[IMR] IMR88-07.TXT July 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). BOLT BERANEK AND NEWMAN INC. ---------------------------- VAX NETWORKING We now have a multicast agent up and running under 4.3bsd and are developing various applications to experiment with wide area multicasting. WIDEBAND NETWORK The Wideband Net IST links which carry Arpanet traffic functioned smoothly during July. Soon, the capacity of these trunks will be increased moderately. This change should improve Arpanet service and, especially, increase the capacity of remaining Wideband IST lines to carry traffic if one of the Wideband IST sites experiences trouble. Westine [Page 1]  Internet Monthly Report July 1988 A multi media/video conference was held between ARPA, ISI, and BBN during July using the Wideband Net. The conference demonstrated the usefulness of the broadcast and stream facilities of the Net. In order to support the heavy use of the Net by multi-site (eventually 4 sites) video conferences and Arpanet traffic, the BSAT software will be reconfigured to run multiple downlink processes. Remote monitoring has been over hauled in the BSAT. We improved the ability to monitor different versions of BSAT host and channel modules and to quickly change the information reported by the BSATs. INTERNET RESEARCH & DEVELOPMENT The Butterfly gateway Telnet X.25 interfaces was certified by Telenet. We are now working to begin installing the Butterfly replacements for the LSI-11 VAN gateways. We have begun to install a new release of Butterfly Gateway software (Release 4.1) which includes a number of improvements. The most important are in the area of remote monitoring and control. The facilities in this area include: Throughput Collection Host Traffic Matrix Collection Buffer Statistics Routing Information Interface Control (Loop, Crosspatch, Enable/Disable, Test) Routing Updates Trap Statistics The release also includes an extensible access control and load sharing mechanisms. SATNET July has been a very busy month for SATNET. We experienced several outages caused by SIMP crashes and related loading problems. We are still working on the causes of the crashes but it is now felt they were related to higher than normal noise on the satellite channels. We are currently working with the personnel at Roaring Creek to resolve this problem. The new international RSRE to BBN direct link has been ordered. Installation and testing are planned for the first of October. Westine [Page 2]  Internet Monthly Report July 1988 The Tanum earth station antenna is scheduled to be repointed to a different satellite on August 20. This will remove them from the SATNET. Work is now underway to provide a direct link from the NTA gateway to the RSRE gateway. This will be needed to provide INTERNET connectivity for NTA-RE. Bob Hinden (Hinden@BBN.COM) ISI --- Internet Concepts Project Jon Postel attended the IAB meeting in Santa Fe, NM, Jul 11-14. Two RFCs and one paper were published this month. RFC 1059: Mills, D., "Network Time protocol (Version 1) Specification and Implementation", University of Delaware, July 1988. RFC 1063: Mogul, J., C. Kent (DEC), C. Partridge (BBN), K. McCloghrie (TWG), "IP MTU Discovery Options", July 1988 RFC 1064: Crispin, M., "Interactive Mail Access Protocol - Version 2", SUMEX-AIM, July 1988. Mockapetris, P., "Development of the Domain Name System", ACM SIGCOMM 88, August 1988. Ann Westine (Westine.ISI.EDU) Multimedia Conferencing Project The multimedia teleconferencing system made its official debut in a conference of more than two sites this month. The Autonomous Networks Task Force used it during a two day meeting among members located at DARPA, ISI, and BBN. All three sites were displayed simultaneously in quadrants of the video screen. At each site, voice from the remote sites was mixed for playback, allowing all sites to talk at once if they wished. The Diamond/MMCONF system was used for shared document display and editing among the three sites. It was felt that the three site conference was effective, but that participants needed to concentrate more, resulting in a more tiring session than in previous teleconferences with two sites. A Westine [Page 3]  Internet Monthly Report July 1988 contributing factor may have been that this meeting included the largest total number of participants of any so far. Two mechanisms were suggested to help with this problem: 1) the ability to zoom in, audibly and/or visually, to help focus the conference during discussions; and 2) a tool running on the workstation showing a queue of people who would like to speak, as an improvement over hands held in the air, as a guide to the chairman. Video zooming is possible now, but may not be easy enough to use except by a non-participant operator. At least one participant was quoted as saying, "A good time was had by all". Eve Schooler, Brian Hung, Steve Casner, Dave Walden (schooler@ISI.EDU, hung@ISI.EDU, casner@ISI.EDU, djwalden@ISI.EDU) NSFNET Project Annette DeSchon nearly completed "bftptool", a window-based user interface to BFTP for use in submitting background file transfer requests. This interface runs on a Sun Workstation. A draft RFC on BFTP was also completed. Bob Braden continued work on editing the Host Requirements RFC, being prepared by an IETF Working