RIPE 59

Test Traffic Working Group
RIPE 59, Lisbon
Wednesday, 7 October 2009, 14:00 - 15:30
Scribe: Rene Wilhelm, RIPE NCC
Jabber: Anand Buddhdev

A. Administrative Matters

Ian Meikle, Co-Chair of the Test Traffic Working Group, welcomed attendees and opened the session. The minutes from RIPE58 were approved.

B. TTM Network Update - Erik Romijn, RIPE NCC

http://www.ripe.net/ripe/meetings/ripe-59/presentations/romijn-ttm-update.pdf

Erik presented the TTM status update which had been prepared by Ruben van Staveren. The measurement network now has 84 active test-boxes. In the past months boxes have been added in Kenya, South Africa, Uruguay, Pakistan, Hong Kong and Iran. No boxes were decomissioned.

C. RIPE NCC NetSense - Franz Schwarzinger & Vasco Asturiano, RIPE NCC

http://www.ripe.net/ripe/meetings/ripe-59/presentations/schwarzinger-netsense-history.pdf
http://www.ripe.net/ripe/meetings/ripe-59/presentations/asturiano-netsense.pdf

NetSense, the new rich user interface to the suite of RIPE NCC Information Services, was unveiled in Tuesday's plenary session. In the tt-wg meeting Franz and Vasco discussed the backgrounds of the design and development process in more detail. Vasco continued with a live demonstration of the NetSense product.

D. Throughput Metrics and Packet Delay in TCP/IP Networks - Timur Sultanov, Samara State Aerospace University (SSAU)

http://www.ripe.net/ripe/meetings/ripe-59/presentations/sultanov-throughput-metrics.pdf

Timur presented the work done at SSAU on measuring available bandwidth and capacity with a TTM box. By measuring delays of packets with changing size, they were able to calculate the two performance indicators. The main problem with the work is the limited number of measurements they could perform; in TTM only one measurement relation could be tested (between tt143 in Samara and tt01 in Amsterdam). He and his colleagues are looking for partners who can help in performing additional measurments.

E. IPv6 Deployment Monitoring - Emile Aben, RIPE NCC presenting on behalf of Rob Smets, TNO.

http://www.ripe.net/ripe/meetings/ripe-59/presentations/smets-ipv6-deployment-monitoring.pdf

Emile presented TNO's work on IPv6 deployement measurements. The basic idea is to include a small bit of javascript on various websites which will make the users' machine try connect back to a test server using IPv4 and IPv6 protocols. The resulting data set will provide insight in endusers' capability to use IPv6.

Daniel Karrenberg commented he likes it very much and would like to participate. However, before doing so he would like to discuss with TNO what kind of data they will publish from these measurements. Daniel is concerned about indivduals' privacy, but would also like to discuss possibilities of sharing the raw data, for other people to do analysis with too. Once he has talked to TNO and the answers and the publication policy are acceptable, Daniel will get the javascript installed on the RIPE NCC websites he is responsible for and will find others to join the program.

Steve Padgett from Google suggested to also check for users who have a broken IPv6 install. Specifically, users who can connect to a site when receiving an A record from the DNS, but cannot connect when receiving both an A and AAAA record. Services are not going to want to be reachable via both v6 and v4 on a single hostname until they can verify users are not going to break.

Maarten Botterman from GNKS consult commented on the data retention policy TNO had in mind. The intention was not to release to anybody outside of the study. They have a reputation to keep and obviously sharing that would be very damaging to the project. They would want to keep further research on the data very limited. However, if it would be possible to do this in collaboration with RIPE NCC, they would certainly go consider it. To get confirmation that the community agrees with TNO sharing the data with RIPE NCC, Maarten asked for a show of hands of people who do not want to see the data shared with anyone outside the project team. One person raised his hand. Maarten said they could talk later.

F. Possible WG Charter Update

http://www.ripe.net/ripe/meetings/ripe-59/presentations/meikle-test-traffic-wg-refocus.pdf

Ian Meikle presented the proposal by the chairs to widen the charter of the test-traffic working group to also include to include topics outside of the RIPE NCC Test Traffic Measurements. Noting the feedback on the tt-wg mailing list was quite positive, he asked the audience if they support this too.

Dave Wilson from HEAnet stated his organisation would like to give their support.

Daniel Karrenberg commented that basically the proposal solidifies the status quo. The working group already has moved in that direction.

Ian Meikle replied the chairs wanted to make sure there were no objections before formalising it in the charter of the test-traffic working group. As a next step he will report on these developments in the Closing Plenary on Friday, to make the broadening of tt-wg more widely known. Discussion on the proposed wording of the charter will continue on the mailing list.