<NIS.NSF.NET> [IMR] IMR89-06.TXT JUNE 1989 INTERNET MONTHLY REPORTS ------------------------ The purpose of these reports is to communicate to the Internet Research Group the accomplishments, milestones reached, or problems discovered by the participating organizations. This report is for research use only, and is not for public distribution. Each organization is expected to submit a 1/2 page report on the first business day of the month describing the previous month's activities. These reports should be submitted via network mail to Ann Westine (Westine@ISI.EDU) or Karen Roubicek (Roubicek@NNSC.NSF.NET). TABLE OF CONTENTS IAB Task Forces APPLICATIONS - USER INTERFACE . . . . . . . . . . . . . . page 3 AUTONOMOUS NETWORKS . . . . . . . . . . . . . . . . . . . page 3 END-TO-END SERVICES . . . . . . . . . . . . . . . . . . . page 3 INTERNET ARCHITECTURE . . . . . . . . . . . . . . . . . . page 5 INTERNET ENGINEEERING . . . . . . . . . . . . . . . . . . page 5 INTERNET MANAGEMENT . . . . . . . . . . . . . . . . . . . page 8 PRIVACY . . . . . . . . . . . . . . . . . . . . . . . . . page 8 SCIENTIFIC REQUIREMENT . . . . . . . . . . . . . . . . . page 8 DSAB . . . . . . . . . . . . . . . . . . . . . . . . . . page 8 Westine [Page 1] Internet Monthly Report June 1989 Internet Projects BARRNET . . . . . . . . . . . . . . . . . . . . . . . . . page 9 BOLT BERANEK AND NEWMAN, INC., . . . . . . . . . . . . . page 9 CERFNET . . . . . . . . . . . . . . . . . . . . . . . . . page 11 CICNET. . . . . . . . . . . . . . . . . . . . . . . . . . page 13 CORNELL . . . . . . . . . . . . . . . . . . . . . . . . . page 13 ISI . . . . . . . . . . . . . . . . . . . . . . . . . . . page 13 JVNCNET, NORTH EAST RESEARCH REGIONAL NETWORK . . . . . . page 14 LOS NETTOS . . . . . . . . . . . . . . . . . . . . . . . page 15 MERIT/UMNET . . . . . . . . . . . . . . . . . . . . . . . page 15 MIDNET . . . . . . . . . . . . . . . . . . . . . . . . . page 15 MIT . . . . . . . . . . . . . . . . . . . . . . . . . . . page 15 MITRE . . . . . . . . . . . . . . . . . . . . . . . . . . page 15 MRNET. . . . . . . . . . . . . . . . . . . . . . . . . . page 16 NCAR/USAN . . . . . . . . . . . . . . . . . . . . . . . . page 16 NEW ENGLAND ACADEMIC AND RESEARCH NETWORK . . . . . . . . page 16 NORTHWESTNET. . . . . . . . . . . . . . . . . . . . . . . page 16 NSFNET . . . . . . . . . . . . . . . . . . . . . . . . . page 16 NTA . . . . . . . . . . . . . . . . . . . . . . . . . . . page 18 NYSERNET . . . . . . . . . . . . . . . . . . . . . . . . page 18 OARNET . . . . . . . . . . . . . . . . . . . . . . . . . page 19 PITTSBURGH SUPERCOMPUTER CENTER . . . . . . . . . . . . . page 19 SAN DIEGO SUPERCOMPUTER CENTER . . . . . . . . . . . . . page 19 SESQUINET . . . . . . . . . . . . . . . . . . . . . . . . page 20 SRI . . . . . . . . . . . . . . . . . . . . . . . . . . . page 20 SURANET . . . . . . . . . . . . . . . . . . . . . . . . . page 21 TEXAS HIGHER EDUCATION NETWORK . . . . . . . . . . . . . page 21 UCAR/BOLT BERANEK and NEWMAN, INC., NNSC . . . . . . . . page 21 UCL . . . . . . . . . . . . . . . . . . . . . . . . . . . page 21 UDEL. . . . . . . . . . . . . . . . . . . . . . . . . . . page 22 UNIVERSITY OF ILLINOIS AT URBANA-CHAMPAIGN/NCSANET . . . page 23 WESTNET . . . . . . . . . . . . . . . . . . . . . . . . . page 23 Westine [Page 2] Internet Monthly Report June 1989 IAB TASK FORCE REPORTS ---------------------- APPLICATIONS -- USER INTERFACE The task force met June 13-15 at Bellcore in Red Bank and Morristown, NJ. Significant topics of discussion included audio servers, floor control in the context of real-time teleconferencing, connection architecture, and broadband testbeds. Further details will hopefully appear in the next monthly! Keith Lantz (LANTZ@ORC.OLIVETTI.COM) AUTONOMOUS NETWORKS ANTF is planning for a small meeting in July in conjunction with ORWG. Plans are underway to begin experimental implementations of various Policy Routing functions. Our next two-day meeting will be in September or October on the East Coast. Deborah Estrin (Estrin@OBERON.USC.EDU) END-TO-END SERVICES The End-to-End task force met on June 7-8, 1989 at Lawrence Berkeley Laboratory (LBL), hosted by Van Jacobson. Major discussion centered on three topics. Congestion Control Scott Shenker described research on congestion control, and in particular on Fair Queueing, at Xerox PARC. Dave Clark and Van Jacobson each described their work on congestion control, and the group then held a lively discussion of these issues. Gateway queueing algorithms control the interactions among the different source hosts' datagram streams; this indirectly controls congestion through the operation of the end-to-end flow control of the hosts. Shenker presented analysis and simulations demonstrating the performance benefits of Fair Queueing (FQ), resulting from the insulation it provides between sources. Westine [Page 3] Internet Monthly Report June 1989 Jacobson advanced the thesis that a "sufficiently complicated net" will exhibit a phase transition leading to large spatial oscillations of traffic. The following questions emerged from the discussion: 1. Is FQ too fair? (How about large server hosts with multiple connections?) 