[IMR] IMR88-06.TXT JUNE 1988 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). BBN LABORATORIES AND BBN COMMUNICATIONS CORPORATION --------------------------------------------------- WIDEBAND NETWORK The Wideband Inter Switch Trunk (IST) lines which carry Arpanet traffic continued to improve their performance during the past month. After correcting both hardware and software problems with the BSATs, these lines have carried heavy traffic loads with few problems. The Wideband Net has also been supporting the development of multi- media, multi-site conferencing. Additional conferencing sites are currently being setup up to provide multi-site conferencing over the Wideband Net. To aid the simultaneous use of the Wideband Net by the Arpanet lines and the video conferences, software was distributed to allow fast, synchronous reconfiguration of the network. Now, sites that are infrequently used can be deconfigured; the channel capacity used by those sites is then freed up for use by other sites. When the infrequently used sites need to use the network, they can be reconfigured. SATNET The SATNET has been very stable through the month of June. We have had no unscheduled outages of the SATNET SIMPS or PSP terminal hardware. The U.S. was isolated from the SATNET because of a problem at the WUI telco office from 6/19/88 to 6/20/88. The Westine [Page 1]  Internet Monthly Report JUNE 1988 availability of the SATNET was above 99% from tests run by ISI for all sites. The CNUCE and UCL gateways had hardware failures this month. The UCL gateway was repaired by replacing a processor node. Because of customs delays in getting a processor to CNUCE service was restored by loading software which allows the gateway it to operate with only 2 processors. The RSRE to Goonhilly link was down from 6/19/88 to 6/29/88 due to problems on the kilostream line. INTERNET R&D The Internet continues to grow. We are currently at about 450 operational networks. We completed the Telenet certification of the Butterfly VAN Gateway gateway X.25 interface and submitted the results to Telenet. From our understanding of the test we think we passed. We await the results from Telenet. We attended the IETF meeting in June and the Open Routing Working Group held a meeting before the IETF. VAX NETWORKING PROJECT A multicast routing daemon is now running under 4.3bsd (although still not fully debugged). We have also begun work to implement the proposed Gateway Discovery ICMP messages on our 4.3bsd test host. VAX NETWORKING PROJECT A multicast routing daemon is now running under 4.3bsd (although still not fully debugged). We have also begun work to implement the proposed Gateway Discovery ICMP messages on our 4.3bsd test host. Bob Hinden (Hinden@BBN.COM) ISI --- Internet Concepts Project Recent investigation into the issue: How might one design network layer controlling protocols so that they are highly resistant to attack either accidentally or by deliberate intervention? This issue, the vulnerability of some current network layer protocols Westine [Page 2]  Internet Monthly Report JUNE 1988 and an example of an attack resistant protocol are discussed in ISI-RR/88-201. Greg Finn (finn@isi.edu) Paul Mockapetris attended the Internet Engineering Task Force Meeting in Annapolis, MD, 14-17 Jun. Four RFCs were published this month. RFC 1055: Romkey, J., "A Nonstandard for Transmission of IP Datagrams Over Serial Lines: SLIP", June 1988. RFC 1056: Lambert, M., "PCMAIL: A Distributed Mail System for Personal Computers", MIT, June 1988. RFC 1057: Sun Microsystems, Inc., "RPC: Remote Procecure Call Protocol Specification", Version 2, June 1988. RFC 1058: Hedrick, C., "Routing Information Protocol", Rutgers University, June 1988. Ann Westine (Westine@ISI.EDU) Multimedia Conferencing Project Work this month was geared primarily towards the July 7th three- site teleconference between DARPA, ISI, and BBN. The system underwent the final evolutionary phase to accommodate multisite conferencing. This meant adding capabilities to the packet video program and video codec for receiving multiple streams of video data and delivering them to separate quadrants on the Image30. To supplement this, the multimedia conference control program was extended with a multipoint protocol for conference connection and disconnection. Furthermore, it now has the ability to operate in an "autopilot" debug mode, for testing of remote conference set up when no user is around at the remote site to help out. A version of echo canceler software on the Evakit, the development system for the NEC7720 signal processing chip, was completed. Currently the software implements an echo canceler with 51 tap weights and includes near-end speech detection and center clipping of residual echo. The near-end speech detection scheme works by setting a threshold around a far-end signal with suitable attenuation and time delay which has the maximum correlation with the near-end signal with out near-end speech. Near-end speech is declared if it falls outside this limit. This version of the Westine [Page 3]  Internet Monthly Report JUNE 1988 software has been tested alone and seems to perform satisfactorily. The next step is to test the software at both ends and this means moving the software into the EPROM version