Group. This RFC has grown to 124 pages in length and now includes a Table of Contents and a "Checklist". Sections on Telnet and trailer negotiation were added this month. Several topics required considerable effort and discussion, especially Telnet end-of-line and TCP SWS avoidance. There is an ongoing discussion in the Working Group of how Multihoming should be handled. Annette DeSchon attended a meeting of the Autonomous Networks Task Force which was teleconferenced between sites at ISI, BBN, and DARPA, Jul 7-8. Bob Braden attended a meeting of the IAB in Santa Fe, NM, Jul 11-14. Bob Braden and Annette DeSchon (Braden@ISI.EDU, DeSchon@ISI.EDU) Supercomputer and Workstation Communication Project Alan Katz started to test the split remote editor and has been working on which runs under GNU Emacs. The editor is written in emacs lisp. Alan Katz (Katz@ISI.EDU) Westine [Page 4]  Internet Monthly Report July 1988 MIT-LCS ------- No report received. MITRE Corporation ----------------- No report received. NTA-RE and NDRE --------------- No report received. SRI --- No internet related activity to report. Zaw Sing Su (zsu@tsca.istc.sri.com) UCL --- Testing of the Campus FDDI net continued. We ran a range of protocol suites, including DECNET, TCP/IP, Appletalk, OSI and Pink Book (X.25 on a LAN/MAN) simulataneously to see if any complex interactions caused problems. The proliferation of broadcasts caused the only real difficulty. Implementation work started on a hidden layer neural network model of distributed routing and congestion control algorithms. The current reinforcement rule is based on a Boltzmann machine principal. John Crowcroft (jon@CS.UCL.AC.UK) UNIVERSITY OF DELAWARE ---------------------- 1. Jeff Simpson is working on policy-based routing algorithms and related technology. He is currently compiling a report and classification of current policies (real or ad-hoc) used in the Internet. Additionally, he is reviewing the policy implementation mechanisms being proposed by various groups/factions. He is also experimenting with various policy Westine [Page 5]  Internet Monthly Report July 1988 restricitions and evaluating the effects on routing performance of both real and random networks. 2. Paul Schragger continues to work on feedback-control and rate-based concepts applied to transport-level communications. He is currently exploring a simple non-linear network model involving rate-based feedback from intermediate and end systems. He is using the MIT network simulator to evaluate various source-control schemes required for stable operation of the transport entity with a nonlinear network model. Work on the simlator itself continues, as well as feedback to MIT on Sun code fixes. 3. The Network Time Protocol (NTP) has been declared an Elective Protocol with RFC-1059 as Draft Internet Standard (DIS) specification document. Interested chimers should tap the ntp@trantor.umd.edu list maintained by Mike Petry, who also has a Unix 4.3bsd daemon conforming to the specification. 4. Five of the original six NSFNET Backbone fuzzballs are back on the air, one (UIUC) with a new WWVB radio clock. Hopefully, the remaining holdout (SDSC) will be back soon with a refurbished WWVB radio clock. Cornell is being lobbied to wiretap an existing GOES satellite clock being used to timestamp seismic data. Together with several other existing and proposed WWV radio clocks, this makes a total of at least nine high-accuracy (WWVB and GOES) and four medium-accuracy (WWV) NTP primary time servers available on the Internet and providing service to over a hundred users. In addition, the new NSFNET Backbone NSS switch components are internally time-synchronized using NTP, but not providing external time service. 5. The U Delaware fuzzballs, including NTP servers, have been rehomed to our research network 128.4, which is gatewayed to ARPANET and SURANET via the campus fiber. Exhaustive testing revealed previously unsuspected low-level packet leaks (about 0.7 percent) in the maze of Proteon routers stitching the campus and SURANET together. Similar leaks involving Proteon routers and serial lines were then found at PSC. There is as yet insufficient data to resolve azimuth of fingerpoint. While at it, several experiments involving the new NSFNET Backbone and random fuzzballs revealed interesting delay anomalies yet to be explained. 6. Dave Mills attended the IAB meeting in Santa Fe. The final report covering issues in reliable telecommunication network synchronization was submitted to the National Academy of Westine [Page 6]  Internet Monthly Report July 1988 Sciences and is now being reformatted as a Department report. Work continues on a draft report discussing issues in the realization of a nationwide gigabit network and, in particular, on channel-access schemes suitable for early deployment and evaluation. Other work in progress includes a paper on DGP in collaboration with Mike Little of SAIC. Dave Mills (Mills@UDEL.EDU) Westine [Page 7]  Internet Monthly Report July 1988 NSF NETWORKING -------------- NSF NETWORKING UCAR/BBN SYSTEMS AND TECHNOLOGIES CORPORATION, NNSC RFC 1063, IP MTU Discovery Options, authored by J.Mogul, C.Kent, C.Partridge and K.McCloghrie is available from the Network Information Center in the directory at SRI- NIC.ARPA. NEW NSFNET