-------------------------------------------------------- CCIRN Measurements Working Group meeting - draft v0.4.5 -------------------------------------------------------- August 22, 1998 Chicago, IL Draft Notes I Participants Guy Almes Internet2 almes@advance.org Shigeki Goto Waseda Univ./APAN goto@goto.info.waseda.ac.jp Keita Fujii Waseda Univ. fuji@goto.info.waseda.ac.jp Satoshi Ono NTT Software Labs. ono@slab.ntt.co.jp Kevin Meynell TERENA K.Meynell@terena.nl Nevi Brownlee The University of n.brownlee@auckland.ac.nz Auckland The University of Waikato Jinhyoun Youn Korea Telecom Lab jyoun@waltz.kotel.co.kr Yasuichi Kitamura Communications kita@crl.go.jp Research Lab Akira Kato WIDE Project kato@wide.ad.jp Matthew Zekauskas Advanced Network matt@advanced.org Services Warren Matthews SLAC warrenm@slac.stanford.edu Martin Horneffer University of Horneffer@uni-koeln.de Cologne Kevin Thompson MCI kthomp@mci.net Murilo Silva Moukiro RNP/BRAZIL murilo@pop-mg.rup.br Alexandre L Grojsgold RNP/BR ALGOLD@CO.RNP.BR Jose Wiz Ribgizo-Filtto zeluiz@nc-rj.rnp.br Yanghee Choi Seoul National Univ. yhchoi@smart.snu.ac.kr Paul Zawda NCSA/STARTAP zawada@ncsa.uiuc.edu Tony McGregor University of tonym@cs.waikato.ac.nz Waikato Greg Miller MCI/vBNS gmiller@mci.net David Martin Fermilab dem@hep.net Marc Blanchet Vaigenie/Canarie Marc.Blanchet@viagenie.qc.ca Andrew Schmidt AMERITECH schmidt@aads.net Steve Shultz NASA NREN/NGI shultz@arc.nasa.gov Jules Aronson NIH aronson@nlm.nih.gov National Laboratory of Medicine Nids Den Otter SURFNET otter@surfnet.nl II Action Items * an ad-hoc meeting will be called during the Internet2 member meeting on Se ptember, 1998. * next meeting will be called during the 43rd IETF meeting at Orlando * the next CCIRN committee meeting will be held at the INET99, San Jose. * the following issues should continue for discussion - software license - privacy/security issues on captured data * the concrete staffs of the WG, a chair, should be selected. III Proceedings (1) the last WG meeting report Shigeki GOTO reported the last CCIRN measurement working group meeting at Ge neva. At the meeting, the following consensus was agreed. - initiation of global collaboration for measurement - the basic measurement parameters Some tools for the measurement are selected such as OC3mon, Surveyor, Waikat o tools and Purgatorio. The major measurement points should be those exchang e points of the research networks. (2) Report: Global Progress Update on Measurement NEVIL BROUNLEE presented about the activity of Waikato. Measuring of uni-dir ectional delay is their target. The tool for this generates sample packets a nd sends their checksums. This tool is running on PC Linux. Currently their target is DS3. He showed some examples of measurements and showed what kind of parameters the tool can monitor. These examples are in http://atm.cs.waikato.ac.nz/atm/delay/ and the contact person is ian@cs.waikato.ac.nz Next one was for the TCP synchronization. The packet over sonet is their res earch target and some chips are built as routers or cards. His next presenta tion was about the NeTraMet. The NeTraMet is a free software and it can moni tor traffic flow. Its target is the bi-directional one and it tries to measu re traffic in real time. He also presented about difference between OCXmon a nd NeTraMet. OCXmon catches all cells but, in case of NeTraMet, with its con figuration file, it can monitor some target flows only. A question for his p resentation was about the software license of NetraMet and he answered he wo uld continue the policy that it should be a free software. A comment is abou t the availability of monitoring flow data with Cisco routers. MATTHEW ZEKAUSKAS presented about the IP performance measurement of surveyor background. The IP performance metrics is the IPPM WG effort of IETF. Its ta rget is the measurement of the one way delay. The surveyors are now set in s ome sites in US and the tool tries to show the difference between the one wa y delay and the round trip delay. The surveyor needs to be synchronized with the GPS clock. The surveyor is not the name of tool on a machine but the inf rastructure of some machines. It needs the central database machine. Each ma chine is a small PC machine with BSDI with a GPS card. He showed some exampl es between some US universities or a NewZealand university. The current stat us of installing surveyors in US is that 28 machines are installed and about 623 paths are monitored now. Because of the facility of UNIX kernel, it cann ot do such real-time monitoring yet. From the following URLs, some informati ons of surveyors are available. http://www.advanced.org/surveyor http://ippm-db.advanced.org/plots At Internet 2, on some sites, some surveyors machines are located. The paths are full meshed things. SHIGEKI GOTO continued to present one way delay on some main backbones in Ja pan after the Geneva meeting. He refers Evi's presentation at Geneva and eva luates those delays. He also presented about AS number statistics in Japan a nd it's very useful data to design some exchange points. KEVIN MEYENELL continued the presentation of TERENA of the Geneva meeting. http://bill.ja.net/ is the site of the JANET traffic accounting and it shows some measurement data. KEVIN THOMPSON presented OC12mon and OCXmons. Both the OC3mon's and the OC12 mon's codes are downloadable. The OCXmon is something like in statistics ext raction mode, and it can extract those user defined kinds of flow. He showed some examples and one of them is about AS pairs table that were shown at ISM A 97. The multicast monitoring is planned, too. GREG MILLER will present his research efforts at the next meeting, I hope. IV Discussion The chair, Shigeki Goto, started about the definition of parameters to be mo nitored first. Jose Wiz Ribgizo-Filtto talked about the fundamental policy o f network monitoring. His thoughts were the monitoring has 2 faces, one is t he network operating oriented one and the other is the research oriented one . STARTAP staffs talked about their experience about monitoring. The chair c oncluded the direction of this WG. At those major exchange points, we should set one or more tools of these 3 tools * ??? * OCXmon * ippm surveyor and the WG should coordinate how to share those monitoring data without fear ing of privacy. Especially, in those major exchange points, the following po ints should be included. * APAN <-> Europe * APAN <-> STARTAP * STARTAP <-> Europe The chair decided to set some contact persons for each areas. * North America: Guy Almes * Asia Pacific: Kitamura/Goto * Europe: Kevin Meynell * Latin America: Jose Wiz Ribgizo-Filtto V Additional Information HANS-WERNER BRAUN could not attend the meeting but he sent us the following URL that shows current status of the NLANAR network analysis activities. http://moat.nlanr.net/Reports/MOAT1stq Yasuichi Kitamura CRL, MPT, Japan kita@crl.go.jp