[IMR] IMR89-11.TXT November 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 INTERNET ACTIVITIES BOARD IAB MESSAGE . . . . . . . . . . . . . . . . . . . . . . page 3 INTERNET RESEARCH REPORTS . . . . . . . . . . . . . . . . page 3 AUTONOMOUS NETWORKS . . . . . . . . . . . . . . . . . page 3 END-TO-END SERVICES . . . . . . . . . . . . . . . . . page 3 PRIVACY AND SECURITY . . . . . . . . . . . . . . . . . page 4 USER INTERFACE . . . . . . . . . . . . . . . . . . . . page 5 INTERNET ENGINEERING REPORTS . . . . . . . . . . . . . . page 5 Westine [Page 1]  Internet Monthly Report November 1989 Internet Projects BARRNET . . . . . . . . . . . . . . . . . . . . . . . . . page 13 BOLT BERANEK AND NEWMAN, INC., . . . . . . . . . . . . . page 13 CERFNET . . . . . . . . . . . . . . . . . . . . . . . . . page 15 CICNET. . . . . . . . . . . . . . . . . . . . . . . . . . page 17 CORNELL . . . . . . . . . . . . . . . . . . . . . . . . . page 17 ISI . . . . . . . . . . . . . . . . . . . . . . . . . . . page 17 JVNCNET, NORTH EAST RESEARCH REGIONAL NETWORK . . . . . . page 19 LOS NETTOS . . . . . . . . . . . . . . . . . . . . . . . page 19 MERIT/UMNET . . . . . . . . . . . . . . . . . . . . . . . page 19 MIDNET . . . . . . . . . . . . . . . . . . . . . . . . . page 19 MIT . . . . . . . . . . . . . . . . . . . . . . . . . . . page 20 MITRE . . . . . . . . . . . . . . . . . . . . . . . . . . page 20 MRNET. . . . . . . . . . . . . . . . . . . . . . . . . . page 21 NCAR/USAN . . . . . . . . . . . . . . . . . . . . . . . . page 22 NEW ENGLAND ACADEMIC AND RESEARCH NETWORK . . . . . . . . page 22 NNSC, UCAR/BOLT BERANEK and NEWMAN, INC., . . . . . . . . page 22 NORTHWESTNET. . . . . . . . . . . . . . . . . . . . . . . page 23 NSFNET BACKBONE, MERIT. . . . . . . . . . . . . . . . . . page 23 NTA-RE/NDRE . . . . . . . . . . . . . . . . . . . . . . . page 24 NYSERNET . . . . . . . . . . . . . . . . . . . . . . . . page 24 OARNET . . . . . . . . . . . . . . . . . . . . . . . . . page 24 PENNSYLVANIA RESEARCH AND ECONOMIC PARTNERSHIP NETWORK . page 24 PITTSBURGH SUPERCOMPUTER CENTER . . . . . . . . . . . . . page 24 SAN DIEGO SUPERCOMPUTER CENTER . . . . . . . . . . . . . page 24 SESQUINET . . . . . . . . . . . . . . . . . . . . . . . . page 25 SRI . . . . . . . . . . . . . . . . . . . . . . . . . . . page 25 SURANET . . . . . . . . . . . . . . . . . . . . . . . . . page 26 TEXAS HIGHER EDUCATION NETWORK . . . . . . . . . . . . . page 26 UCL . . . . . . . . . . . . . . . . . . . . . . . . . . . page 27 UDEL. . . . . . . . . . . . . . . . . . . . . . . . . . . page 28 UNIVERSITY OF ILLINOIS AT URBANA-CHAMPAIGN/NCSANET . . . page 29 WESTNET . . . . . . . . . . . . . . . . . . . . . . . . . page 29 Westine [Page 2]  Internet Monthly Report November 1989 IAB MESSAGE INTERNET RESEARCH REPORTS ------------------------- AUTONOMOUS NETWORKS ------------------- The ANRG had a one day BBN-ISI videoconference on November 9. We discussed charging, fault-isolation, and application level relays in the context of administratively autonomous networks. Deborah Estrin (Estrin@OBERON.USC.EDU) END-TO-END SERVICES ------------------- The End-to-End Research Group met for a one-day teleconference on November 8. Topics discussed included the following. TCP TIMESTAMPS Van Jacobson reported experimenting with a TCP that is able to avoid the 3-way SYN handshake when timestamps are available. This approach was validated previously by the Mercury Project at MIT. HIGH-SPEED TRANSPORT PROTOCOLS Dave Feldmeier of Bellcore summarized briefly his research on a transport protocol for high-speed networking. He said that his proposed protocol is "not lightweight", will operate over datagram, virtual circuit, and fast-packet-switching networks, and will support the full range of qualities of service, from bulk transfer to constrained delays. Little information was given about how these things will be accomplished, and as this is corporate rather than university research, the papers on his work could not be made available to the group. As background, Feldmeier also described some of the Bellcore research on fast packet switches, and in particular the Sunshine and Dawn projects, and on resource reservation schemes. Clark, who is a participant in Dawn, added more information. Westine [Page 3]  Internet Monthly Report November 1989 PERFORMANCE ISSUES A. How can we find experimental evidence for the postulated "phase change"in the behavior of a complex network as congestion occurs? B. To what extent can local behavior in switches -- e.g., Fair Queueing, or the declumping advocated by some of the fast-packet-switching research -- change global network behavior? Shenker and Jacobson went a few more rounds on this. C. What would be the effect of Fair Queueing on large multi-thread hosts, e.g., large mail gateway machines? Some fear that the widespread introduction of Fair Queueing could result in collapse of the Internet mail delivery system. We need to understand the potential effect on email and DNS operation. To help answer such questions, we need data from real networks to analyze; in particular, we need timestamped packet traces. The group discussed several approaches to obtaining such data. NETWORK TIME The group discussed briefly the possible implications for network protocols of having universal clock synchronization, e.g., obtained with NTP. It was suggested that protocols could be simplified in a number of ways. Bob Braden (Braden@ISI.EDU) PRIVACY AND SECURITY -------------------- The minutes of the October 24-26, 1989 meeting of the Privacy and Security Research Group (PSRG) were distributed on November 22, 1989. Work is continuing on the implementation of Privacy- Enhanced Mail, with beta-testing of TIS's initial implementation to begin in December. In response to the workshop on requirements for the Commercial IP Security Option (CIPSO) conducted at Interop 89, the PSRG has initiated work on a framework document on labelling. Russ Housley will act as coordinator. Westine [Page 4]  Internet Monthly Report November 1989 The next meeting of the PSRG is scheduled for January 17-19, 1990 in California. This meeting will be held in conjunction with a meeting of the End-to-End Research Group, and a joint meeting of the two Research Groups is planned. Lyndalee Korn (lkorn@BBN.COM) USER INTERFACE -------------- No report received. INTERNET ENGINEERING REPORTS ---------------------------- IETF Chair: Phill Gross (NRI) The IETF met on 30 Oct - 4 Nov at the University of Hawaii. The meeting was be hosted by Torben Nielson (UH) and NASA. Highlights included 19 working groups meeting, an open steering group meeting, a focus on Pacific rim network connectivity, and the completion of the Point-to-Point Protocol Specification (published as RFC 1134). The Proceedings are in final preparation. We plan to mail Proceedings to attendees by the end of December. The IETF Steering Group (IESG) held an open meeting at the Hawaii IETF plenary. At that meeting, we decided, beginning this month, to include reports from each area in the monthly IETF report to the Internet. The eight area reports are below. APPLICATIONS AREA REPORT Interim Director: Phill Gross (NRI) There are currently two working groups in this area; - Telnet (Borman, Cray) - Domain Name System (Mockapetris, ISI) Each of these WGs has produced RFCs, but continue to work on related topics in their areas. We need to begin some activity in other application areas as well. Some proposal have been made: - Common protocol for remote printing that could be used by both TCP/IP and PC-based networks. Westine [Page 5]  Internet Monthly Report November 1989 - Remote backup facility from both TCP/IP and PC-based networks. - Electronic mail (bitmaps for SMTP, standardizing addressing hacks) - Usage of DECnet Naming Service in TCP/IP networks - Network FAX Some of this work would need to be coordinated with prospective application services (e.g., RCP) under the Host and User Services Area. The highest priority issue in this area is the designation of a permanent director. OSI AREA REPORT Directors: Ross Callon (DEC), Robert Hagens (Univ. of Wisconsin) The OSI Area has expanded rapidly. The initial OSI working group (OSIIWG) has been converted into a general OSI WG (OSI-General). In order to meet the challenge of operating OSI in a dual environment, the following new working groups have formed, or will be forming shortly. List of Working Groups Name: OSI-General Chair: Callon & Hagens Scope: + Forum for OSI-related issues not covered by an existing WG + Initial starting point of any OSI issue Name: OSI-X.400 Chair: Hagens Scope: + 822/X.400 gateway issues (including RFC 987 and successors) + Follow work of NIST X.400 groups Name: OSI-X.500 Chair: Deutsch (tentative chair) Scope: + X.500 and DNS interactions + Evaluation of any missing pieces in X.500 + Naming service requirements in a dual environment + Follow work of NIST X.500 groups Westine [Page 6]  Internet Monthly Report November 1989 Name: OSI-NSAP-ADMIN Chair: TBD Scope: + Produce NSAP administration guidelines Name: OSI-RA (Registration Authority) Chair: TBD Scope: + Produce X.400/X.500 name registration guidelines + Follow work of NIST and ANSI registration groups Name: OSI-MIB Chair: TBD Scope: + Definition of MIB variables for dual protocol hosts Readers interested in a summary of the work of the OSIIWG should consult the Current Meeting Report for the OSI-General working group. OPERATIONS AREA REPORT Interim Director: Phill Gross (NRI) There are currently two active WGs in the Operations area: - JoMAAN (Hares, Merit) - Benchmarking Methodology (Bradner, Harvard) JoMAAN is a long standing and productive WG. Benchmarking methodology is a new WG with the goal of developing standard methods for measuring performance in, for example, routers and bridges. We have identified the need for at least one more near-term WG -- a TCP/IP Installation Guide. A prospective chair has been identified and we will attempt to hold the initial meeting at the February IETF at FSU. I had not included an Operations Area in my original plans for an IETF steering group. Instead, I had included the important topic of User Services as a distinct Area. In our early IESG discussions, we realized that there were several broad topics that generated what amounted to long-term standing WGs, and these often were operations-oriented topics. Therefore, our original thinking was to broaden out the User Services Area to be an Operations Area, which would include network operations Westine [Page 7]  Internet Monthly Report November 1989 (e.g., JOMAAN), network information services (e.g., user services working group), and network connectivity planning (e.g., Interconnectivity working group). However, the User Services WG has been very active, and we finally decided to move those efforts under Craig Partridge's Host Services Area rather than let