BACKBONE (MERIT) The NSFNET backbone is now a production network. By the beginning of July, the thirteen nodes comprising the backbone were announcing routing information to their regionals, making it possible for user data to flow across the network. The AT&T circuits have been shut off and the backbone is now entirely linked by MCI T1 lines. All of the old NSFNET backbone fuzzballs are turned off. Final tuning is under way to maximize speed and reliability as of July 20. Reaching production stage has been a gradual process. Each regional site established a schedule for moving its campus and institutional networks from the old backbone to the new one. Most sites started with one or two local networks of "friendly users." As their confidence grew, each site added more traffic by increasing the number of subnetworks routed through the backbone . Currently, over 200 networks are able to pass traffic on the new NSFNET backbone. The number of academic and research networks able to send traffic through the NSFNET backbone is increasing by an average of two new networks per day. Significant work has been done to improve the speed of the new backbone over the old. In addition, hardware upgrades are now being developed which are expected to further increase network speed. Westine [Page 8]  Internet Monthly Report July 1988 Reaching production stage has required multiple steps and the cooperation of many individuals. We are very excited by the progress of this project. None of these achievements would have been possible without the cooperation at all levels and across all areas of the NSFNET endeavor. We would like to extend a special thanks to all the regional site people for their hard work and extra efforts to make this project a success. by Laura Kelleher (lkelleher@merit.edu) NSFNET BACKBONE (CORNELL UNIVERSITY THEORY CENTER) No report received. NSFNET BACKBONE SITES & MID-LEVEL NETWORK SITES UNIVERSITY OF ILLINOIS AT URBANA-CHAMPAIGN/NSCAnet Cutover to the NSSnet from FuzzNet was completed July 20 when uiuc.nsf.net was turned off at 12:00 CDT. The new NSS is handling an immense amount of traffic, about six times the peak fuzz figures. Eight additional Minnesota networks, on hold since April, are now being EGP'ed to the NSS. A total of 43 nets are being announced at EGP metric 0 and another 18 at EGP metric 3. Copies of our gated.conf files and PostScript images of UIUC's network topology are available for anonymous FTP from uxc.cso.uiuc.edu:pub/net. Concomitant with the increase in traffic is frequent exhaustion of virtual circuits with the ACC-6250 X.25 link to PSN 94. ACC says that the 128 circuit fix will be available "in a month or two". Ross Veach has been working on adding "long-distance" EGP features to gated. The goal is to spread routing knowledge selectively through a regional so that bad routing decisions can be avoided. For UIUC and Minnesota, each has an ARPANET gateway that's to be preferred for net-10 and all other nets not carried on the NSSnet. Minnesota traffic for NSSnet advertised nets is sent to UIUC. By having the two gated programs at UIUC and Minnesota EGP peer with each other, the regional nets in between don't have to propagate 400+ nets via RIP. Using RIP would also capture all of Minnesota's Arpanet bound traffic and send it to UIUC as IGP routes are favored by gated over EGP routes. Westine [Page 9]  Internet Monthly Report July 1988 The fuzzball has been re-born as truechimer.cso.uiuc.edu at 192.17.174.40 providing time of day, standard time intervals and other related NTP information. It ticks at stratum 2 until mid-August when the Spectracom 8170 WWVB clock will be installed. by Paul Pomes (paul@uxc.cso.uiuc.edu) JOHN VON NEUMANN NATIONAL SUPERCOMPUTER CENTER The JVNCNet monthly report has been reduced in size to comply with requests from some readers. The full report (about 12 pages long) can be obtained from the JVNCnet NIC. Monthly Status Overview: * The overall uptime for the gateways (33 gateways) and links (T1s, 6kbps and satellite) this month was 91.68% (worst case, this number considers that all gateways are unreachable when JVNCA is down the measured uptime when JVNCA was available (97.2% of the time) was for an average on all the gateways 94.76% available. This represents a lower uptime than last month (96% over this months'91.68%). This increase in down time was primarily due to a number of routing loops and electric storms that caused interruption of the services at JVNC. * Traffic on the JVNCA gateway has been slightly higher this month than last month, with a total number of packets in and out (of ONE of its ethernet interfaces) of 125,248,680 packets. Plans are moving ahead for the connection of the JANET (Joint Academic Network) in the UK to the NSFNet at JVNC. * Our "new" NSFNet node is up and running. The NSS is on an ethernet segment connected to two CISCO routers on the JVNC internal network. These routers are exchanging EGP information with the NSS and the PSN, and provide to be the most intelligent routers on our system. * Number of networks being announced from the JVNC-arpa gateway: with egpmetric 0 = 15; and with egpmetric of 3 = 18. Westine [Page 10]  Internet Monthly Report July 1988 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 For more information contact: