variable variable Low variable variable variable
Dist Sched4 No No No No No No No No   Yes No Yes No
Public access to data Yes Yes N/A6 No Yes Yes   Yes No Yes Yes Yes Yes
Community HENP Digital Divide NSF HPC Internet 2 Eu ISPs HENP, Grid Global Internet Net res Computer science Net res, eng, debug Internet 2 Net100 Grid Grid
Focus         Analysis Desktop config Testing Tool develop for WAN Infra Partial path analysis eval uation Forecast Drill down, vizual
Tools ping ping 1 way delay 1 way delay ping Web100 ping   zing & ping owamp ping RTT ping
    trace route trace route   trace route crafted tests trace route trace route trace route trace route trace route    
  ABwE sprobe mcast bwctl / iperf path load / rate
        iperf     iperf iperf2   iperf bandwidth  
    FTP Rev Path Tree scriptroute FTP netest
Bad Good


1 Not currently, but possible
2 Also does Treno
3 Estimates how hard it is to add a new host
4 Distributed Scheduling refers to the ability for multiple monitoring hosts to coordinate sending probes so do not interfere
5 Refers to the ease of adding new tests, e.g. can the user do it via a scripting language, can the central admin do itů
6 No data is available from Surveyor at the moment (Jan '04)
7 Refers to clients, only need Java. Servers requires web server access
8. Refers to monitored hosts, monitoring host takes 1-2 days to install/configure etc.
9 Can read various host internal counters such as iocount in PlanetLab host
a e.g.
b You provide ssh public key, they distribute to all hosts you have a slice on
c New site needs hardware & install software and a trained admin, follow on users fill out form and provide ssh key
d MonALISA was initially aimed at monitoring Grid cpu resources
e Scriptroute runs on PlanetLab so we use PlanetLab's numbers
f NetFlow

See Also:

Internet Active End-to-end Measurement Infrastructures Internet Measurement Tool Evaluations Interesting Sites for Network Monitoring
Network Monitoring Tools CAIDA Network Measurement Tool Taxonomy