of the NEC7720. Currently, difficulties with the EPROM version have been encountered. Brian Hung, Steve Casner, Dave Walden, Eve Schooler (hung@ISI.EDU, casner@ISI.EDU, djwalden@ISI.EDU, schooler@ISI.EDU) NSFNET Project Annette DeSchon started work on a "BFTP Tool", which runs in the SunView(TM) window environment. This tool will provide a window based user interface for use in submitting file transfer requests to a BFTP server running on a Sun workstation. Work on the associated BFTP RFC is progressing. Annette DeSchon (DeSchon@ISI.EDU) Supercomputer and Workstation Communication Project Alan Katz continued work on a split editor that runs under GNU Emacs. Alan added a remote dired, rlogin, and ftp modes. Alan attended the Fourth Annual Workshop on Networking and Supercomputers sponsored by NCAR and held in Boulder, CO, June 7-9. Alan Katz (Katz@ISI.EDU) MIT-LCS ------- After almost a year of work, our network simulator has reached a stable stage and we would like to share it with friends who are interested in playing with it. Whoever is interested may contact Andrew Heybey (atheybey@ptt.lcs.mit.edu) about getting the code. They should be pre-warned, however, that the simulator is still an experimental system and therefore we do not make any guarantee about it being bug-free. We will try our best to fix any bug reported. We will also appreciate any comments/suggestions about how to further improve the simulator. Lixia Zhang (Lixia@XX.LCS.MIT.EDU) NTA-RE and NDRE --------------- No report received. Westine [Page 4]  Internet Monthly Report JUNE 1988 SRI --- Internet Research No Internet related progress to report besides my attendance at the Open Routing Working Group and IETF meeting at the Naval Academy in Annapolis, Maryland, 14-17 June. Zaw-Sing Su (ZSu@SRI.COM) UCL --- Infrastructure. The FDDI Campus net is now operational. In principal, it is connected to the internet via the CS gateways. In practice, we are access controlling it, until we work out policies on routing. Research. Enhancements to the MAC Routing work include a simple management protocol (not SNMP), and a simple system to display bridge network topologies on a PC. Meetings. Jon Crowcroft attended the Security meeting at RIACS, NASA Ames, to represent European views on policy based routing issues. John Crowcroft (jon@CS.UCL.AC.UK) UNIVERSITY OF DELAWARE ---------------------- 1. Jeff Simpson has narrowed the scope of his research to the study of policy-based routing algorithms and related technology. He has amassed an amazing library of past papers and reports in this area, including the work of the Internetters, ANSI and NBS. He is currently working on an issues paper on policy models and mechanisms to stimulate further discussion and refinement. The Mike Minnich family produced a new Internetter, class of 2010. 2. Paul Schragger continues to work on feedback-control and rate-based concepts applied to transport-level communications. Westine [Page 5]  Internet Monthly Report JUNE 1988 Paul spent an exciting day with Dave Clark's group at MIT discussing these concepts and planning further mischief. He has brought up a snazzy network simulator from MIT on one of our colorful Suns and is evaluating it for use in support and refinement of these models. We have ordered a copy of the commercial OPSNET simulator as well to serve the same purpose. 3. The Network Time Protocol (NTP) community continues to flourish, with new primary servers online at MIT and planned for DECWRL and NASA/AMES. Two NSFNET Backbone sites, SDSC and UIUC, have announced plans to support primary servers using existing fuzzballs and new or refurbished WWVB clocks. The tables in three of the existing primary servers were expanded from about 60 to over 100 peers to support the recent surge in new chimers. 4. Effective 5 July our U Delaware Unix systems will no longer be connected to the ARPANET, but will continue service via SURANET. Our frisky fuzzballs, including NTP servers, will remain connected to the ARPANET and everything else in town, but will not carry regular campus traffic. We are in process of rehoming everything on a single class-B net, which of course is breaking every little thing. 5. Dave Mills attended an RADC meeting on the Dissimilar Gateway Protocol (DGP) in Colorado Springs, CO, an IAB telemeeting in Washington, DC, an INENG meeting in Annapolis, MD, and a NAS National Security Telecommunications Committee meeting in Woods Hole. A draft report covering issues in reliable telecommunication network synchronization was submitted and is now being revised. A paper on the history and application of the intrepid Fuzzball was revised and submitted for the SIGCOMM 88 Symposium. A draft report discussing issues in the realization of a nationwide gigabit network is now in preparation. Other work in progress includes a paper on DGP in collaboration with Mike Little of M/A-COM Government Systems, which is spelled SAIC as of 1 July. Dave Mills (Mills@UDEL.EDU) Westine [Page 6]  Internet Monthly Report JUNE 1988 NSF NETWORKING -------------- UCAR/BBN LABS NNSC The NNSC published and distributed issue 4 of NSF Network News. Send requests for copies to nnsc@nnsc.nsf.net. We are still increasing the number of sites for which we are providing backup domain service. Craig