Network Operations: phone: (609) 520-2000 e-mail: JVNCnet-noc@jvnca.csc.org hours: Primary: 9am-5pm Monday to Friday Secondary: 5pm-9am Monday to Friday, 24hrs, weekends and holidays Network Informations: phone: (609) 520-2000 e-mail: JVNCnet-nic@jvnca.csc.org hours: 9am-5pm Monday to Friday by Sergio Heker (heker@jvnca.csc.org) NATIONAL CENTER FOR ATMOSPHERIC RESEARCH AND UNIVERSITY SATELLITE NETWORK PROJECT USAN now connects to the Internet via the NSS and most routing on USAN is now dynamic, with RIP the IGP. The WESTNET-EAST cisco gateway that is on the USAN backbone EGP peers with the NSS, and windom, a Sun 3/280 with only one ethernet interface, serves as a gated agent peering with the NSS for the USAN AS. Westine [Page 11]  Internet Monthly Report July 1988 The fuzzball has been disconnected from the serial lines but still is in operation as a timeserver for WWVB clock information. Some problems still exist in exchanging routing information with the fuzzball and windom so not all sites can yet reference the clock by Don Morris, morris@windom.ucar.edu PITTSBURGH SUPERCOMPUTING CENTER I am the new gateway/IP network engineer at PSC. I will be working with Gene Hastings and Dave O'leary installing and maintaining PSC's IP network facilities. Since we have missed the last several reports, and this is my first, this report reflects changes back to an indefinite beginning. Our machine room has been completely overhauled. The NSFnet NSS required additional floor space, air conditioning and power beyond that already available. We took the opportunity to completely recable the machine room as well. Our routing stub was converted from subnet 1 of 128.182 to class C network 192.5.146. The machines of interest are: psc-gw1.psc.edu (192.5.146.1) microvax gated agent with: PSN14 (10.4.0.14) and PREPnet (129.250) psc-gw2 (192.5.146.2) microvax gated agent with (FUTURE): T1 to SURANET and 2nd Arpanet connection. psc-gw3 (192.5.146.3) microvax gated agent with: Carnegie Mellon (128.2) and the NSS egp peer. mi-rtr (192.5.146.6) Our "internal" Proteon with remote links to: SURANET (slated to be moved to psc-gw2) UPghRTR (192.5.146.7) Proteon 4200 connecting to Univ. of Pittsburgh. fuzzball (192.5.146.42) Keeping time at a new address. NSS-5 (192.5.146.254) PREPnet is the new Pennsylvania regional network. It came online mid June, connecting PSC, University of Pittsburgh, Carnegie Mellon University, University of Pennsylvania, Drexel, Temple, and Lehigh. We also installed "virgin" Mt Xinu on our 3 microvax gated gateways. One of our design goals was to make the gateways as symmetric as possible: except for 4 files they are identical. Our arpanet gateway has been having moderate stability problems, panics with "mbuf map full". It seems to be related to the large routing tables rather than the acc driver. There have been a few reboots on the other gateways attributed to the same problem. Westine [Page 12]  Internet Monthly Report July 1988 During June The CMU/TEK TCP/IP on the VAX/VMS front ends to the Cray XMP became unusable. We have not determined if this was due to upgrading the front ends from VAX 8650's to 8810's or due to changes in the backbone. As a work around, we are using DECnet- ULTRIX DECnet-Internet Gateway software running on some microvax II's to convert TCP/IP services to DECnet. We are also testing Wollongong, SRI MultiNet, and other TCP/IP implementations for the front ends. Both SRI and Wollongong have passed our initial robustness tests. Testing with real users is scheduled to begin later this week. Some of the other products are still under consideration. Our direct T1 to SURANET net will be moved from our internal Proteon to its own dedicated routing stub with a gated agent. This change has been imminent, but keeps getting snagged on various problems. This connection provides backup for a large class of NSS/local routing stub failures and carries direct traffic to some of our heaviest users. When we resolve the gateway stability problem we plan to act as an additional NSF/ARPA net gateway. by Matt Mathis (mathis@faraday.ece.cmu.edu) SAN DIEGO SUPERCOMPUTER CENTER NSFnet II is operational at SDSC. All traffic that was routing through our fuzzball is now flowing over the Merit/MCI/IBM system. Our Fuzzball was turned off for about a week - it has now been resuscitated to act as a clock for Dave Mills. Our ARPAnet PSN has a second operational trunk. We now have trunk links to ISI and UCLA. OPUS, our SUN 3/50 running gated, is now almost out of the packet forwarding game. Only a couple of minor systems still route their packets through it. Our p4200 (and our VAX's) peer with OPUS via EGP to collect NSFnet routing information. Thus our p4200 connected nets get direct service from the p4200 to the NSFnet NSS. Our Proteon p4200 is now connected to (just) U Hawaii, UC Santa Barbara, UC San Diego and the Salk Inst. We are expecting a cisco systems router in the next 2 weeks. It will be used as an evaluation unit for CERFnet's routers. by Paul Love (loveep@sds.sdsc.edu) Westine [Page 13]  Internet Monthly Report July 1988 BARRNET No report received. CERFNET (California Education and Research Fed. Network - Initial Status Report) The California Education and Research Federation (CERF) has submitted a proposal to the National