those efforts languish under an area without a director. At this time, we feel it is better to leave these activities under Craig, rather than attempting to eventually move it under the Operations area. To emphasis the committment of IETF to the important user services issues, we have changed the name of Craig's area to Host and User Services. It is a high priority matter to designate a permanent Director for the Operations Area. SECURITY AREA REPORT Interim Director: Phill Gross (NRI) This is an incredibly important area that demands immediate attention. I am very happy to announce that Steve Crocker (Trusted Informations Systems) has acepted the invitation to direct this IETF area. Steve has a rich history in the ARPAnet and Internet. Steve was the author of RFC 1 in April 1969, and in fact authored approximately one third of the RFCs published in that first eventful year. More recently, Steve has been involved in protocol verification techniques and security related matters. Next month Steve will report in this space. There is currently only one WG in this area: - IP Authentication (Schiller, MIT) However, this WG has essentially completed its initial objective of developing an IP Authentication option, and has moved on to developing a method for SNMP Authentication. Therefore, in the interest of keeping those objectives distinct, this WG has been asked to split into two -- IP Authentication and SNMP Authentication. Both will be chaired by Jeff Schiller until further notice. The goal will be to conclude the IP Authentication portion expeditiously, so full attention can be given to other matters. We have identified the need for at least one additional near-term WG -- the Secure Configuration WG. The goal will be to draft a short RFC documenting the proper ways to configure a new system to minimize the known windows for attack (e.g., turn off STMP debug, Westine [Page 8]  Internet Monthly Report November 1989 etc). We have tentative agreement from the CERT to join us in this WG. NETWORK MANAGEMENT AREA REPORT Director: David Crocker (DEC) The Network Management IETF Area has recently seen a flurry of activity and coalescence. Each of its three major areas has had development progress. In addition to the technical work, there appears to be a degree of stabilization to the specification process, for network management. The Area is broadly divided into three technical domains: SNMP- related protocol issues, CMIP-related protocol issues, and MIB- related data structure definitions. The MIB-related work further sub-divides into Transmission media -- broadly defined as anything below the IP layer -- and the rest of the MIB. The SNMP Working Group has tried to keep the SNMP protocol and the SMI framework for data structures completely stable, in order to minimize operational impact, so that the focus of their work has been to upgrade the core MIB, with 100 variables, up to about 170 variables. The core MIB was restricted to pure TCP/IP issues, except for very minor host-specific information. This emphasis has been retained in the upgrade, which is called MIB II. The NetMan Working Group, sometimes referred to as the CMOT Working Group, has renamed itself to OSI Internet Management Working Group. The group is continuing to pursue long-term use of the OSI CMIP protocol. A current debate is between the currently-published CMOT specification, versus a revision which would use the full CMOT, full SMI, and the full OSI upper stack -- as opposed to operating over the lightweight Presentation Layer, as currently defined. This would rely upon ISO advancing CMIP to full International Standard, which is expected to happen shortly. Though not yet formed, another group is developing. The range of specification efforts has the potential for unexpected and undesirable interactions, such as between MIB variables. Consequently, I am creating an advisory group to assist with coordination of the MIB specification(s) and to resolve any technical conflict. I hope to have it in place by the next IETF. Westine [Page 9]  Internet Monthly Report November 1989 HOST AND USER SERVICES AREA REPORT Director: Craig Partridge (BBN) This area combines two distinct activities: work on improving the quality of host-based services from the transport layer up to, but not including, applications; and work on developing and improving the quality of user services available on the Internet. In the area of Host-Based services, the plan for the next year is to actively encourage new work on standards for various support services such as remote procedure call, external data formats, distributed file systems, and network graphics. Under User Services, the key focus is on improving the services that already developed, and encouraging and fostering new activities such as the publication of the User Directory, the FYI notes, and the SIGUCCS project, that hold promise for improving user services offerings in the Internet. Report from the last IETF meeting: The major activities in Host-Based Services were the meeting of the Host Configuration Working Group (which is progressing faster than expected towards developing a configuratihon protocol) and the Ad- Hoc TCP Options WG (which got bogged down a bit at its meeting). In User Services, the major news is that Karen Bowers, after getting the User Services WG started under IETF, has taken on new responsibilities at NRI which make it difficult for her to continue as USWG