RIPE 39

Notes TT-Traffic WG May 1 14:00
Chair Keith Godber.
Scribe Olaf Kolkman.
______________________________________________________________________


15 attendees.
Around 8 non RIPE NCC participant.

A. Agenda

No modifications

-------------------- --------------------

B. Henk Uiterwaal,
TTM status and plans.
See slides on the web.

Staff update: 3 people out, 3 people in, student for 2001, Total 5.2
FTE and 0.3-0.7 FTE Student

TB operation structure: Since April 1 Sanctroos is lead network
engineer coordinating operations.

PAM 2001 workshop on passive and active measurements.
- more than 120 ppl attended
- 2 papers by the RIPE NCC.
- procedings available www.ripe.net/pam2001 and paper copy.

The paper "Providing Active Measurements as a Regular Service for ISP's"
contains a technical overview of TTM.


Service Fee
- January 2001 possible problems with tax authority delayed service fee.

- March 2001 Tax authorities cleared situation.
Service fee can be charged.

- Operational costs about 360keuro/year, order 100 test boxes.
3000 EU for first box
1500 EU for 2-9 boxes.
Numbers revised per year.

- Charging starts July.

- Clearing house construction. Surpluses are refunded.
Same model as LIR services, but completely independent.

- Existing sites will be asked to sign a service contract.

- Special deal for network research institutes.

Network expansion.
- TTM glossy; lots of feedback, long time to get through.

- major ISPs and research networks are interested.

- Aim for 150-200 test boxes by late 2001

Access to data
- accessing scheme changes from 1 password for all to new
4 class scheme.
Important new scheme: customer access e.g. for SLA verification.
- Beta test untill 15 May.

Technical developments:
- Automated connectivity test.
- automated warning in about 6 hours: 1 e-mail with instructions.
2 days: tt-ops then contact TBhost.

Clocks:
Trimble clocks had problems replaced by trimble.
Clocks and Clockcard : we are not a reseller but contact us when
interested.

Site Survey tool.
Will be made available to TB hosts...

Tech feedback from RIPE 38
1 u high: possible
GbEther: possible
telco power: looked into.

Analysis machine has been upgraded. Goal to have plots available at
9 am.

Data analysis will be presented at RIPE 40. Jitter added to existing
lists.
Still in process.



Chair asked if pricing is reasonable. No response from audience, Henk will
post this to the list.
-------------------- --------------------


C. Rene Wilhelm: Passive calibration of an active measurements.

Two types of measurements, active and passive, combined: inject traffic
and watch it go over the net.

TTM boxes and DAG cards measurements combined.
see http://dag.cs.waikato.ac.nz

testboxes generate traffic that are also traced by the DAG cards.

Measurements of the two method match closely (within order of 250
microsec, 2 orders of magnitudes below delay ). The systematic
results small and understood.

Details between the measurements where explained: separation of
transmit and receive components.


Questions on details of the presentation.
- bandwidth is not known
- Different kinds of packets can be constructed to study effects
of different protocols (e.g. TCP, multicast, etc)

-------------------- --------------------
AOB
Daniel Karrenberg, demo on local testbox access tool

- Daniel started with the Weather forecast for next days.. Wednesday
there will be rain.


The RIPE NCC started to look at a web-interface running on the test-boxes,
providing:
+ looks like a GUI.
+ update your administrational details.
+ real time traffic graphs.
+ life testbox monitoring.
status of the system and status of the GPS.
- Daniel gave a demo of the alpha mainly focused on the charts showing the

status of the GPS clocks and how the graphs can be used for antenna
position
optimizing.
- suggestions for additions are welcomed.



meeting closed at 15:20.