2. How rapidly is steady state reached, both in theory and practice? For example, if steady state is reached very slowly, theoretical stability results may not be very useful for real systems. 3. For a steady-state traffic pattern, can we devise a flow control algorithm that will prevent long-lived oscillations? 4. Can we demonstrate the phase transition phenomenon, by analysis and/or simulation? 5. Do the gateways need to estimate time constants and actively anticipate congestion, in order to damp oscillations? VMTP Dave Cheriton reported on some further thinking he has been doing on VMTP. He outlined some changes that would modify the protocol without changing the external model (user interface). The proposed changes are designed to improve efficiency (lower overhead and delay) for larger transfers, and embody a partitioning of functions between those that are performed seldom and those required for actual data streaming. VMTP now supports the socket routine "select", to provide an asynchronous interface. IP Multicasting Steve Deering reported on progress of IP multicasting work. * An IP multicast Mazewar program will be out very soon. * IP multicasting is being installed into NTP at Maryland, and into the ISIS system at Cornell. Sun RPC has been modified to use multicasting. Westine [Page 4] Internet Monthly Report June 1989 * There is serious interest within the OSPFIGP WG of IETF on incorporating multicast routing. This appears to be a natural extension, although it is too large an addition to be included in the initial OSPFIGP version. One problem with using IP multicasting at present is the general scarcity of multicast-capable interface hardware. We hope that the vendors will soon recognize this need and fill it. The group began discussion of plans for performing a major demonstration of multicasting, perhaps using the upcoming Defense Research Internet lines for a national testbed. Suggested applications include an interactive game (e.g., Mazewar), teleconferencing, and the multicast distribution of data (e.g., netnews). Finally, Scott Shenker of Xerox PARC and Lixia Zhang of MIT have joined the task force. Bob Braden (Braden@ISI.EDU) INTERNET ARCHITECTURE The Workshop on the Future of the Internet System Architecture and TCP/IP Protocols was held 1-2 June 1989 at the University of Delaware. About 77 souls attended the workshop, which included a day of individual presentations and a day of panel discussions. The proceedings are now being edited for early publication in a special issue of the ACM Computer Communications Review. Dave Mills (Mills@HUEY.UDEL.EDU) INTERNET ENGINEERING 1. Two drafts documents are now being circulated by the Point- Point Protocol Working Group. One is a white paper on requirements. The second is the preliminary protocol specification. See 3. below for information on obtaining copies. 2. The IETF will meet at Stanford University on July 25-28, 1989. Eighteen Working Groups are scheduled to meet. Send to ietf-request@venera.isi.edu for a copy of the preliminary agenda. Westine [Page 5] Internet Monthly Report June 1989 Following a suggestion at the April IETF meeting, another refinement was suggested to the agenda format to give even more time for Working Groups activity. The suggested schedule was: Days 1 and 2 9 am - 12 WG Morning session 1 pm - 4pm WG Afternoon session 4 pm - 5:30 Technical Presentations (in Plenary) Day 3 9 am - 12 WG Morning session 1 pm - 5:30 pm Technical Presentations (in Plenary) Day 4 9 am - 12 WG Reports This gives an additional period for WG sessions, making a total of five session, but retains the overall time available for technical Plenary presentations. This goal is to reduce the number of overlapping WG meetings. This format will be tried as an experiment at the next several meetings. 