Partridge chaired the Management Information Base Working Group meeting of the Internet Engineering Task Force in Annapolis. by Karen Roubicek (roubicek@nnsc.nsf.net) NSFNET BACKBONE PROJECT CORNELL UNIVERSITY THEORY CENTER Scott Brim attended the IETF meeting and a meeting at RIACS, and will be attending an Autonomous Networks Task Force Meeting this week, primarily in relation to issues of policy-based routing. Jeff Honig continued development of the gated program in preparation for the deployment of the new backbone. Doug Elias made presentations at the SuperComputer Workshop in Boulder Colo. and to MERIT in Ann Arbor on the NSFNet traffic studies. He also presented a session on System Science Tools at the ACM Conference on Simulation in Pittsburg using the NSFNet traffic data for his model. During the period from October 1987 to May 1988, the amount of traffic carried on the backbone was four times the amount handled during the same time period from 1986 to 1987. The Cornell node alone experienced an eightfold increase in the traffic handled. The data indicates that Suranet contributed approximately 25% of the increase to Cornell's traffic. by Martyne M. Hallgren (martyne@tcgould.tn.cornell.edu) NEW NSFNET BACKBONE Saturday evening, June 11, provided an historic moment for the new NSFNET backbone. At that point, all thirteen nodes had achieved on-line status and were able to send packets across the backbone. After two weeks of testing and debugging, several sites were fully operational and began announcing reachability to their regionals. By July 6, the following sites were announcing routing information to their regionals: The Merit Computer Network, NCAR, NCSA, SURANET, SESQUINET, NYSERNET, JVNC, WESTNET, and NorthWestNet. The Westine [Page 7]  Internet Monthly Report JUNE 1988 other four nodes will be phased in during early July. Upgrades to the Nodal Switching System (NSS) include Version 1 software and some hardware improvements. Other changes include improved overall network management capabilities. As of July 1, the Merit Network Operations Center (NOC) hours are 8:00 a.m. - 12:00 a.m. EDT, Monday through Friday and 8:00 a.m. - 6:00 a.m. EDT, Saturday and Sunday. By July 30, the NOC plans 24- hour operations. An article on monitoring ("Monitoring Data Exchanges between the NSFNET Backbone Network and its Attached Regional Clients") is available in the NSFNET directory on the Information Services machine by anonymous FTP of the file: monitor.mtg-6-88 by Laura Kelleher (Laura_Kelleher@merit.edu) NSFNET BACKBONE SITES & MID-LEVEL NETWORK SITES UNIVERSITY OF ILLINOIS AT URBANA-CHAMPAIGN/NCSANET Ross Veach attended the Baltimore IETF on June 15-17. His chief interest were the discussions relating to Son of NSFnet. Ed Krol, Charley Kline, and Ross Veach met with the Merit group in Ann Arbor, MI on June 20 to discuss operations and routing issues. UIUC has split the main building Ethernet into two networks. The new one is a Class C net (uiuc-remote) for connecting all gateways with off-campus connections. This includes the NSS, uxc (Arpanet), and various IP routers that reach either internal UIUC networks or the regionals. Work is in progress at NSC to port gated 1.3.1.48 to their EN641 router. UIUC will beta-test its capability to peer with the NSS. Nine regional nets are being announced to the NSS: UIUC, UIC(hicago), UofChicago, Northwestern U, Notre Dame, UofWisc Parkside, UofMinn, Mayo Clinic, and ETA. The MIDNets will not be announced until p4200s can be convinced not to believe EGP routes over IGP routes. NSS backbone to Arpanet traffic is being held off pending resolution of ongoing routing discussions. by Paul Pomes (paul@uxc.cso.uiuc.edu) Westine [Page 8]  Internet Monthly Report JUNE 1988 JOHN VON NEUMANN NATIONAL SUPERCOMPUTER CENTER This report is designed to inform the JVNC Consortium and JVNCnet network members as well as the Internet community of monthly status of the JVNCnet network. The data used on this report is collected using a number of techniques developed at JVNC, together with data from the JVNC operations group. Network brief: The John von Neumann National Center's high speed network (JVNCnet) connects seven north-east states plus two mid-west states. JVNCnet's 13 Consortium Institutions, plus JVNC and 10 non Consortium Sites form the 24 node network. The topology of the network (see attached diagram) is a combination of tree and double rings providing redundancy and high bandwidth access to the JVNC center, the NSFnet backbone and the ARPANET. The high speed links are mostly T1 lines (1.544 million bits per second), the rest are 56,000 bits per second (both terrestrial and satellite). The routing switches are a combination of VAXs, UB routers and CISCO routers. The network is operated from the JVNCnet Network Operations Center (NOC) located at the John von Neumann Center, which is staffed from 9 a.m. to 5 p.m., Monday through Friday and covered by the JVNC Operations group during off hours. Monthly Status Overview: The overall uptime for the gateways this month is 95.36% (worst case, this number considers that all gateways are unreachable when JVNCA is down, which is why it is called worst case). The measured uptime when JVNCA was available (99.24% of the time) was for an average