chair. Joyce Reynolds (ISI) has agreed to be the new chair of the USWG, effective immediately. INTERNET SERVICES AREA REPORT Director: Noel Chiappa (Consultant/Proteon) The Internet Services Area is a fairly active one, with many groups focusing on particular issues. A major recent accomplishment is the production of the standard Point-to-Point Protocol document , and an initial options document. Both of these documents are now available. The basic protocol document has been submitted to the RFC editor as a proposed standard. Further work on the Point-to- Point protocol will be accomplished in the Point to Point Extensions working group, which will add support for additional protocols to the basic specification. Westine [Page 10]  Internet Monthly Report November 1989 Of the ongoing groups, the Connection IP group and the Performance and Congestion Control group, have documents well under way. A major long term issue in this area is congestion control. The initial round of mechanisms being discussed in the Congestion Control group may not be sufficient, and further work in the area (perhaps involving some research on the IRTF side) is needed. Router Discovery, MTU Discovery, and IP over FDDI have been formed and are quite active. Discovery of available routers is a pressing problem. Currently the community has to either configure hosts with router addresses, or use a variety of non- standard techniques to find them; something standard is clealy needed quickly. Several new groups are in the process of being formed, and will be announced shortly. The most important is a Router Requirements group, to redo RFC-1009 and bring it up to the standards of the Host Requirements RFC's. A group is being put together to standardize use of IP with the Appletalk environment. Among other things, it will document the KIP protocol, which has come into wide use. A working group is being set up to address IP over Multi-Media Bridges. There are a number of vendors who wish to offer multi- media bridges, but there are a number of technical issues to be solved before the IP protocol family will operate over such devices. Multi-cast has been in an interim state in the architecture for years. A number of RFC's on the issue have appeared, but they need to be brought forward to the "recommended" state and mandated for use. Other topics will be receiving attention soon. The Internet needs a clear standard for the use of variable length subnet masks. The original subnet RFC did not deal with the details of this issue, and it needs to be regularized. Detection of dead nodes, particularly dead routers is a pressing problem. Most hosts fail to recover gracefully from routers that crash, and although the Host Requirements RFC discussed the issue, more work is needed. A group is being created to consider the issues involved in integrating the Switched Megabit Data Service, SMDS, into the IP architecture. This new service to be offered by the phone companies will provide true packet service (i.e. no connections or connection setup) over a T3 rate interface. It looks like the world's largest LAN, and presents some scaling problems for the IP architecture; clearly, ARP cannot be used in its existing form! Westine [Page 11]  Internet Monthly Report November 1989 ROUTING AREA REPORT Director: Bob Hinden (BBN) The major issue in this area is the topic of a standard internal gateway routing protocol (IGP). The IESG discussed this in detail at the open meeting in Hawaii. We plan to make this tpoic the focus of a special meeting at the next IETF meeting at Florida State University (Feb 6-9, 1990). Because of its importance and its early promise, we have also decided to form a WG to specifically examine at the experimental Border Gateway Protocol (BGP). One possible outcome would be for BGP to eventually replace EGP as the exterior gateway routing protocol. Another possible outcome might be that the better parts of BGP could become a basis for a new or better EGP. Phill Gross Westine [Page 12]  Internet Monthly Report November 1989 INTERNET PROJECTS ----------------- BARRNET ------- No report received. BOLT BERANEK AND NEWMAN INC. ---------------------------- TERRESTRIAL WIDEBAND NETWORK AND ST/IP GATEWAY During November, the Terrestrial Wideband supported 8 video conferences. These included meetings of the IETF Open Routing WG, the IETF User Doc WG, the IRTF Autonomous Networks RG, the IRTF End-to-End RG, and the Internet Engineering Steering Group. Professor Leonard Kleinrock (UCLA) held discussions with DARPA on November 22. Dr. Barry Boehm held an ISTO meeting with his staff at DARPA on November 27. This month we worked out a plan with DARPA to use the Terrestrial Wideband and the ST gateway to support large scale distributed war gaming as implemented by the SIMNET system. We are working to have sufficient infrastructure in place by March 1990 to support a 5- site SIMNET demonstration involving 800 simulated vehicles. Software development of an ST host for this wargaming application began this month. In the last week of November, we ran a successful experiment with 4-way conferencing. INTERNET RESEARCH The open routing architecture paper is mostly completed and will be available for comment in December. The major elements of the architecture are clustered policy gateways (for redundancy and data reduction); centralized policy, route, and address servers; source routing; and route set-up. The paper discusses these