3. The "IETF:" and the "INTERNET-DRAFTS:" directories at the SRI-NIC.ARPA contain information about IETF activities, available by anonymous FTP. These directories have been reorganized and brought up-to-date. The "IETF:" directory contains: - a general description of the IETF (history, organization) - a summary of active Working Groups within the IETF - IETF meeting dates/locations - upcoming meeting information - the upcoming meeting agenda, and - README file with an overview of directory contents. In addition, the directory contains the charters, status updates and current meeting reports for each Working Group. These WG files are named in the following fashion: <WG NAME>.charter <WG NAME>.status <WGNAME>.report This Working Group information corresponds to the information from the most recent IETF Proceedings. The "INTERNET-DRAFTS:" directory contains draft documents that will be submitted ultimately to the RFC Editor to be considered Westine [Page 6] Internet Monthly Report June 1989 for publishing as RFCs. Review and comments are encouraged and should be addressed to the responsible persons or Working Group whose names and email addresses are listed on the first page of the respective draft. There is also a file containing an index and abstract of all draft documents in the directory. The following drafts are now available : <DRAFT-IETF-HOSTREQ-HRLL-00.TXT> Requirements for Internet Hosts -- Communication Layers, edited by Robert Braden for the Host Requirements Working Group, 16 June 89 <DRAFT-IETF-HOSTREQ-HRUL-00.TXT> Requirements for Internet Hosts -- Application Layer, edited by Robert Braden for the Host Requirements Working Group, 22 May 89. <DRAFT-IETF-PPP-REQ-00.TXT> Requirements for an Internet Standard Point-to-Point Protocol, edited by Drew Perkins for the PT-PT Protocol Working Group, June 1989 <DRAFT-IETF-PPP-IPDATAGRAMSTX-00.TXT The Point-to-Point Protocol(PPP): A Proposed Standard for the Transmission of IP Datagrams over Point-to-Point Links, edited by Drew Perkins for the PT-PT Protocol Working Group, June 1989 <DRAFT-UCL-KILLE-X400RFC822-00.TXT Mapping between X.400 (1988) and RFC822, Steve Kille, University College London, 5 June 89. For more information, please contact ietf-request@venera.isi.edu. Phill Gross (gross@SCCGATE.SCC.COM) Westine [Page 7] Internet Monthly Report June 1989 INTERNET MANAGEMENT No report received. PRIVACY During June, three RFCs were revised in preparation for release during July. The revisions were required because of differences in X.400 and X.500 naming conventions and because of changes to the X.509 spec which resulted from suggestions put forth by this task force. The RFCs undergoing revision are : Privacy Enhancement for Internet Electronic Mail: Part I: Message Encipherment and Authentication Procedures, Part II: Certificate-Based Key Management, and Part III: Algorithms, Modes, and Identifiers. Lyndalee Korn (lkorn@BBN.COM ) SCIENTIFIC REQUIREMENTS No report received. DSAB ---- No report received. Westine [Page 8] Internet Monthly Report June 1989 INTERNET PROJECTS ----------------- BARRNET ------- No Report Received. BOLT BERANEK AND NEWMAN INC. ---------------------------- SATNET Last month, we reported the retirement of the SATNET and its planned replacement by two point-to-point links from Europe to the US. In January, the UK to US link was installed. As of the end of June, Contel and Italcable were checking out the final piece of the point-to-point link from CNUCE (Italy) to DARPA (US). The tail circuits at each end have passed testing and only the international satellite circuit remains. As soon as the entire circuit is available, the gateways at each end will be connected up and final systems integration and test will be done. In the meantime, CNUCE can be reached via BITNET and the VAN Gateway at BBN. TERRESTRIAL WIDEBAND NETWORK During June, we installed WPSs at SRI and New York and transitioned the gateways at SRI to the new network. Work is underway to connect the RADC gateway to the NY WPS. We also installed a gateway at NRL and connected it to the Washington WPS. Over the summer, we will be continuing to install Wideband Packet Switches (WPSs) as power at the POPs and tail circuits between the backbone and the gateways become available. Throughout the month, the network continued to behave well with minimal software or hardware problems. A line-error problem appeared at the end of June between NY and BBN. This is currently being worked on and is expected to be corrected by the replacement of a bad CSU at the NY POP. Current sites (going from east to west): BBN BBN WPS NY WPS DARPA, NRL Washington WPS ISI LA WPS SRI SRI WPS Westine [Page 9] Internet Monthly Report June 1989 Future installation work will include: RADC, Ft. Monmouth (NY WPS) -- The RADC gateway and NY WPS have been checked out. A bad DSU was found at RADC and is being replaced. The site at Ft. Monmouth will be connected to the NY WPS as soon as the tail circuit is available. We do not yet have a confirmed date for the on-site wiring but the