on all the gateways 96.09% available. Traffic on the JVNCA gateway has been higher this month than last month, with a total number of packets in and out (of one of its ethernet interfaces) of 125,237,270 packets. Plans are moving ahead for the connection of the JANET (Joint Academic Network) in the UK to the NSFNet at JVNC. We are still re-routing all traffic between Penn State and NCAR and Princeton and NCAR via JVNCnet and the University of Colorado. The users are very satisfied with the access that they have to NCAR over JVNCNet and UC. We received requests from users to look at the possibility Westine [Page 9]  Internet Monthly Report JUNE 1988 of rerouting traffic from their sites to other NSF supercomputer centers. The communication rooms and facilities have been in the process of upgrade for the last two months, the first phase was the construction of an annex to allow for the installation of the new NSFNet equipment, the annex to the communications room complies with all the requirements in terms of electric power and air conditioning as well as space. This phase has been accomplished. The work continues in the upgrade of the communication facilities. Our "new" NSFnet node is up and running, it is at the present time connected to MERIT, it is happily "talking" EGP with a router at JVNC, and is getting ready to pass data. The internal JVNC network has been redesigned to allow for better access to the different facilities available (supercomputers, front ends, graphics stations, printers, external networks, etc...), and the implementation is moving along. Once fully implemented we will be able to account for utilization of all network resources within the JVNC LAN and interaction with the external networks as well as the JVNCNet regional network. The JVNC Network Operations Center (JVNCNet NOC) provides support to the JVNCNet network as well as the internal network of the JVNC. The NOC is staffed by the network staff from 9 a.m. to 5 p.m. Mon- Fri, and is on call between 5 p.m. and 9 a.m.. The computer operators monitor and perform minor troubleshooting tasks between 5 p.m. and 9 a.m. and provide for backup network operations center. Routing: A program was written, and runs in the background on JVNCA and JVNCB to check for routing loops and clears the problem when it occurs, the program runs every 5 minutes. A way to avoid the routing problems seen from outside sites is by filtering the information listened to from the site, to accept only the routing information from the directly attached campus only. A plan is under current evaluation on that effect. A program runs every 10 minutes on JVNCA to detect number of changes on the routing database of the "gated" program. These changes are recorded as and represent changes in routing protocol, gateway or metric (in milliseconds). During the month of June there were certain days where the routing activity has been out of normal, getting peaks of 1600 route changes. These data could be correlated with events on the NSFnet backbone or/and the ARPANET, and represent a degree of stability (or instability) in those networks and the JVNCNet network. We Westine [Page 10]  Internet Monthly Report JUNE 1988 have seen in the past that most of these changes occured as metric changes for routes on the NSFnet backbone. Access to JVNC: Arpanet: The JVNC PSN ( PSN 16 ) is fully operational. It is at the present time connected to JVNCA and will soon be connected to two routers (part of the JVNC-LAN). All the gateways within the JVNC system, that are under JVNC staff administration default now to JVNCA which shares its routing information with the ARPANET and NSFNet (phase I backbone). JVNC is announcing to the ARPANET 14 networks with egpmetric of 0 and 16 networks with egpmetric of 3. Some of these networks have been disconnected from the ARPANET and JVNC is providing their only access to it. NSFnet: The phase I backbone is still operational. The "new" backbone, (phase II) is expected to come on line by July 1st. At this point in time it seems to be moving ahead according to plans. JVNC has been exchanging routing information with the new backbone for a few weeks, and test traffic has been moving across. Bitnet: The JVNC Bitnet gateway (JVNCC) is performing well for mail delivery between Internet and Bitnet. Tymnet/Telenet: Tymnet without news. Dial-in: No significant news. JVNC Local Area Network: In the process of modification to a new configuration. Systems included are 5 ethernet segments (each one with different IP subnet address), built with redudant paths. Appletalk connections for the whole building with gateways to Internet. For more Information Contact: Network Operations: JVNCnet NOC, electronic mail address: "JVNCnet-noc@jvnca.csc.org" phone number: (609) 520-2000, x448 Westine [Page 11]  Internet Monthly Report JUNE 1988 Network Informations: JVNCnet NIC, electronic mail address: "JVNCnet-nic@jvnca.csc.org" phone number: (609) 520-2000, x448 --------------------------------------------------------------------------- JVNCnet Network Topology ------------------------ Boston U.