elements both in terms of their functionality and in terms of their instantiations in internet entities. The next steps after completing the architecture paper is to specify the Interior Gateway to Policy Gateway Protocol and the mechanisms for dissemenination of outage data. Westine [Page 13]  Internet Monthly Report November 1989 INTERNET O&M The biggest event of the month was the partition of the Arpanet from 11/4 to 11/13. This partitioning was experienced by many of the Internet users as routing problems, causing many complaints to the NOC, to gw-fire, and to the gateway group. The reason becomes clear when you consider the following picture: host 10.e.f.g | _____________________________________________________ net 10 | | | | | | | | | | trunk X X | | X down | | | | isolated O O | | O psn | | | | | | | | | | ^ ^ ^ ^ ^ BMIL gws (mailbridges) - -- - -- - | | | | | | | | | | | | | | | | | | | | ______________________________________________________ net 26 | | ___________ | random- | | gateway | ----------- | | ( random-) ( net )------ host r.s.t.u If a user on host 10.e.f.g is trying to get to host r.s.t.u on random-net, his message may get all the way to r.s.t.u. But random-gateway may "know" that the route back to net 10 is through BMILISI (or BMILAMES, or BMILMTR). It will send the return message to BMILISI, which will send it out to its PSN, which will drop it because the trunk is down. So even though the source and destination are up, and there is a path between them, there is a problem. Westine [Page 14]  Internet Monthly Report November 1989 Gateways in general have a special problem when a directly- connected network is partitioned. The gateway routes at the network level, and cannot recognize a partition in order to re- route traffic to another gateway with access to a non-isolated node. During this partitioning, it was suggested that we unplug the Arpanet interface cables from the backs of BMILISI, BMILAMES, and BMILMTR. This would have let those gateways know that the Arpanet interfaces were down, so they wouldn't have advertised routes to the Arpanet through themselves. However, this would have adversely affected the hosts on PSNs 22, 51, 111 and 17 (behind 111). They were not experiencing any routing problems, but they would have if we had done this. Though there is no solution to the general problem, the Arpanet topology is being made more fail-safe by the addition of a microwave link between BBN and MIT, so in the event the T1 trunk from east to west coast goes down, the VSAT links at MIT will carry all the cross-country traffic. Bob Hinden (Hinden@BBN.COM) CERFNET ------- Network Performance The installation of NYSERnet version of Simple Network Management Protocol (SNMP) on CERFnet has aided network operations tremendously. A graphical display of CERFnet illustrated on the SUN workstations alerts the staff of network troubles in near real- time. The California State University (CSU) Chancellor's Office T1 link to the University of California at Irvine was down from Friday, November 24 until Wednesday, November 29. Because of this the CSU campuses did not have access to CERFnet. After several days of extensive troubleshooting, including swapping out the CSU/DSUs at each end and testing the circuit, the link was brought up. The cause of the problem is still being investigated. A full report will be issued in the upcoming weekly network managment report. Interested parties may obtain CERFnet weekly network management reports via anonymous ftp to sds.sdsc.edu in the subdirectory [.cerfnet_stats]. Type cd [.cerfnet_stats] to change to the subdirectory. Westine [Page 15]  Internet Monthly Report November 1989 Installations CERFnet is adding a new backbone member. The University of California's Office of the President located in Oakland, California will have a 512 kilobits-per-second link to the San Diego Supercomputer Center in the near future. Five new industrial members will be brought up on CERFnet. A division of Xerox Corporation in San Diego is scheduled to be brought online in early December. Xerox will have a 56 kbps link to the SDSC. Also, a T1 link from Science Applications International Corporation of San Diego to SDSC is scheduled to be brought up in mid-December. Scheduled to be brought up in late January is a 56 kbps link to Quotron Systems Incorporated of Los Angeles. Quotron will be connecting to CERFnet through UCLA. Supercomputing Solutions of San Diego will be brought online in late January. Supercomputing Solutions will have a 56 kbps link to SDSC. A 56 kbps link between Science Horizons of Encinitas and SDSC will be brought online in early February. Minutes CERFnet held an open session on October 24 to discuss the need for CERFnet provided dial-up service. A solid three hour discussion was held with participants representing over 25 academic and commercial institutions. Among the items discussed were participants need for dial-up service, types of services to offer, i.e. e-mail and file transfer, the different technologies that are available, and how much dial-up should cost. Results of this session will be summarized and used to design a CERFnet dial-up service. Other Activity CERFnet recently announced the First Annual Summer Research Fellowships. Up to two $5,000 fellowships are available to support full-time graduate student research in the summer. The December 1989-January 1990 issue of CERFnet News was completed in November. The issue