current estimate is for some time in August. CMU (Pittsburgh WPS) -- Negotiations concerning on campus wiring are proceeding, but we do not yet have a confirmed date. POP equipment installation and wiring will be scheduled once the CMU circuit is available. NCSA (Chicago WPS) -- A T1 circuit from the Chicago POP to NCSA, Urbana was ordered 6/13. Circuit and Chicago POP facility upgrade are expected to be available in July. WPS and gateway installation plans will be finalized as soon as there is a confirmed date for the circuit. TERRESTRIAL WIDEBAND TELECONFERENCING SYSTEM The conversion from the satellite-based Wideband to the terrestrial network is complete. The system will be available for use in 2-site and 3-site conferences on July 10. INTERNET R&D The VAN Gateway is now running EGP with CNUCE over Telenet. We have seen a number of outages on the line between BBN and RSRE. We are looking to see if there any common failure patterns. So far we have not seen any. The latest outage occured when RSRE (the building, not the gateway) was struck by lightning. We finished an overview paper on the Open Routing work for "Connexions." This paper discusses the major architectural decisions of the protocol and why those decisions were made. It is a complement to the architecture paper which goes into greater depth on what we plan to do and less depth on why. Bob Hinden (Hinden@BBN.COM) Westine [Page 10] Internet Monthly Report June 1989 CERFNET ------- SITE OUTAGES: The California State University's Office of the Chancellor's SWRL Facility was temporarily down on June 14th. The AGS cisco gateway box suffered a major malfunction. (This is the AGS that was used in the CERFnet pilot.) The new MGS cisco box designated for the SWRL facility was shipped out immediately to replace the loss. CIRCUIT OUTAGES: On June 5th, the University of California, Los Angeles (UCLA) link to San Diego Supercomputer Center (SDSC) was down from 10:00 a.m. to 11:00 a.m.. This was due to fiber vendor at UCLA mistakenly running test on the CERFnet link and rendering it out of service.for this period of time. Traffic to NSFnet from Hughes, UCLA and Los Nettos were consequently effected. On June 20th, Caltech reported a circuit outage. Apparently the ethernet used between Caltech CERFnet cisco Box and their ethernet, was disconnected from the cisco box for an indefinite period of time. However, they did not lose any connectivity because the Los Nettos cisco box which is connected via a different ethernet was able to route the CERFnet traffic. Also, on June 19th and 20th, a major outage occurred on CERFnet. The PSN (the ARPAnet connection at SDSC) went down on the 19th at approximately 18:00 hours. The link was down until 14:14 hours on June 20th. Our PSN is used to advertise Net 10.0.0.0 which is the default network in the ciscos used for CERFnet. When Net 10 announcements disappeared the cisco boxes did not know where to route packets because many sites were unknown. The reason for this outage is attributed to both ends of the SDSC PSN link to ISI and UCLA going down. OTHER NETWORK ACTIVITY: The installation of the CERFnet Backbone was completed in June. The University of California, Irvine (UCI) was installed on June 1st. Susie Arnold of CERFnet commented about the installation, "Because UCI is hooked to SDSC through a connection involving a 3-Com Bridge Box, there were some circular routing paths running for awhile." This bridge box has since been removed from the loop. UCLA, followed UCI and was brought on-line June 7th. The installation went smoothly. Again initially, routing issues Westine [Page 11] Internet Monthly Report June 1989 presented the only minor problem and were quickly solved. However, once the link was in use other routing problems arose. Routing between CERFnet and Los Nettos (at UCLA and Caltech) was affected. Several routing loops were discovered running between CERFnet, Los Nettos, and outside networks, such as NSFnet. These needed to be evaluated and eliminated accordingly to ensure efficient routing of all traffic. Walt Prue of Los Nettos and Susie Arnold of CERFnet, continued to work on routing at UCLA and Caltech between the two networks throughout June. During this process, Caltech suffered some difficulties when both of its cisco's, moon- doggie and cisco-kid, routed to each other, apparently each believing that the other was the best route for reaching UCLA. These problems have since been resolved and routing between the two networks is functioning well. Cooperation between Los Nettos and CERFnet continues to be advantageous. The CERFnet circuit between cisco routers at UCLA and SDSC has been upgraded from 56 kbps and is now