Science Foundation to establish a regional network in California. This network, CERFnet, will be established to enhance both supercomputer access and the environment for academic computing in the region. CERFnet will have a mesh architecture with a high speed backbone through the most heavily used paths. There is a mixture of 56 kilobit, 512 kilobit, and 1.544 megabit lines. The initial members are listed below. (All lines are T1 unless otherwise indicated). -The academic institutions actively participating are: The Agouron Institute California Institute of Technology The California State University system including (all 56k): Chancellor's Office California State University campuses at: - Bakersfield - Los Angeles - Chico - Northridge - Dominguez Hills - Sacramento - Fresno - San Bernardino - Fullerton - San Francisco - Hayward - San Jose - Humboldt - Sonoma - Long Beach - Stanislaus San Diego State University The California Polytechnic State Univ. campuses at: - Pomona - San Luis Obispo Information Sciences Institute Occidental College (56k) Research Institute of Scripps Clinic Salk Institute of Biological Studies San Diego Supercomputer Center University of California Office of the President University of California campuses at: - Irvine - San Diego - Los Angeles - Santa Barbara (512k) - Riverside (56k) Westine [Page 14]  Internet Monthly Report July 1988 University of San Diego (56k) University of Southern California -The industries actively participating are: Northrop Research and Technology Center Science Applications International Corporation (56k) There are a couple of interesting features of CERFnet. First, some members of CERFnet are also members of Los Nettos. Los Nettos is a regional network locaLos Angeles. An important feature of Los Nettos is that the circuits and gateway systems be high speed to encourage research that requires high bandwidths. Currently, all connections are expected to be at least 1.544 megabit capacity. Los Nettos will be a member of CERF in its own right, in addition to individual institution memberships. Second, the California State University system has an existing state-wide X.25 network. This will be upgraded in its internal bandwidth and a TCP/IP gateway added at each campus. This will provide each campus in the system with full TCP/IP service. Third, CERF has asked NSF to provide a separate dedicated link into the NSFnet backbone at ISI. This link is proposed to have equal capacity to those of the other directly connected regionals. This link has been requested because of the large amount of backbone that is anticipated from the Los Angeles Basin institutions. (Additionally, the CSU X.25 network will interconnect at ISI.) At this time, there are several devices that CERFnet is evaluating for the gateway/routers. Because there are so many variations in the network, it may be possible that more than one vendor's devices will be used in the network. At this time, cisco Systems is the only company that can provide the diversity of hardware that will be required in CERFnet and provides a superior, although proprietary, routing protocol. In order to facilitate CERFnet's access to the ARPAnet, an additional DDN X.25 card will be installed in a cisco gateway. Initially this will be at the San Diego Supercomputer Center, enabling a simple connection to the existing ARPAnet Packet Switching Node located at SDSC. It may be moved to ISI at a latter date or a second card installed there. A test network using the cisco X.25 & DDN/X.25 capabilities went up during the month of August between the California State University Chancellor's Office, Cal Poly San Luis Obispo, Cal State Fresno, Cal State Northridge and the San Diego Supercomputer Center. It will be extended to Los Nettos when possible. Westine [Page 15]  Internet Monthly Report July 1988 by Susan Estrada (estradas@luac.sdsc.edu) MERIT/UMNET The IP gateway box for the MTS mainframe operating system, of which several are attached to Merit, is in provisional release to users and is proving fairly reliable. MIDNET No report received. MRNET No report received. NORTHWESTNET Our connection to the new NSFnet backbone is now in production and working quite reliably. Users have experienced substantially improved service, except in access to the ARPANET. The University of British Columbia is now online connected through a NorthWestNet router to the UW campus network, and from there to NSFnet. by JQ Johnson (jqj@hogg.cc.uoregon.edu) Westine [Page 16]  Internet Monthly Report July 1988 NYSERNET Clarkson Albany | | | | | | Rochester ========== Cornell ======= Syracuse ====== Rensselaer || || / || || || / || || || / || || || NyserNet || || ..... Oswego || NISC || Buffalo ... Alfred || || | || || | || || | || || | || || | || || | Stonybrook ____ Columbia ====================== N Y U ===:\ | / | | __/ | || | Brookhaven ____/ / | ___/ | || | / | / | || | AOA === NYNEX S+T NSMAC _____ C U N Y _______ Poly || | || | \ || | || | \ || | || | \ || | Compass -------- Garden City | \_____ Rockefeller......SKF | | || | | || | | || Binghamton _______________________________| White Plains CO _________________________________________________________ | | | KEY: | | | | Line Speed Representated 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 July 1988 JULY EVENTS: - EGP peering and traffic flow with the new NSFNet - release of the 2.2 version of NYSERNet's SGMP - installation of NYSERNet's ARPANET gateway - release of the 3.0alpha version of NYSERNet's SNMP - completion of the ANSI Z39.50-1988 ASN.1 encoder/decoder - addition of a new member: SmithKleinFrench by Martin Lee Schoffstall (schoff@nisc.nyser.net) No report received. OARNET No report received. SESQUINET No report received. SURANET The following SURAnet sites are presently on-line: University of Alabama at Birmingham Alabama Supercomputer Network University of Alabama at Tuscaloosa Catholic University of America Clemson University Continuous Electron Beam Accelerator Facility University of Delaware Department of Energy/Oak Ridge Operations Office University of Florida Florida State University Fox Chase Cancer Center Emory University Gallaudet University George Mason University Georgetown University George Washington University Georgia Institute of Technology University of Georgia Johns Hopkins University University of Kentucky Louisiana State University University of Maryland NASA/Goddard National Bureau Of Standards Westine [Page 18]  Internet Monthly Report July 1988 National Cancer Institute/Frederick Cancer Research Center National Institute of Environmental Health Sciences National Institutes of Health National Radio Astronomy Observatory National Science Foundation Naval Research Laboratory Oak Ridge National Laboratory Old Dominion University Supercomputer Research Center (IDA) University of Tennessee Triangle Universities Computation Center Duke University North Carolina State University University of North Carolina Vanderbilt University Virginia Commonwealth University US Geological Survey University of Virginia Virginia Polytechnic Institute University of West Virginia College of William & Mary It is anticipated that Mississippi State University will be added to the Network within several days. The following networks are being advertised from SURAnet to the NSFnet: 128.4 DCN 128.173 VaTech 192.5.39 UDEL 128.8 UMD 128.175 UDEL 192.5.45 FCCC 128.60 NRL 128.186 FSU 192.5.57 UDEL-CC 128.61 GeoTech 128.192 UGA 192.5.214 DEC 128.82 ODU 128.220 JHU 192.5.82 FSU 128.109 TUCC 128.227 UFL 192.5.215 GMU 128.140 Emory 128.239 Wm & Mary 192.5.219 Clemson 128.143 Virginia 129.6 NBS 192.16.75 Georgetown 128.150 NSF 129.57 CEBAF 192.16.176 LSU 128.163 UKY 129.59 Vanderbilt 192.26.10 Gallaudet 128.164 GWU 129.66 UAB 192.31.192 IDA/SRC 128.167 SURA 129.71 WV net 192.31.193 Catholic 128.169 TENN 129.143 NCIFCRC 192.33.115 NRAO 128.172 VCU 130.39 LSU 192.41.177 SURA Westine [Page 19]  Internet Monthly Report July 1988 SURAnet is presently using GATED to exchange routing information with the NSFnet backbone. We have obtained an early release of the Proteon p4200 gateway software Version 8.1. When this is running GATED will not be required. We plan to test Version 8.1 within the next few days. by Jack Hahn (hahn@umdc.umd.edu) WESTNET 1. Both Westnet "East" and Westnet "West" have been routing traffic across the new NSFnet T-1 Backbone, and have observed a marked improvement in performance, compared to the old 56 kbps backbone. 2. A meeting of the Westnet Steering Committee was held in Denver on 7 July 1988. A variety of topics were covered, with the greatest time being spent on the establishment of a proper NIC and NOC for Westnet. It was announced that Ms. Carol Ward will be assuming a full-time position under the direction of David Wood at the University of Colorado at Boulder to perform NIC and NOC functions, and provide other technical support for Westnet. 3. The 56 kbps circuit between the cisco Gateways at NCAR and the University of Colorado at Boulder, experiencing significant congestion during the day, was ordered to be upgraded to T-1 in early August. 4. Pat Burns, the PI of Westnet, is now on sabbatical until July 31, 1989. He is still continuing to function in his Westnet capacity as part of his formal duties during his sabbatical. He can be contacted until August 1 1989 at: Pat Burns Supercomputing Research Center 4380 Forbes Blvd. Lanham, MD 20706 e-mail: pburns@super.org phone: 301/731-4910 by Pat Burns (pburns@super.org) Westine [Page 20]  Internet Monthly Report July 1988 TASK FORCE REPORTS ------------------ APPLICATIONS -- USER INTERFACE User Interface Task Force: Summary of June 1988 Meeting The task force met June 22-23 at Olivetti Research Center in Menlo Park. The major topics of discussion were group work, user interface architecture, video, and real-time requirements of voice and video. The resulting highlights were: - Andy Schulert of On Technology (and formerly principal architect of Open Dialogue at Apollo) added his support to the argument that, in the world of "network computing", "user interface toolkits" would be better placed on the same machine with the window server. This effectively means encapsulating them in a separate server (as is being doing at Olivetti, for example) or in the window server itself (as several groups are doing with NeWS). The principal motivation is the same as that for NeWS and Stanford's VGTS before it, namely, to reduce network traffic. A possible side effect is that it may even be possible for the client- server protocol to be strictly synchronous without employing batching or