---Harvard---MIT---Brown---Wesleyan | | | Dartmouth---Northeastern | Yale | | | Umass (Amherst) | | | | | | ============ | ----------------|| ||------------- || || IAS_--------------------|| ||------------U. of Penn Montclair State---------|| JVNC ||------------Penn State NYU---------------------|| ||------------U. of Colorado Columbia----------------|| ||------------Princeton U. of Arizona-----------|| ||------------Rutgers Rochester---------------|| ||------------NJIT- -Stevens ============ | ------- -UMDNJ ---------------------------------------------------------------------- Westine [Page 12]  Internet Monthly Report JUNE 1988 JVNC Local Area Network Configuration ------------------------------------- Regional Network LAN Front-ends LAN | | =============== ==================== | | -------------------------------------- | JVNC Internal routers |------- Telenet NSFnet ----| |------- ARPANET | |------- Dial-in -------------------------------------- | | ================ ==================== | | Graphics LAN External networks LAN --------------------------------------------------------------------------- by Sergio Heker (heker@jvnca.csc.org) NATIONAL CENTER FOR ATMOSPHERIC RESEARCH AND UNIVERSITY SATELLITE NETWORK PROJECT The USAN Satellite system was officially down for a four day period to install the long awaited ComStream radio updates. The network was totally down for only 24 hours however, with the remaining time devoted to technician travel, BERT testing and site debugging. During the ComStream changeover, the Institute for Naval Oceonography was added as a new member of the USAN Satellite net. by Don Morris (morris@windom.ucar.edu) PITTSBURGH SUPERCOMPUTING CENTER No report received. SAN DIEGO SUPERCOMPUTER CENTER The last RT for the new NSFnet NSS has arrived at SDSC. Our ARPAnet PSN is now announcing via EGP our own two IP networks (the local Ethernet and the Apollo ring), UC Santa Barbara, UC Ivine, and the Salk Inst. We have given some thought to ARPAnet gateways here at SDSC, and have decided that it would be a good thing to get our current one, a much used uVAX, out of the gateway Westine [Page 13]  Internet Monthly Report JUNE 1988 game. We are considering a couple of alternatives: - Use the new NSFnet gear. This would mean getting an IBM X.25 card for the PSPs. This would greatly simplify the routing out of SDSC both for ourselves and our attached nets. - Use the existing NSFnet fuzzball. We can move the ACC 5250 card from the uVAX into the fuzzball, hook it up to the PSN and have a dedicated box for the link. OPUS, our SUN 3/50 running gated, is now announcing a few nets via EGP into the new NSFNET backbone (but is not listening to any). We have discovered a bug in V7.4b of the Proteon code which distributes static routes via RIP even though it was not told to do so. This is wreaking havoc on BARRnet, since we had a static route to net 31 (the Univ. of Calif. library automation system). At present we have a filter in the p4200 which prevents traffic to/from net 31 (which does not make our librarian happy) until we get new software from Proteon. We still are looking forward to upgrading the p4200 with Proteon's newer CPU, Ethernet cards(a pair), and level 8.0 software. by Paul Love (loveep@sds.sdsc.edu) BARRNET No report received. MERIT/UMNET While Merit has supported the MIT-SLF protocol on its many asynch lines for almost a year now, support for SLIP (Serial Line Internet Protocol), typically used in Unix systems, has just been implemented as an additional protocol. This will allow users who wish to use an inexpensive asynchronous serial line for an Internet gateway to connect a port on a host to a port on a Merit SCP for this purpose. The SCP will act as a subnet gateway in order to pass IP datagrams at a speed up to 19.2Kbps. We are beta testing this week and expect to support SLIP on any of the 10,000 Merit ports within a few weeks. Westine [Page 14]  Internet Monthly Report JUNE 1988 We are currently considering the technical feasibility of supporting Dial-Up IP in the same way CSnet has announced this capability. This is a modification of SLIP with added identification/authentication. by (Laura_Kelleher@merit.edu) MIDNET No report received. MRNET We are still receiving user reports of problems encountered while using NSFnet/ARPANET; however there have been problems with our ARPANET interface that may have caused some of these problems. We have also begun to measure utilized bandwidth on the 56kb line from MRNet to Illinois, and on the 56kb line from MSC to ARPANET. In general, during June NSFnet appeared fairly reliable, although some users reported that FTP transmissions were slow (which is one of the reasons we are beginning to measure utilized bandwidth). by Ken Carlson (kgc@uf.msc.umn.edu) NORTHWESTNET NWnet operated in full production mode throughout May and June with its primary link to NSFnet being a leased 56 Kb land line from U of Washington to SDSC. We experienced an unusually high rate of hardware and line problems around the network; most were diagnosed and repaired within a day. We now have DECnet in production on all Proteon routers on the network. Although DECnet connectivity exists throughout the network, there is currently very little applications use; the vast majority of traffic is still IP, though we expect this to change somewhat in the next few months as we move to a genuine dual- protocol environment. Because of area-number management problems, DECnet