operating at 512 kbps. UCOP recently conducted emergency testing and problem isolation work on the IDNX node sites running T1 at UCLA, UCOP, and UCSB on June 21st between 6:00 and 7:00 hours. (This was necessary because of recent upgrades to T1 circuits.) The UCOP IDNX was restarted, causing a short outage of all circuits between northern and southern campuses, and circuits at UC Berkeley and UC Davis. However, southern campuses were not effected. Some additional testing was required because the T1 between UCLA and UCSB continued to experience excessive errors UCOP scheduled testing late on June 22nd. This testing effected all traffic to UCSB, however circuits to other campuses were not effected. The line from SDSC to UCLA was moved at both ends to the correct cisco box at UCLA and SDSC. Routing to UCLA is following the same path that it did previous to the switch. The preferred route is via the T1 link to Caltech, across the Ethernet to the Los Nettos cisco box, across the T1 link to the UCLA Los Nettos cisco box. UCSB was to be brought on-line on June 23rd but the V.35 cable which connects the IDNX to the cisco box wasn't correct. The cable problem is being resolved and a new installation date for UCSB is pending the arrival of the correct cable. The CERFnet newsletter was distributed to over 200 recipients. Karen Armstrong (armstrongk@Luac.Sdsc.Edu) Westine [Page 12] Internet Monthly Report June 1989 CICNET ------- No Report Received CORNELL ------- No report received. Scott Brim (swb@chumley.tn.cornell.edu) ISI --- INTERNET CONCEPTS PROJECT Sufficient study of the SUN UNIX operating system has been completed and the experimental Source Quench congestion control modifications to IP are now being written. Greg Finn (finn@isi.edu) Paul Mockapetris attended the NASA OSI Naming Transition Workshop in Washington D.C. 13-16 June 1989. Four RFCs were published this month. RFC 1103: Katz, D., "A Proposed Standard for the Transmission of IP Datagrams over FDDI Networks", Merit/NSFNET, June 1989. RFC 1104: Braun, H-W., "Models of Policy Based Routing", Merit/NSFNET, June 1989. RFC 1105: Lougheed, K., (CISCO), Y. Rekhter, (IBM), "A Border Gateway Protocol (BGP), June 1989. RFC 1106: Fox, R., "TCP Big Window and Nak Options", Tandem, June 1989. One ISI Paper was published. ISI/SR-89-235: Katz, A., and S. Casner, "Supercomputer Workstation Communication, USC/ISI, June 1989. Ann Westine (Westine.ISI.EDU) Westine [Page 13] Internet Monthly Report June 1989 MULTIMEDIA CONFERENCING PROJECT We assisted in the preparation of a demo used by Craig Fields, Director of DARPA, in testimony for hearings on Senator Gore's "National High-Preformance Computer Technology Act of 1989". We collected timing data for actual transfers of images across the ARPAnet and WBnet for comparison to gigabit speeds. LBL incorporated this data into the demo program they prepared to display still images and "movies" at these different speeds. Development of an integrated conferencing environment continues. File distribution modes optimized for conferencing scenarios were added to MBFTPTOOL. Refinements were made to the packet video host's operator interface and its coordination with the packet voice host in our efforts to make the system more operational and robust. The teleconference room at ISI is back in service after the reconstruction of the 11th floor. The new room is larger and has improved acoustical treatment and lighting. The multisite conference capability is back in service as well, now on the Terrestrial Wideband Net. Delays are much lower than with the satellite network. Requests to use the facility should be sent to video-conf-requests@bbn.com. Steve Casner attended the User Interface Task Force meeting at Bellcore. Dave Walden, Eve Schooler, Steve Casner (djwalden@ISI.EDU, schooler@ISI.EDU, casner@ISI.EDU) FAST PARTS Alan Katz continued working on T.REX, a proposed Remote Execution protocol and continued learning C++ and the Interviews toolkit. Alan attended Xhibition '89 in San Jose., June 25-28. Alan Katz (Katz@ISI.EDU) JVNCNET, NORTH EAST RESEARCH REGIONAL NETWORK --------------------------------------------- No report received. Westine [Page 14] Internet Monthly Report June 1989 LOS NETTOS ---------- TRW was installed June 16th, and was made operational on June 19th. Unisys is planning to become a Los Nettos member. We have been working with CERFnet closely to optimize routing and to avoid routing loops. Walt Prue attended the Calinet meeting at Stanford June 22nd. There was much discussion about how to make routing work when a planned north-south line between Los Nettos/CERFnet and BARRnet is brought up. The routing is made complicated by the fact