contains articles on how Simple Network Management Protocol is working on CERFnet, CERFnet NIC services, an overview of The Hitchhikers Guide to the Internet, the new CERFnet Westine [Page 16]  Internet Monthly Report November 1989 fellowships, the CERFnet Report column, and the Internet Notes column. Electronic distribution will begin December 4. Hard-copy distribution will begin December 18. Back issues of CERFnet News are available via anonymous ftp to sds.sdsc.edu in the subdirectory [.cerfnet_news]. by Karen Armstrong (armstrongk@sds.sdsc.edu) CICNET ------- No report received. CORNELL ------- No report received. ISI --- INTERNET CONCEPTS PROJECT Bob Braden added a much-requested feature to the packet monitor program statspy: the ability to count bytes as well as packets. A new release including this feature was nearly completed this month. Work is also underway to add to statspy the extended language features mentioned in the SIGCOMM '88 paper: boolean expressions and case statements. Bob Braden held a one-day video teleconference for the End-to-End Research Group on November 8, and presented an internal ISI seminar on ATM. He attended a one day meeting at ISI of the group responsible for the engineering of the DRI, and continued to help DARPA with planning experiments for the DRI open-gateway testbed. Bob Braden Greg Finn has begun comparison testing of the IP/SQ algorithm within SUN's UNIX. Jon Postel attended a Calinet meeting in Oakland, CA, 16 November. Joyce Reynolds attended the IETF meetings at Hawaii, October 30-Nov 3. Westine [Page 17]  Internet Monthly Report November 1989 Four RFCs were published this month. RFC 1125: Estrin, D., "Policy Requirements for Inter Administrative Domain Routing", USC, Computer Science Department, November 1989. RFC 1132: McLaughlin, L., "A Standard for the Transmission of 802.2 Packets over IPX Networks" The Wollongong Group, November 1989. RFC 1133: Yu, J., and H-W. Braun, "Routing between the NSFNET and the DDN", Merit Computer Network, November 1989. RFC 1134: Perkins, D., "The Point-to-Point Protocol: A Proposal for Multi-Protocol Transmission of Datagrams Over Point-to-Point Links", CMU, November, 1989. Ann Westine (Westine.ISI.EDU) MULTIMEDIA CONFERENCING PROJECT Use of the teleconference system is accelerating! Eight teleconferences were held this month for various IETF and IRTF groups and for DARPA staff members. Half of these were three-site conferences. A series of experiments were started with the sound I/O facilities on the NeXT machine, as part of our investigation of that workstation as a platform for our real-time voice and video software. New versions of MMCC, the multimedia conference control program, and MBFTPTOOL, the multiple site background FTP program, were installed at the conference sites. The latest version of MBFTPTOOL is available via anonymous FTP from the "pub/" directory on venera.isi.edu in compressed tar-file format, BFTP.213.tar.Z. Steve Casner attended the IETF ST/CoIP Working Group meeting at UH. Annette DeSchon, Dave Walden, Eve Schooler, Steve Casner (deschon@ISI.EDU, djwalden@ISI.EDU, schooler@ISI.EDU, casner@ISI.EDU) FAST PARTS No internet-related progress to report. Westine [Page 18]  Internet Monthly Report November 1989 JVNCNET, NORTH EAST RESEARCH REGIONAL NETWORK --------------------------------------------- No report received. LOS NETTOS ---------- Jon Postel and Walt Prue attended the Calinet meeting held in Oakland November 16. The North-South link was discussed again. Completion is delayed until 1990. A few possible options on how to use the link were discussed. The next Calinet meeting is scheduled for 29 March 1990 at ISI. Unisys (Camarillo) was brought up as the latest Los Nettos member 29 November 1989. Unisys is connected through UCLA. We have a signed contract from NOSC to add them to Los Nettos. NOSC will be connected through ISI. We experienced a routing problem with two of our networks. There were no links down to indicate any problem existed. However SNMP based route monitoring tools immediately alerted us to the existence of a problem. SNMP based diagnostic tools helped us to quickly isolate the problem. Ping is not enough to verify proper operation of a network. Walt Prue (Prue@ISI.EDU) MERIT/UMNET ----------- No report received. MIDNET ------ MIDnet has added several new members in the last couple of months: McDonnell Douglas Corp. in St. Louis EROS Data Center in Sioux Falls, SD University of South Dakota in Vermillion, SD University of Missouri-Rolla University of Missouri-St. Louis University of Missouri-Kansas City Westine [Page 19]  Internet Monthly Report November 1989 Activity in MIDnet is increasing quite a lot of late. Much of this is due to the extensive archives currently available at Washington University (anonymous ftp to wuarchive.wustl.edu). by Dale Finkelson (dmf@westie.unl.edu) MIT-LCS ------- "A New Architecture for Packet Switching Network Protocols" by Lixia Zhang is now available as technical report MIT/LCS/TR-455. Copies can be obtained ($15.00 each) from: MIT Laboratory for Computer Science Publications Office 545 Technology Square, Room 113 Cambridge, MA 02139 (617) 253-5894 Andrew Heybey prepared a new release of the MIT network simulator. Work continues on an X-based DUA tool for use in conjunction with the Quipu X.500 directory software. Chuck Davin (jrd@PTT.LCS.MIT.EDU) MITRE Corporation ----------------- We have completed a draft