pipelining, as used by X of the VGTS, for example. On the other hand, this eliminates the parallelism that can be achieved by using several workstations. Indeed, experience with the V-System at Stanford and with NeWS suggests that users can often get better response by moving even very low level processing onto remote workstations. In short, the jury is still out. - Greg Foster of Xerox PARC noted that the Colab group is moving away from "electronic meeting rooms" per se to the more general concept of "shared workspaces". The idea is to provide support for the sorts of serendipitous meetings that occur throughout the day---in the lounge, by the coffee pot, and in people's offices---rather than require people to pick up all their notes and move to a dedicated room to have a meeting. (Phil Gust of HP Labs later added another view of this: "seamless migration from one locale to another.") Moreover, much to the delight of the editor, Greg offered that the Colab room itself is used primarily for communal access to existing applications (as displayed on the 8' rear-projection screen). While this does not mean that special-purpose "groupware" is useless---indeed, the idea generator Cognoter appears to be used regularly by several different groups---it lends yet more support on the argument Westine [Page 21]  Internet Monthly Report July 1988 that there is much to be gained by providing multi-user access to existing tools. - To summarize the discussion of real-time issues: Reasonable-quality voice requires 8000 bytes/sec throughput (note that this is "voice", not high-fidelity audio). Unfortunately, on generic (UNIX-based) workstations, it is not reasonable to generate an interrupt on every byte (sample); rather, it is necessary to provide on-board packetization (or employ multiple processors). As many companies have discovered, most UNIX-based systems are also incapable of delivering the necessary response to storage and retrieval requests; the level of response determines the amount of buffering that is required in the audio server. As for video, T1- (also CDI-) quality implies 200 Kbytes/sec throughput. As with voice, this presents no problem bandwidth-wise, but rather with regard to response (to request to store or retrieve individual frames). However, true broadcast TV requires roughly 60 Mb/s and studio quality requires 140 Mb/s bandwidth, either of which present significant problems for UNIX. - Also with respect to video, Steve Casner reported on the Video Working Group Meeting held May 10 at the MIT Media Lab. The focus of the meeting was on "video in workstations" -- what's the state of the art, and how will it be used. It's too early to define the taxonomy of uses for video (and advantages over other media). Participants in the meeting are working individually on video display in a window on the workstation, video bandwidth compression, transmission across packet LANs and WANs, and integration of video with other media. However, no group has yet brought all these pieces together to form, for example, a multimedia workstation conferencing system. For video to be used as other media are now, there is a strong need for authoring tools. People have grown accustomed to high-quality production in video, making the task even more difficult. The need is equally strong for reading tools; structured multimedia documents will have dual control, both by the author defining how the pieces form the whole and by the reader selecting the parts of interest (e.g., text to explain an event in the video or a video segment to illustrate an example in text). Full minutes of the meeting will be posted to UI-INTEREST. Westine [Page 22]  Internet Monthly Report July 1988 - As for voice server work, projects continue at the Sun New Media Workstation Group to create a full-function PC-based server using the TI speech board and at Olivetti Research Center on a software "audio server" that does for audio what window servers do for graphics (and then some---e.g. computer-controlled mixing and routing). The design specification on the latter will be available shortly. Complete minutes will be posted to UI-INTEREST momentarily. The next meeting will be September 12-13 at ANSA Headquarters in Cambridge, England. Keith Lantz (LANTZ@ORC.OLIVETTI.COM) AUTONOMOUS NETWORKS The Autonomous Networks Task Force held a two-day three-way video conference July 7-8 !!! The three sites were ISI in L.A., BBN in Cambridge, and DARPA in Washington D.C. The meeting began with a report on Barry Leiner's June workshop (Leiner brought together government agency representatives from the FRICC and members of the technical community to discuss policy drivers for new internet protocol mechanisms.) We analyzed the draft policy statements articulated at the Leiner workshop and determined that there was much more work to be done to produce statements of policy that could be used to drive the design and selection of appropriate technical mechanisms. This turned into one of the three action items that members of the task force will pursue during the next few months. In particular, we are drafting a set of detailed questions that must be addressed in every autonomous region's policy statement(s). With these questions in hand we plan to meet individually with members of the FRICC to