that the Arpanet and the NSFNet are connected to both regionals already. Additionally CSUnet and the U.C.DLA-Net are both connected to both regionals but do not wish to be transit networks. SNMP was brought up on several nodes of the network. We are using the MIT development kit. We have found SNMP useful for debugging routing problems and for obtaining traffic statistics programmatically. Walter Prue (Prue@ISI.EDU) MERIT/UMNET ----------- No report received. MIDNET ------ No report received. MIT-LCS ------- No progress to report this month. Lixia Zhang (Lixia@XX.LCS.MIT.EDU) MITRE Corporation ----------------- No report received. Westine [Page 15] Internet Monthly Report June 1989 MRNET ----- No report received. NCAR/USAN --------- No report received. NEW ENGLAND ACADEMIC AND RESEARCH NETWORK ----------------------------------------- NEARnet added three new members in June: Bull HN, Samsung Software America, and the Mass. Microelectronics Center. Reconfiguration of the link between NEARnet and the JvNCnet was completed. Operation of the network was stable, with six sites using 10MB microwave links to communicate (BBN, BU, Harvard, Lincoln Labs, MIT, Thinking Machines Corp.). The Northeast Network Coordinating Committee, consisting of representatives from NEARnet, JvNCnet, NYSERNET, and PREPnet, met on June 1 to discuss common problems and issues. The group will work together on routing, addressing, redundancy, and robustness, as well as other questions that affect all of the networks. by Laura Breeden (breeden@SH.CS.NET) NORTHWESTNET ------------ No Internet-related progress to report this month. by JQ Johnson (jqj@hogg.cc.uoregon.edu) NSF BACKBONE (Merit) ------------------- Expansion Proposal Approved In June, the National Science Foundation approved $6 million in funding for Merit's NSFNET expansion proposal to add nodes to the NSFNET backbone. This approval raises the total amount authorized for the five-year project to $20 million. In the near future, the NSF will solicit proposals from organizationswishing to become nodes on NSFNET. The NSF, in consultation with Merit, will then choose the new node sites by October or November of this year with implementation following, probably early in 1990. Westine [Page 16] Internet Monthly Report June 1989 NSFNET Reconfiguration update Phase A of the NSFNET reconfiguration has been completed and production trafficis running at T1 rates with redundancy at all nodes. The move to a more robust network is nearing completion. Interagency Connections Within the past six months efforts have been underway to design closer interconnection points between agency peer networks in order to rationalize routing and improve the manageability of those connections. Two major points of interconnection between the NSFNET, the ARPANET/MILNET, and the NASA ScienceNetwork were proposed and authorized as the first phase of this process. On the west coast the NSN connection at Ames via the Split-E-PSP connection to Palo Alto became operational on June 22. On the east coast the NSN connection to the NSFNET at College Park was implemented on June 15. Support from Milo Medin and Tony Villasenor of NASA, Bill Yundt of BARRNet and Jack Hahn of SURAnet was critical to this endeavor. The phaseover of networks from old to new gateways will continue throughout July. These changes represent significant progress architecturally as well as from the perspective of managing peer network interconnections, with implications for security and policy- based routing. The direct peer-to-peer interaction represents a transition in technology to establish EGP sessions with the mailbridges providing access to the ARPANET and MILNET. With the implementation of a split E- PSP, EGP peers do not need to be co- located with an NSS (Nodal Switching Subsystem). Changes in "routed" to support the new configuration were made in cooperation with IBM. Comparison of Traffic Counts Traffic on the NSFNET continues to increase with June packet counts topping theone billion mark for the first time: Westine [Page 17] Internet Monthly Report June 1989 ----------------------------------------------------------- May 1989/June 1989 Packets In Packets Out May 964,789,380 975,139,613 June 1,067,845,778 1,085,872,371 % increase 11% 11% ----------------------------------------------------------- Network connections are also increasing and now stand at: ----------------------------------------------------------- Primary routes 630 Secondary routes 384 Tertiary routes 59 ----------------------------------------------------------- July Internetworking Seminar The Internetworking Seminar, "From the Campus to National Networking: Connecting to the Rest