technical report on our Internet Engineering work for the past year. It will be titled "Evaluation of Internet Performance -- FY89." (The previous year's report was called "A Model of Internet Performance Measurement"). The work was done in the DCA Internet Engineering Net, which is joined to NSFNET and the MILNET in a transit LAN here at MITRE. The DCA Internet Engineering Net is a large and growing facility based on T1 links. Its host and gateway systems are extensively instrumented, with Ethernet-based tools similar to tcpdump, and with NETMON, an internal protocol performance tracing facility for BSD UNIX. Most of the systems now have OSI stacks as well as DoD. The DCA Net is dedicated to Internet development and evolution projects, including OSI transition. Projects are mainly those sponsored by the Defense Communication Engineering Center, but this year we have provided testing facilities for the University of Maryland OSPF implementation, and other such cooperation is welcomed. In the year's work, we implemented and evaluated the DEC-NIST OSI Congestion Avoidance algorithms. Our implementation is for the University of Wisconsin TP4-CLNP, which is the basis for the OSI- POSIX stack being integrated into Berkeley 4.4. We have begun to Westine [Page 20]  Internet Monthly Report November 1989 integrate our code into Berkeley's alpha release, so that the algorithms can be present as a per-connection option by the time of the beta release. Rick Wilder gave an informal presentation on this work at the IETF OSI Interoperability Working Group in Hawaii. The tech report will be publicly released in early 1990, along with a tech report on our measurement tools and techniques, called "Design and Implementation of an Instrumented Gateway". The other performance evaluation results, on Source Quench and Random Drop, will be discussed in next month's Internet Report. A distribution of NETMON is available by anonymous FTP from aelred-3.ie.org and comments are welcomed. We have recently started a mailing list about our tools, ie-tools(-request) @gateway.mitre.org. Allison Mankin (mankin@gateway.mitre.org) MRNET ----- Work continues on upgrading MRNet's link to the NSFnet backbone. The T1 line has been installed, but is not yet operational. Jeff Wabik is working with UIUC to coordinate the migration to the new link, new routers, and an Autonomous System Number for MRNet. The movement to an independent system number for MRNet should migrate much of the work for maintaining MRNet routing from UIUC. The Minnesota private colleges' proposal for connection to MRNet continues to be approved in principal. Nonetheless, the implementation of their connection awaits formal approval of funding. 3M connected to MRNet at 56k bps. The four incumbents for the MRNet Executive Committee were the only people accepting nomination. Elections for 1990 Executive Committee members will be completed shortly. The following organizations are in various stages of applying for MRNet membership: o Minnesota private colleges (10 schools) o Minnesota State University System (7 schools) o Secure Computing Technology Corp. o Management Graphics Westine [Page 21]  Internet Monthly Report November 1989 o Unisys The next MRNet general meeting is tentatively scheduled for January at MSC. Jeff Wabik, Craig Finseth, and Stuart Levy plan to offer their "Internetworking with TCP/IP" seminar again sometime in February. by Timothy J. Salo (tjs@msc.umn.edu) NCAR/USAN --------- No report received. NEW ENGLAND ACADEMIC AND RESEARCH NETWORK ----------------------------------------- NEARnet added Motorola Cambridge Research Center to the network during November. NEARnet and Microwave Bypass Systems, manufacturer of 10Mb/s microwave equipment, provided a 10Mb/s connection from an OSF sponsored show to the Internet. A demonstration of Transarc's Andrew File System featured use of the connection. Operation of the network continued to be stable. by John Rugo (jrugo@bbn.com) NNSC, UCAR/BOLT BERANEK and NEWMAN, INC. ---------------------------------------- NNSC/UCAR/BOLT BERANEK AND NEWMAN INC. Craig Partridge gave a presentation discussing the IAB, IETF, and Gigabit Networking at the Westnet Technical workshop. Craig Partridge and Karen Roubicek attended the IETF. The NNSC published Chapter M of the Internet Resource Guide, as well as revisions to Chapters 1, 3, and 6 of the guide. Requests to be added to or removed from the distribution list should go to resource-guide- request@nnsc.nsf.net. The guide is also available via anonymous ftp at , cd resource-guide. by Karen Roubicek Westine [Page 22]  Internet Monthly Report November 1989 NORTHWESTNET ------------ No report received. NSF BACKBONE (Merit) ------------------- During November 1989, NSFNET backbone traffic passed the two billion mark. This represents an increase of 7% over September 1989 and a 550% increase over November 1988. In addition, as of 11/30/89 total network connections stand at 1897. Border Gateway Protocol Border Gateway Protocol (BGP), a prototype Inter-Domain routing protocol, is now in use between the Merit mid-level network and the NSFNET backbone. BGP is performing functions compatible to the Exterior Gateway Protocol (EGP) which is currently used across the NSFNET. The expansion of BGP to its full architectural model as a large-scale border gateway protocol still requires further experimentation and evaluation. Research Network All circuits on the NSFNET Research