CAIDA-WIDE Workshop: Minutes
This page contains the minutes of CAIDA/WIDE Workshop on March 21, 2003.
Presentations
- Brad Huffaker (CAIDA):
- skitter project: forward ip toplogy mapping and performance
- some root name servers allow monitoring of incoming queries
- SAINT paper on distance metrics
- PAM paper on DNS root servers placement
- Nevil Brownlee (CAIDA):
- passive measurements with NeTraMet
- round trip times (RTT) to root and gtld servers, 3 years of data
- performance plots on the web
- distributed.net montoring (Chris Yarnell, ISC)
- Andre Broido (CAIDA):
- rfc1918 spectroscopy
- RTT properties
- Duane Wessels (CAIDA/Measurement Factory)
- simulating BIND behavior in controlled laboratory environment
- Jun Murai (Keio University):
- measurement groups
- development of DNS (with COMNET and ISC): v6, client-side, security
- WIDE Internet: transpacific 10GB link to Seattle, 2 other links to US, and APAN (commodity)
- agenda:
- SRI - use DNS structrure to store geographical infromation
new namespace, RFId spaces, (.geo )
https://www.researchgate.net/publication/224693404_RFInD_An_RFID-based_system_to_manage_virtual_spaces
http://www.icann.org/tlds/report/report-iiib3-09nov00.htm
- lack of study on DNS performance impact
- how to think about analysis of Internet traffic in general (new challenge of testbed)
- mobility testbed. 2000 automobiles using mobile IPv6 - need new measurement methodologies for that
- large multicast testbed using satellite (UDLR schemes) over Asia
- concern about using skitter in future DNS root
measurements
* roots doing anycast
* lack of response to ICMP/UDP
- SRI - use DNS structrure to store geographical infromation
- Yoichi Shinoda (Japan Advanced Institite of Science and
Technology):
- interests: simulation and validation
- BED research facility design overview:- destinations - 500 hosts w 800Mhz Pentium PC
- companies use for research and to verify products (e.g., interoperability)
- project to implement IP traceback
- study of multicast performance information
- can be used for more general measurement tasks
- phd student appointed for this type of work
- Suguru Yamaguchi (Nara Institite of Science and
Technology):
- AI3 satellite Internet testbed network for Asia in WIDE
- partnerships with universities in Asia (Thailand) to conduct joint research
- how to extend UDLR infrastructure to the rest of Asia?
- Hiroshi Esaki (University of Tokyo):
- R&D activities and testbed operation
- steering committee member (with Jun Murai) of Japan Gigabit Networ (JGB)
- ipv6 PC testbed with commercial ISPs
- international testbeds
- auto ID - RF tag
- Hideki Sunahara (Nara Institite of Science and
Technology):
- Internet CAR: connect people inside/outside vehicles to the Internet
https://www.semanticscholar.org/paper/IPCar-%3A-Building-the-Probe-Car-System-with-the-Sunahara-Sato/9a9cdd8be1e19bb77c2549c71ad07c4164863a2c (in Japanese)
- Osamu Nakamura (Keio University):
- current status of NSPIXP (management work)
- fast connection to PAIX (Los Angeles) and to San Jose
- most of infrastructure is dual stack
- due to protocol/performance restrictions, domestic testbeds are v4 routed via Cisco and v6 routed via FreeBSD Cisco (no support for OSPF-6)
- 12.2.15.TANGO
- multicast capability and performance on v6 (v4 multicast is almost wire speed, but v6 is not)
- 30Mbps 300 pps multicast is needed for 'school on the Internet'
- Akira Kato (University of Tokyo):
- operational M-root operations: 2-3 Mbps load daily
- M-peer (backup) in Osaka - working with NSPIXP 2 to make distributed server, no formal name yet
- tests of v6 connectivity: NSPIXP6 (layer 2 IX) is a logical XP set over the two exchange points (in Tokyo with GigE and in Osaka with OC12, shared with wide backbone), uses Foundry/Cisco
- Kenjiro Cho (Sony Computer Science Labs. Inc.):
- WIDE traffic repository
- aguri aggregagtoin based traffic profiler
- DNS (server selection algorithm)
- Yuji Sekiya (University of Tokyo):
- performance to root and gTLD nameservers (PAM paper)
- Joao Damas (Internet Software Consortium):
- current F-root architecture diagram
- global servers: Palo Alto, San Francisco (PAIX)
- local servers: San Jose (MAE-WEST), Madrid (ESPANIX), New York
(Telehouse), Hong Kong
- global servers: Palo Alto, San Francisco (PAIX)
- kc claffy (CAIDA):
- overview of research programs and results at CAIDA
- routing analysis (BGP dynamics analysis, atoms, modeling techniques)
- connectivity analysis (as_core and as_rank visualizations, IXP)
- traffic analysis (OC48 traces, interarrival time spectroscopy)
- tool development: passive analysis tools and modules
* netramet (dns performance, flow statistics)
* coralreef (DOS attacks, fragmentation) - security issues (quarantine, modeling)
- bandwidth estimation tools and tests
Discussion
- funding (J. Murai)
- research problems:
- IP -> latitude/longitude mapping by NetGeo
- routine AS112 log collection from the M-root
- joint CAIDA/WIDE report on AS112 analysis (A. Broido, A. Kato)
- expand analysis into SOA and RFC1918 areas (A. Broido)
- how to instrument anycast root servers for measurements? How to define the location of the anycast root server?
- study the load distribution by anycast (B. Huffaker)
- triangulation models for skitter data (B. Huffaker)
- identify 10 ccTLD servers to monitor (N. Brownlee, K. Cho, Y. Sekiya)
- at PAM meeting (April, 2003):
- discuss methodology for comparing measurements of nameserver response time (N. Brownlee, B. Huffaker, k claffy, K. Cho)
- discuss "April data collection" and determine the logistical requirements
- Plan another WIDE/CAIDA workshop in 6 months: either before or
after the 58th IETF meeting (November 9-14, 2003, Minneapolis):
- try to organize it in Los Angeles, possibly at ISC (k claffy)
- prepare announcements and agenda for this workshop (M. Fomenkov)