of the World" is scheduled for July 27 and 28 in Denver. To obtain additional information please telephone 1-800-66-MERIT or send electronic mail to 'julyseminar@merit.edu'. by Patricia G. Smith (patricia_g._smith@um.cc.umich.edu) NTA-RE and NDRE --------------- No report received. NYSERNET -------- Mark Fedor and Martin Schoffstall attended the 2nd adhoc network management meeting held by Vint Cerf at NRI on 12 June 89. The outcome of this meeting will be published in an RFC by Vint. The X.500 "master" Directory Management Domain (DMD) for the US was moved to osi.nyser.net (192.33.4.20). Osi.nyser.net (a SUN3/260) is attached via X.25 to TELENET at 9.6kbps as well as having dual T1 access to NYSERNet. Additional X.25 PDN connections are currently being sought. After successful completion of a hardware upgrade of c.nyser.net (which runs a root DNS) "c" will begin running as a slave DMD for the US. Westine [Page 18] Internet Monthly Report June 1989 The NYSERNet White Pages pilot project which is based on X.500 achieved a number of milestones in June. The objective of this pilot is to incorporate about 1/2 a million records into the directory from the various NYSERNet institutions and enable users to readily access them. Fred is a new (beta) non-graphical DUA/client application which has some backward compatability in syntax to the WHOIS protocol (this is somewhat controversial). Additionally a new Administrator's guide to QUIPU has been produced and distributed to the participants. Documentation can be found via anonymous ftp on nisc.nyser.net in "wpp/pilot-ps.tar.Z". A one day information session in August is scheduled for NYSERNet technical representatives involved in the WhitePages pilot. Martin Schoffstall (schoff@rebel.nyser.net) OARNET ------ No report received. PITTSBURGH SUPERCOMPUTER CENTER ------------------------------- No report received. SAN DIEGO SUPERCOMPUTER CENTER ------------------------------ A 56K link to LANL, via cisco's, has been put into operation. Currently only traffic between LANL, SDSC, and UCSD is being carried. The PSN (26) at SDSC will be removed during July. To that end, we have routed traffic that was aimed at it to ISI's PSN via CERFnet and Los Nettos. We elected to do this now rather than retain a host link until the fall as our path is now all T1 (3 hops) vs the 56k direct link. The routing for SDSC is much simplified. An EN641 from Network Systems has been ordered to provide a gateway between our Ethernet and the Hyperchannel. This will support our Unicos XMP and, later, YMP. Westine [Page 19] Internet Monthly Report June 1989 OPUS, our hardworking SUN 3/50 is at last out of the networking area. GATED is now in production use from MultiNet and our VMS systems. SDSC continues to be active in CalNet, Farrnet and IETF by Paul Love (loveep@sds.sdsc.edu) SESQUINET --------- No report received. SRI ---- DDN NIC The ARPA -> MIL name transition is proceeding on schedule. So far we have completed the rename process for more than 1,100 networks, and 26 hosts, into the MIL domain. We have been assigning an average of 55 new connected IP network numbers per month; by the end of June there were a total of 1,648 numbers assigned. Domain registrations are also steadily increasing; we've been registering an average of 46 new domains each month. There are now 984 domains registered; 41 of those are top-level domains and 913 are second-level domains. Mary Stahl (STAHL@NIC.DDN.MIL) Internet Research Jose J. Garcia-Luna attended the FRICC Workshop on Coordinated OSI Transition, which took place on June 14th at the Marriot, Wash Dulles Airport. The meeting was chaired by the IETF chair, Phill Gross, and focused on the issues important for implementing OSI migration with regard to directory services in the Internet. Jose J. Garcia-Luna and Zaw-Sing Su attended INARC meeting June 1-2 at University of Delaware, chaired by Dave Mills. Zaw-Sing Su (ZSu@SRI.COM) Westine [Page 20] Internet Monthly Report June 1989 SURANET ------- No report received. TEXAS HIGHER EDUCATION NETWORK ------------------------------ THEnet's Network Information Center machine, "nic.the.net" was upgraded from a VAXstation II to a VAXstation 3500. by Tracy LaQuey (tracy@sirius.cc.utexas.edu) UCAR/BOLT BERANEK and NEWMAN, INC., NNSC ---------------------------------------- The NNSC collected several entries for the Internet Resource Guide, which will start distribution during the first week of July. For more information about the guide, send messages to resource- guide@nnsc.nsf.net. Craig Partridge attended the INARC Meeting and the End-2-End Task Force meeting, and Karen Roubicek participated in an informal meeting of the IETF User