============================================================ NANOG UPDATE Susan Harris JANOG9 January 25,2002 ============================================================ Thank you for receiving a wonderful warm reception. I'm glad to meet you and to speak with you. Special Thanks to... Ikuo Nakagawa Rie Shimada Toshio Tachibana Takashi Arano Kuniaki Kondo JANOG9 sponsors! ------------------------------------------------------------ NANOG is holding the meeting 3 times for the year like JANOG. The topics which has placed especially the focus are.. * Internet operations (IP and other protocols) * Lage-scale backbone network technologies * Education for ISP and cooperation We'll take up topics about technologies realizable within at least six months. If we'll foresee too much, it will become wishful observation and the nesessity of breaking in till the protocol not being enacted in IETF will come out. ------------------------------------------------------------ NANOG is a non-profit organization and doesn't represent the industry. Whoever want to participate in NANOG is welcome. You know NANOG has a ML. Although it is presumption, there is over 10,000 subscribers of the ML at least. Merit network is the coordinator of NANOG. It is also the organization to which I belong. I have worked as a technical writer there. Merit network is in the campus of University of Michigan at Ann Arbor in Michigan. Merit network is a non-profit organization too, and having produced Radius is also known. In 1966, Merit was founded in order to develop a system which connect the main computer of three universities (University of Michigan, Michigan State University, and Wayne university mutually). Now Merit connects Michigan's 4-year public universities to the Internet. Also Merit operates MichNet which is Michigan's statewide backbone network. Now Merit is concerned with a management of Michigan's 4-year public universities. Also Merit operates Michinet which is a local area network. ------------------------------------------------------------ Many Japanese people participate in NANOG. Mr. Mizukoshi who is one of the persons asked me "Why was it possible to have made the organization that gather such all people, operator, researcher, vendor and so? " When I answer it simply, this is a accident of History! ------------------------------------------------------------ From 1987 to 1995, Merit managed NSF's network. NSFNET was the high-speed network built for the first time in the United States. NSFNET is nonprofit. The main members were people concerned with education. NSFNET was tree-tiered architecture, 13 regional networks was hanging down from NSFNET, and several 100 campus networks was hanging down from the bottom of it further. [NSFNET Regional Networks] *BARRNet *SDSC *JVNCnet *SESQUINET *Merit *SURAnet *MIDnet *Westnet *NCAR/USAN *PSC *NCSA *NWNet *NYSERNet For the NSFNET backbone service , merit staff met quarterly with the representative of regional network operator. This meeting developed into NANOG. ------------------------------------------------------------ NANOG had also been exposed to the wave of a big change. It had overlapped also with the time of a big change of the Internet. It was such a time NSFNET(non-profit) was replaced with commercial ISPs. *a new NSFNET architecture by mid '90's *NSFNET backbone to be replaced with many commercial backbones *commercial backbones connected at Network Access Points(NAPs). These things were NSFNETS's goals from the beginning. In the 90s, the people of commercial ISP also came to participate in the meeting. And regional-Techs meeting grow, include borader base of vendors, operators, R&E. That is why we re-charters as NANOG in 1994 to reflect broader role. ------------------------------------------------------------ In 1994-1995, regional operators changed their backbone from NSFNET to commerciall. And NSFNET decommissioned in 1995, these backbone was absorb by MCI and Sprint. NANOG was funded by National Science Foundation also for a while. Now funded by registration fees, host/supporting organizations, vendor contributions. ------------------------------------------------------------ Then, I think you have a question about NANOG Meeting , it was not able to be asked even if you want to do. I want to receive such the question from you. *What is the audience like? -like Janog's audience, who would like to discuss very openly. *What is the discussion, and utterance jump out? -There is a person who said, "That was the worst talk I've ever heard!" -There is a person who said to a FBI, "Are you calling me a liar?" ------------------------------------------------------------ [graph : attendee occupation 10/01] => ISP(43%), Vendor(27%), R&E(7%).... [graph : Meeting attendance] You can grasp that the attendance increase favorably from this graph. Last time, the meeting held about one month after September 11, when that Terrorism occurred. Therefore, many people canceled an attendance. Also Japanease people canceld. It seems that there was an unavoidable situation that an overseas trip was forbidden by their company. The meeting safely held in California. Though there were many cancellations, we got many attendee from local people. [List : Countries Represented Oct.2001] In this time, although there were 13 attendee from Japan, there was very few attendee than usual. There are always 30 attendee from Japan. Japan is always the country of the maximum attendee except United States. ------------------------------------------------------------ With help from the University of Oregon, poeple who couldn't come actually are able to participate the meeting by multicast system at remote. NANOG of last year, there were 20-30 concurrent viewers and 50-100 unique viewers. There were also many RealMedia users up to 146 concurrent viewers. In many cases, multicast has a problem establishing a native link. We were faced with various technical problems, so the infrastructure of multicast is often put at Internet2 GigaPOPs. And we provide IPv6 connectivity via a tunnel back to Merit in Michigan. ------------------------------------------------------------ Although we encourage strongly our attendee to use the Squid cache at the meeting hall, there are few users as yet. *Meeting Schedule | Morning | Afternoon | Evening ------+---------+-----------+---------- Sun. | Moving | Tutorials | Tutorial | | | or Host Party ------+---------+-----------+---------- Mon. | General | General | Beer`n | Session | Session | Geer,BOFs ------+---------+-----------+---------- Tue. | General | End mid- | Moving | Session | afternoon | ------+---------+-----------+---------- We'll hold Beer'n Geer party at night. Vendors who became sponsor will provide free beers for attendee. In return, we'll permit vendors to exhibit their products at