access is currently limited to academic NWnet members only. Engineering design is in process to expand the network to include 4 new non-academic members. Also, a voice grade leased line is now in place from a NWnet router at the U of Washington to the University of British Columbia. As soon as some line noise problems are resolved, this will provide the latter institution with TCP/IP access to NSFnet through NWnet. Westine [Page 15]  Internet Monthly Report JUNE 1988 The Boeing Network Managment Center is experimenting with SGMP on a Sun workstation for network monitoring and control. As of June 28, the Seattle NSS is up and running reliably at the U of Washington with NWnet networks advertised by EGP to the new NSFnet backbone (which, as of that date, appears to be down more than it has been up, doing "destructive testing" until the cut-over date of July 1). All appears ready for the big cutover; more data next month. by JQ Johnson (jqj@hogg.cc.uoregon.edu) Westine [Page 16]  Internet Monthly Report JUNE 1988 NYSERNET Clarkson Albany | | | | | | Rochester ========== Cornell ======= Syracuse ====== Rensselaer || || / || || || / || || || / || || || NyserNet || || ..... Oswego || NISC || Buffalo ... Alfred || || | :..... Fredonia || || | || || | || || | || || | || || | Stonybrook ____ Columbia ====================== N Y U ===:\ | / | | __/ | || | Brookhaven ____/ / | ___/ | || | / | / | || | AOA === NYNEX S+T NSMAC _____ C U N Y _______ Poly || | || | \ || | || | \ || | || | \ || | Compass -------- Garden City | \_____ Rockefeller | | || | | || | | || Binghamton _______________________________| White Plains CO _________________________________________________________ | | | KEY: | | | | Line Speed Represented as: | | ------------------ ----------------------------------- | | NYTel RCI Kbps line type examples Planned | | --- --- ------ --------- -------------- ------- | | | | T1 DS1 1,540 double || = // \\ ~ | | | | DDS DS0 56 single | _ / \ ^ | | | | (sub-rate) 9.6 dots : . , ` | |_________________________________________________________| Westine [Page 17]  Internet Monthly Report JUNE 1988 JUNE EVENTS: - NYSERNet/NISC moved to new headquarters - installation of NYSERNet to NSFNet Gateway (N2NGW.NYSER.NET), but no traffic - completion of IDEA11-01 SNMP INOC, no UNIX agent/server yet (15July88 est) - continued work on ANSI/ISO Z39.50 and application bridge for same - Mark Fedor and Marty Schoffstall participated in Annapolis IETF - NYSERNet Site Technical Meeting (nysertech) at NYU - Bill Schrader and Marty Schoffstall gave RIB presentation to FRICC technical group - Board of Directors meeting in Albany by Martin Lee Schoffstall OARNET No report received. SESQUINET The complete initially proposed SesquiNet configuration. The following campus networks are being served, and are advertised via EGP to NSFnet and (currently via UIUC) the Arpanet core: Baylor College of Medicine 128.249 BCM-Technologies 192.31.88 Houston Area Research Center 192.31.87 Prairie View A&M University 129.208 Rice University 128.42 Texas A&M University 128.194 Texas Southern University 192.31.101 and the University of Houston 129.7 The new NSFnet backbone node at Rice University became operational during the last week of June, and has proved quite reliable. FTPs of 96kb/s across the new NSFnet are typical. All our current problems relate to sites on NSFnet that have not yet gotten their routing through the new backbone to work yet. Our Arpanet line finally went down during the last week of June. During the first part of the month it had been very stable and useful, and we will miss it. Our current primary core gateway advertiser is now uxc.cso.uiuc.edu, thanks to Ross Veach. We will probably shift this to one of the gateways that will serve to exchange packets between the Arpanet and the new NSFnet backbone some time this month. Westine [Page 18]  Internet Monthly Report JUNE 1988 During June 15-17 I attended the IETF meeting at Annapolis. My primary agenda there were issues of routing exchange between NSFnet and the mid-level nets. During June 22-24 I attended a workshop at NASA/Ames on Policy- based Routing. This workshop dealt both with the technical aspects of policy-based routing and with specific policy constraints of the federally supported research networks represented by the FRICC. We will soon resume our testing cisco's support for dual protocol (IP and DECnet) routing. by Guy Almes (almes@rice.edu) SURANET No report received. WESTNET 1. The IBM NSS is working at the University of Utah, in a sporadic manner. Connectivity, once established, is very stable and fast - with significantly less delay than over the ARPANET. 