Services Working Group held during the FARNET meeting. by Karen Roubicek (roubicek@nnsc.nsf.net) UCL ---- 1. We have put up the Nottingham (J.Onions) Remote Operations version of the Network Time Protocol (NTP). This is peering over X.25 (JANET) with Nottingham, and with several US sites over the Internet. We will shortly be getting a Rugby Clock Receiver so that we can run a stratum one clock for the UK. Typical ntpd output, Day 1: Neighbor address 126.0.0.1 port:10123 local address 126.0.0.1 Reach: 0377 stratum: 2, precision: 250 dispersion: 48.000000, flags: 1301, leap: 0 Reference clock ID: [128.243.20.1] timestamp: a8533dff.b435bd50 hpoll: 10, ppoll: 10, timer: 1024, sent: 47 received: 47 Delay(ms) 1234.00 1455.00 1035.00 974.00 1235.00 1234.00 1252.00 1255.00 Offset(ms) -30467.00 -30314.00 -30417.00 -30427.00 -30286.00 -30290.00 -30283.00 -30286.00 Westine [Page 21] Internet Monthly Report June 1989 delay: 974.000000 offset: -30427.000000 dsp 48.000000 Neighbor address 128.8.10.1 port:123 local address 128.16.5.50 Reach: 0353 stratum: 1, precision: 246 dispersion: 16.000000, flags: 9301, leap: 0 Reference clock ID: WWVB timestamp: a8533f92.ef5c0000 hpoll: 6, ppoll: 10, timer: 64, sent: 188 received: 159 Delay(ms) 1154.00 1307.00 1285.00 1067.00 974.00 956.00 984.00 984.00 Offset(ms) 3.00 -55.00 -120.00 -4.00 36.00 46.00 49.00 59.00 delay: 956.000000 offset: 46.000000 dsp 16.000000 2. Work continues on designing fair sharing algorithms for the potential UK-US high bandwidth link. John Crowcroft (jon@CS.UCL.AC.UK) UNIVERSITY OF DELAWARE ---------------------- 1. The Workshop on the Future of the Internet System Architecture and TCP/IP Protocols was held 1-2 June 1989 at the University of Delaware. Paul Schragger and Mike Davis spent some days transcribing and editing the workshop proceedings. Since this project produced rather more output than originally anticipated, work continues on further editing and preparation for publication. 2. Final touches are now being added on a detailed analysis and commentary of the Border Gateway Protocol (RFC1000). A plan for proposed experiments using resources of the Research Internet Backbone is nearing completion. The plan is based on relatively simple modifications to selected hosts and gateways which allow participation of research nets located far from the Backbone itself. 3. Mike Davis brought up the current version of SNMP, which now runs in assorted Suns and PCs. We are now watching our campus nets and selected SURA paths to the NSF backbone Point of Internet Presence (PIP, hereby introduced). Our system staff brought up traceroute, which immediately boggled minds demonstrating tortuous routes used by our clockworking friends. 4. The various Fuzzball time servers scattered throughout the Internet are logging increasing numbers of ICMP messages in patterns which suggest occasional massive routing loops Westine [Page 22] Internet Monthly Report June 1989 lasting in the order of a half hour. Most of these appear to involve contorted ARPANET/NSFNET paths and occasionally involve MILNET. Logging procedures have been modified to select additional related information for subsequent analysis. 5. New NTP sites have come online in Australia, Hawaii and Norway. It is reported that Japan, New Zealand and Finland will shortly be next and soon the first WACC (Worked All Continents Clocks) certificate should be awarded. Following suggestions by several people, the NTP spec was amended to revise the authentication mechanisms and move the text to an appendix. Also, a new appendix on control messages was added. Finally, a precision oven-controlled crystal oscillator arrived and began begging for an interface it a handy Fuzzball time server. Dave Mills (Mills@UDEL.EDU) UNIVERSITY OF ILLINOIS AT URBANA-CHAMPAIGN/NCSANET -------------------------------------------------- No report received. WESTNET -------- 1. Hewlett Packard in Fort Collins, Colorado (Westnet is secondary to BARRNET) and the Idaho National Energy Laboratory in Idaho Falls, Idaho became connected this month. 2. We expect Idaho State University in Pocatello, Idaho to become connected during the month of July. This will result in a minor topological change in the circuits in Idaho. 3. We are investigating upgrading all of Westnet's interstate circuits to T-1 for next year, and add a redundant link to run between the University of Arizona in Tucson and New Mexico Technet in Albuquerque. There is a need for faster response time during interactive sessions, and for greater bandwidth for research applications. by Pat Burns (pburns@csupwb.colostate.edu) Carol Ward (cward@spot.colorado.edu) Westine [Page 23] Internet Monthly Report June 1989