elicit their policy requirements for usage and interconnect of research network facilities. With the discussions of policy requirements as context we switched our focus to the discussion of several proposed, policy routing mechanisms; in particular, we heard presentations on several schemes developed by members of the IETF Open Routing Group (Nakasis, Callon, Ciappa/Gardner/Zhang), Paul Tsuchiya's Landmark Routing, and Dave Clark's Policy Routing proposal. We found it difficult to 1) evaluate each scheme in terms of its ability to support interconnected autonomous networks, and 2) to compare the schemes to one another. In order to facilitate the first point Westine [Page 23]  Internet Monthly Report July 1988 we began generating a list of generic scenarios/cases that a rich policy routing mechanism should be able to support. A second activity of ANTF members during this work period will be to consolidate/expand/refine the list. To facilitate the comparison of alternative schemes we generated a third action item: to coordinate the preparation of brief summaries of each of the proposed schemes and attempt to make the terminology used more uniform. Our next meeting will be held sometime before the end of the year to discuss all three items mentioned above: refined policy statements, list of generic test cases, and summaries of proposed mechanisms. A note on the medium: I think we were all somewhat surprised at how successful the three-way video conferencing medium was! Personally, I want to thank the support personnel from BBN and ISI for their very successful efforts on our behalf. I hope that we will have the opportunity to use this medium on a regular basis. If any of the other task forces or working groups are planning a three-way meeting of this sort I would be happy to offer my two cents of advice on what I found was helpful in the conduct of the meeting (or would have been had a realized it earlier). Deborah Estrin (Estrin@OBERON.USC.EDU) END-TO-END SERVICES At its July meeting, the IAB created a formal step in the process of defining a new Internet protocol standard: a "Draft Internet Standard". This is a proposal that has undergone extensive review and is expected to become a standard, but requires some period of widespread experimentation before full standardhood is granted. The Draft Internet Standard category and the ISO DIS category are analogous. In a separate action, the IAB agreed to designate the IP Host Group Multicasting specification as described in RFC-1054 as a Draft Internet Standard. Bob Braden (Braden@ISI.EDU) Westine [Page 24]  Internet Monthly Report July 1988 INTERNET ARCHITECTURE The task force will be sponsoring an Internet research workshop in conjunction with the IAB meeting in January. The workshop will include invited presentations by research contributors from throughout the Internet community and especially the IAB and its task forces. Dave Mills (Mills@HUEY.UDEL.EDU) INTERNET ENGINEERING I was gently pinged (actually more of a `ding', as in a ball- peen hammer rapped lightly against a heavy metal gong) that my June submission to the Internet Monthly Report was incomplete. I should have been more careful to mention that status and technical reports were made on both the new NSFnet and Arpanet. Hans-Werner Braun (U Mich) gave a report on the deployment status of the NSFnet and Jacob Rekhter (IBM) presented routing details. Marianne Lepp and Mike Brescia (both from BBN) presented status and performance information on the Arpanet and Internet. Marianne discussed the plans for evolving Arpanet into DRI. I also neglected to mention that Dave Mills (U Del) gave a very timely report on NTP. Mike Kramer and Eddie Singh (both from NYNEX) reported on efforts to develop and commercially offer a metropolitan-area-network technology. Although these were all listed in the final agenda, I neglected to specifically mention them in my summary of "major points". My apologies for the oversight. I've heard it said that three words often get reporters into trouble--"can't", "always" and "never". They soon learn to use mushier replacements like "might not", "often" and "seldom". In the same vein, the wording in my report ("THE major points were...") could be construed to imply that those things that I negligently forgot to mention were not "important" in their own right. This is not true, of course, since there is not enough room on a crowded IETF agenda to schedule "unimportant" issues. However, just to be safe in the future, I'll retreat to a more benign phrase like "SOME of the important points that I currently have time to summarize include..."! Phill Gross (Gross@GATEWAY.MITRE.ORG) Westine [Page 25]  Internet Monthly Report July 1988 INTERNET MANAGEMENT No report received. PRIVACY During July, minutes from the 15-16 June Privacy Task Force meeting at DEC in Littleton, MA were distributed to the task force membership, as well as to Jim Bidzos and Ron Rivest of RSA Data Security, Inc. An appendix to the minutes contained a revised draft of one section of the key management RFC currently in preparation, proposing a definition for certificate contents. Dave Solo of Sparta joined the Task Force. John Linn (Linn@CCY.BBN.COM) ROBUSTNESS AND SURVIVABILITY No report received. SCIENTIFIC COMPUTING No report received.