2. After analyzing some traffic statistics, we have orderer a T-1 line to be installed between the University of Colorado at Boulder and NCAR. We have also ordered Dowty Information Systems CSU's, and cisco T-1 interfaces. We expect to have this line running at T-1 by the end of July. 3. We are preparing for the Westnet Steering Committee meeting to be held in Denver, on July 7th. 4. Westnet West has been slowly migrating onto the Internet, thanks to the efforts of Mr. Lee Hollar at the University of Utah. Access appears stable, and trouble free. by Patrick J Burns (pburns@super.org) Westine [Page 19]  Internet Monthly Report JUNE 1988 TASK FORCE REPORTS ------------------ APPLICATIONS -- USER INTERFACE The task force met June 22-23 at Olivetti Research Center in Menlo Park. Primary topics of discussion were progress reports on voice server work; summary reports of the video working group and "shared X" meetings; real-time requirements of voice and video; user interface architecture; and periodic revelations from the Colab project at Xerox PARC. In addition, overviews of relevant work were presented by observers George Champine (DEC/MIT Project Athena), Greg Foster (Xerox PARC), Phil Gust (HP Labs), Dick Phillips (Los Alamos National Laboratory), and Andrew Schulert (On Technology). A more detailed summary will be posted in next month's issue. Keith Lantz (LANTZ@ORC.OLIVETTI.COM) AUTONOMOUS NETWORKS No internet-related progress to report. Deborah Estrin (Estrin@OBERON.USC.EDU) END-TO-END SERVICES IP Multicasting: * Steve Deering published RFC-1054, "Host Extensions for IP Multicasting", in May. This is the up-to-date specification which will soon be proposed as an Internet standard. * Steve also announced the availability of host code to implement RFC-1054, as extensions to the "4.3+" (4/4/88) Berkeley networking code. It is available for anonymous FTP from Stanford. Eric Nordmark ported the code to Sun OS 4.0, and a release will be available shortly from Stanford. For more information, contact: deering@pescadero.stanford.edu. * BBN Labs is making an important contribution to IP multicasting development writing the gateway forwarding code for IP multicast, for Butterfly gateways and for BSD-based gateways. Westine [Page 20]  Internet Monthly Report JUNE 1988 * Steve Deering will present a paper "Multicast Routing in Internetworks and Extended LANs" at the August SIGCOMM'88 meeting. This paper proposes extensions to two common internet routing protocols -- distance-vector routing and link-state routing -- to support low-delay datagram multicasting. VMTP: * The VMTP Unix implementation is being revised to bring it up to specification and extend the implementation. It has been ported to execute within the Sun OS 4.0 kernel, by Eric Nordmark of the Swedish Institute of Computer Science. This revised implementation should be available for distribution by the end of the summer, if not sooner. * Dave Cheriton will present a paper "Exploiting Recursion to Simplify RPC Architectures", which covers the main ideas behind some major simplications to VMTP, at SIGCOMM'88. Performance: * Bill Nowicki of Sun has experimented with the retransmission algorithms of NFS over UDP. A modified application of the Van's TCP timeout scheme is being applied within NFS to provide dynamic adaptation to Internet characteristics, reducing the number of useless retransmissions. Dynamic transfer size adjustment is more difficult, but the prototype implementation adapted to a 9.6K bps line and to the Arpanet in experiments. Bob Braden (Braden@ISI.EDU) INTERNET ARCHITECTURE No activity to report this month. Dave Mills (Mills@HUEY.UDEL.EDU) INTERNET ENGINEERING The Internet Engineering Task Force (IETF) held a three day meeting at the U.S. Naval Academy in Annapolis, Maryland on June 15-17. The meeting was hosted by Terry Slattery. The final agenda, as conducted at the meeting, is given below. The major points of the meeting included: Westine [Page 21]  Internet Monthly Report JUNE 1988 1) Thirteen of the seventeen Working Groups met and reported. The final agenda below gives the thirteen that met at this IETF. The IETF Chair stressed the importance of WG progress and reporting. Working Groups that do not seemed to be making significant progress toward their stated goals or not not report their activities in a timely fashion will be dissolved. A new WG (to be co- chaired by Drew Perkins (CMU), Phil Prindeville (McGill) and Russ Hobby (UC Davis)) was formed to address the issue of a standard Serial Line IP (SLIP). 2) Members of the Federal Research Internet Coordinating Committee (FRICC) attended most portions of the 3 day IETF. The FRICC met during the Working Group sessions to discuss, among other issues, the status of the Research Internet Backbone (RIB). Steve Wolff (NSF) , Mark Pullen (DARPA) and Bill Bostwick (DOE) combined to present details of the FRICC and RIB to the IETF. 3) Two presentations were made which provided an overview of the TCP/IP networking landscape in Canada. John Curley of the National Research Council of Canada (roughly the equivalant of the U.S. NSF) gave an overview of their funding plans. Philip Prindeville of McGill University gave an overview of several networking proposals. 4) Van Jacobson presented an insightful analysis on self- organizing behavior of allegedly random processes in networks. In particular, he presented data showing how DECnet routing updates quickly became synchronized. He used this to supprt his allegation that under certain conditions the Arpanet may be behaving like a giant token ring, leading to degraded performance because of lock- step behavior. 5) Internet Problem Description Forms were given to all WG chairs, for their meetings at the IETF. The completed forms are still being collected. 