the party. But a purpose of the party is not commercial, therefore we limit the number of company to eight. And they can exhibit only their product on a table. *[List : Previous Local Hosts] ------------------------------------------------------------ *Presentations In what form is a presentation proposed? -Many talks generated by Call for Presentations. Of course, we have program commitee. But role of program commitee differs from JANOG. The members not responsible for particular topic area. Main task is review about proposals and slides. We'll hold NANOG soon (February). I'd like to talk about TOPICs that is taken up at the meeting. ------------------------------------------------------------ *DNS Damage - Measurements at a Root Server -Evi Nemeth, CAIDA She measure on F root server. *Malformed A queries were 14% of the load *20% of queries asked for non-existent TLDs. ->*Includes lots of internal company name *Private address space sneaks out as source addresses and query targets. *Denial of service attacks ->Denial of service attacks often use the DNS as reflectors But, Performance of the root servers amazing given the bogus query load. ------------------------------------------------------------ *New Developments in Peering for Tier- 2 and Content Providers -Jeb Linton, EarthLink He asserts that there are 2 major changes in peering industry. *Seven Tier-1 provider plan to use common colocation space for Next- Generation" peering at OC- 48 and higher speeds. *Prices for transit services have gone down. =>Therefore, traditional peering methods, such as legacy NAPs and private line peering, are no longer less expensive than transit. And furthermore we set pannel discussion. NAP IXP affairs is going to be updated (LINX@London) *Topic of International IPv6's IX is included. ------------------------------------------------------------ *Analysis of IS-IS Behavior -Cengiz Alaettinoglu & Steve Casner, Packet Design *Analysis of IS-IS packet traces *About IGP convergence problems *Recipe for achieving sub-second IGP convergence ------------------------------------------------------------ *Problematic inter-domain routing issues. -Olaf Maennel and Anja Feldmann *They made new public domain tool, character. *They used MRT tools too. =>for analysis of flap dampening problems. ------------------------------------------------------------ *Native Multicast service on GIANT. -Agnes Pouele, DANTE & Jan Novak, Cisco This presentation will publish about Native multicast service on GIANT. GIANT is a new European network of 10GbE. That target is 28 countries in Europe. GIANT will carry native multicast traffic by February 2002. ------------------------------------------------------------ *Global Crossing's operational experience with MPLS He'll talk about following topics. *MPLS for IPv6 *MPLS/VPN ------------------------------------------------------------ All these abstracts are on the web. http://www.nanog.org/mtg-0202/topics.html Slides aren't available before the meeting. ------------------------------------------------------------ Future Presentations *PANEL: Intelligent route controle technology(NEXT) =>About optimal route selection by BGP. *used specialized boxes. it provide optimized egress routes to multihomed enterprises But Vendor not yet ready to discuss in next meeting. I think concrete topic will come out in the next meeting, May or June. ------------------------------------------------------------ *Status of four- byte AS number implementations *The drain problem of AS number *Registry plans for dealing with the problem *Examples of implementation. ------------------------------------------------------------ *Native multicast peering at exchange points =>There are pros and cons. I think we'll discuss this topic in the meeting, May or June. ------------------------------------------------------------ First back-to-back meetings, fall this year. This goal is that we provide additional operator input to ARIN. Both organizations will offer Sunday tutorials. And NANOG will meet from Sunday to Tuesday as usual. ARIN will meeting from Wednesday to Friday. Furthermore, in order for NANOG and ARIN to promote the re- lation, ARIN IP analyst is available at help disk throughout for offering informations. That's all. ============================================================ - Q&A - Q(Sekiya) : Each time, you explain hopeful topics for the next meeting at the last of a meeting. Who and how did decide these? A : The program commitee decide it. Number of the commitee are 12 or 13. I am a chair of the program comitee. All of members are specialist. You can know their name by following site. http://www.nanog.org/general.html The highest target of program commitee is that we obtain various presentation from a broad field. These presentators include researcher, vendor, operator and so on. I know your going answer. "Would be accept suggestions or top- ics?" I know your going answer. Q(Sekiya): It's good. But, I'd like to ask you an another ques- tion. How do you choose program commitee? A: Those who are known very well, the specialist in a true meaning are chosen. ---- Q(Maemura) : Please teach me that you feel the role of NANOG have changed through your experience of NANOG for seven years. A: We try to be little different from original role as possi- ble. But all difference is, we are much bigger from 100 to 700. There is much more commercial interest now. Recently we have approach directly from marketing departments. I think they know NANOG from web and they consider NANOG good advertisement. Therefore we have to explain, "Not as represented company, Do you have operators and so on can present abstract telling us and he or she want to talk about?" Also more vendor interest in Beer'n Geer, we could be much more larger. But we restrict to eight companies. And we try to preserve spirit of the NSFNET. We openly ex- change information and protect from company's interest. ---- Q(homine?) : If Community is continued for along time, young person will not enter and it will be faced with an aging prob- lem in many cases, How about it in NANOG? A: There are so youngs like you in NANOG. I'm exception:-) ---- Q(Tachibana) : Could you make some comment about mental at- titude when selecting talks for NANOG? A : It is difficult one. Very careful screening about talks is need to make sure from not their product's gain and to make sure speaking from own experience. ---- Q(sekiya):Although it is a vague question, what is the biggest problem when participants increase? A : From view point of staff, registration ,management of the people, finding space and finding hotel that have a big hole room. We don't plan for increasing attendee without any limited. We think that is about 800 maximum as ever. ------------------------------------------------------------