6) Proceedings of the March IETF in San Diego were distributed in Annapolis and are now available from the NIC. Final Agenda for the June 15-17 IETF ----- ------ --- --- ---- ----- ---- This is the final agenda, as conducted at the meeting, for the June 15-17 IETF at the U.S. Naval Academy in Annapolis, Westine [Page 22]  Internet Monthly Report JUNE 1988 Maryland. There was be a total of 13 Working Group meetings in the first 1.5 days. In addition, there was an NSFnet Technical Group meeting and a FRICC meeting, both on Thursday morning. WEDNESDAY, June 15 9:00 am Opening Plenary (Introductions and local arrangements) 9:30 am Working Group Morning Session o Host Requirements (Braden, ISI) o SNMP (Rose, TWG) o Open Routing (Hinden, BBN) o Open SPF IGP (Petry, UMD and Moy, Proteon) o TELNET Linemode (Dave Borman, Cray) 12:00 Lunch 1:30 pm Working Group Afternoon Session o Host Requirements (Braden, ISI) o Landmark Routing (Tsuchiya, MITRE) o Short-Term Routing (Hedrick, Rutgers) o Open INOC (Case, UTK) 5:00 pm Recess THURSDAY, June 16 9:00 am Opening Plenary 9:15 am Working Group Session o Management Information Base (Partridge, BBN) o Authentication (Schiller, MIT) o PDN Routing (Rokitanski, DFVLR) o Performance and Congestion Control (Mankin, MITRE) o Domains (Mamakos, UMD) o NSFnet Technical Group (Choy, NCAR) o FRICC (Bostwick, DOE) 11:30 am Lunch 1:00 pm Opening Plenary Statement (Gross, MITRE) 1:15 pm Network Status Reports Westine [Page 23]  Internet Monthly Report JUNE 1988 o Status of the New NSFnet (Braun, UMich/Rekhter, IBM) o FRICC Initiatives (Bostwick, DOE/Pullen, DARPA/Wolff, NSF) This talk presents an overview of the networking initiatives supported by the Federal Research Internet Coordinating Committee (FRICC). In particular, the Research Internet Backbone (RIB) will be discussed. o Arpanet Report (Lepp, BBN) o Canadian Research Networking (Curley, NRC of Canada) The National Research Council of Canada has been developing a plan for an IP-based network for Canadian researchers. This report will give an overview of this and other networking activities in Canada. o Switched Multi-Megabit Data Service (SMDS) (Kramer and Singh, NYNEX) This talk will focus on work currently underway on the development broadband Public networks, highspeed data communications and the evolving underlying technologies to support such services. In addition to SMDS, Broadband ISDN and FDDI will be discussed. 5:00 pm Recess FRIDAY, June 17 9:00 am Internet Report (Brescia, BBN) 9:30 am Working Group Reports and Discussion o Authentication o CMIS-based Network Managament o Domains o Internet Host Requirements o Internet Management Information Base o Open SPF-based IGP o Open Systems Internet Operations Ctr o Open Systems Routing Westine [Page 24]  Internet Monthly Report JUNE 1988 o PDN Routing Group o Performance and Congestion Control 12:00 Lunch 1:00 pm Working Group Reports and Discussion (cont.) o Short Term Routing o SNMP Extensions o TELNET Linemode 1:30 pm Technical Presentations o TCP Performance and Other Unconfirmed Rumors (Van Jacobson, LBL) ``Phill, ...self-organization... synchronization...Smolunchowski's equation... anyway, if you want to hear these wild, unsubstantiated theories, I'll be glad to talk about them.'' o Bellringing, Clock Punching and Gongferming (Mills, UDel) The latest on the Network Time Protocol (NTP). o Cray TCP Performance, An Update (Borman, Cray) Would you believe a 150 Mbps TCP? o Issues in Canadian Networking (Prindeville, McGill) An overview of a proposed Canadian Internet Architecture with special attention to specific problems (eg, routing metrics under multi-protocolsupport and multi-gateway reachibility. 4:45 pm Concluding Plenary Remarks 5:00 pm Adjourn Phill Gross (gross@gateway.mitre.org) Westine [Page 25]  Internet Monthly Report JUNE 1988 INTERNET MANAGEMENT No report received. Vint Cerf (Cerf@A.ISI.EDU) PRIVACY The IAB Privacy Task Force had a productive two-day meeting at Digital Equipment Corporation's facility in Littleton, Massachusetts, 15-16 June 1988. Attendees were: Dave Balenson, Curt Barker, Don Brinkley, Morrie Gasser, Russ Housley, Steve Kent, John Linn, Dan Nessett, Mike Padlipsky, Rob Shirey, and Steve Wilbur. Jim Bidzos, of RSA Data Security, Inc., attended the 15 June session. Russ Housley, of Xerox Special Information Systems, was a new member as of this meeting. RFC-1040 implementation activities were reviewed, including the successful exchange of encrypted messages between the National Bureau of Standards and University College London. A partial and preliminary draft (distributed to task force members in advance of the meeting) for a key management RFC to complement the privacy-enhanced electronic mail processing procedures defined in RFC-1040 was discussed. The roles of users, notaries, ordering representatives, and of RSA Data Security in its planned capacity as certificate issuing authority were considered, and will be noted in subsequent revisions to the key management RFC. The next PTF meeting was scheduled for the afternoon of 13 September through 15 September at Xerox Special Information Systems, Arlington, VA. John Linn (Linn@CCY.BBN.COM) ROBUSTNESS AND SURVIVABILITY No report received. Westine [Page 26]  Internet Monthly Report JUNE 1988 SCIENTIFIC COMPUTING No report received. TACTICAL INTERNET No report received. TESTING AND EVALUATION No report received.