Network are now being monitored with the Cylink Network Management System (CNMS) instead of the Integrated Digital Network Exchange (IDNX). CNMS gathers information about the digital circuits by querying the Advanced Channel Service Units (ACSUs) which are part of the Nodal Switching Subsystem (NSS). MERIT/NSFNET Staff Activities Bilal Chinoy attended the 5th Annual Supercomputer Conference in San Diego where he gave a presentation on network performance benchmarking. Susan Calcari took part in several activities during CAUSE 89 in San Diego including a pre-conference seminar on "Introduction to Internetworking/NSFNET Overview." Susan also took part in a panel discussion regarding regional and national networking and funding for the National Research and Education Network. Susan Hares gave presentations on routing at two regional network meetings: NorthWestNet in Seattle and Westnet in Salt Lake City. RFC 1133, "Routing between the NSFNET and the DDN" by Jessica Yu Westine [Page 23]  Internet Monthly Report November 1989 and Hans-Werner Braun has been published. This document is a case study of the implementation of routing between the NSFNET and the DDN components (MILNET and ARPANET). Kenneth Latta recently completed a one-month residency at the IBM International Technical Support Center in Raleigh, NC. Other November activities included completion of a NetView-Information Management software bridge. by Patricia G. Smith (patricia_g._smith@um.cc.umich.edu) NTA-RE and NDRE --------------- No report received. NYSERNET -------- No report received. OARNET ------ No report received. PREPNET-Pennsylvania Research and Economic Partnership Network -------------------------------------------------------------- No report received. PITTSBURGH SUPERCOMPUTER CENTER ------------------------------- No report received. SAN DIEGO SUPERCOMPUTER CENTER ------------------------------ Since our last report: Both our Proteon p4200 and cisco AGS have been updated to the latest rev levels. We continue to have linemode problems with Telnet under UNICOS - usually caused by non-cooperating clients. Our old PSN (26) finally did get removed. Westine [Page 24]  Internet Monthly Report November 1989 The latest version of Multinet has been installed. With it, our VMS systems can be watched by SNMP. During December, we are planning two Flag Days during the month to move the Center from our current Class C address (192.12.207) to our Class B address (132.249). The center point of 132.249 will be our NSC EN641. S. Arnold and P. Love attended the last IETF meeting. S. Estrada attended the November CaliNet meeting. SDSC continues to be active in these plus FARnet. by Paul Love (loveep@sds.sdsc.edu) SESQUINET --------- No report received. SRI ---- DDN NIC Over 200 new IP network numbers have been assigned in November, which brings the total of all IP networks assigned to 3,648; that total includes 2,084 connected networks and 1,564 unconnected networks. Domain registration was slightly higher than usual, with 60 new domains being registered this month. There are now a total of 1,234 domains registered, including 44 at the top-level, 1,152 at the second-level, and 38 third-level MIL domains. Mark Lottor, Mary Stahl, Jose Garcia-Luna, and Zaw-Sing Su attended the IETF meeting at University of Hawaii. Lottor made a presentation of recent results from his domain-tree-walker program and related information. by Mary Stahl (Stahl@NISC.SRI.COM) Westine [Page 25]  Internet Monthly Report November 1989 Internet Research Jose Garcia-Luna participated in the Open Distance Vector Routing Working Group at the last IETF. He presented a loop-free routing algorithm based on distance vectors, and simulation reults comparing it with the distributed Bellman-Ford algorithm and topology-broadcast algorithms. Jose Garcia-Luna (Joaquin@NISC.SRI.COM) SURANET ------- SURAnet continues to increase in the number of sites connected and in the number of networks advertised to the NSFnet. At present there are 75 sites online and 124 networks are being advertised to the NSFnet. The current list of sites and networks can be obtained via anonymous FTP from noc.sura.net, password guest, cd pub. File name is "online". by Jack Hahn (hahn@umd5.umd.edu) TEXAS HIGHER EDUCATION NETWORK ------------------------------ On 15 September, UT System OTS brought up a T1 data circuit between UT Austin and Rice University in Houston, replacing a 56 Kbps circuit. This circuit increased the capacity of the link by a factor of 24, giving THEnet better connectivity to the NSFNET backbone and the Internet in general. This circuit is the first leg of an Austin-Dallas-Houston T1 triangle. On 1 November, UT System OTS brought up a T1 data circuit between UT Austin and UT Dallas, replacing a 56 Kbps circuit. This circuit provides better connectivity between UT Dallas and UT Austin, and between UT Dallas and the NSFNET backbone. This circuit is the second leg of an Austin-Dallas-Houston T1 triangle. On 20 November and 21 November 1989, UT System OTS personnel traveled to Edinburg and Brownsville to install the routing equipment necessary to (1) install a 56 Kbps DECnet circuit between UT Pan American at Edinburg and UT Pan American at Brownsville, and (2) install a 56 Kbps DECnet circuit between UT Pan American at Edinburg and UT San Antonio, replacing the 9600 bps data circuit between UT Pan American and Southwest Texas State University. The Westine [Page 26]  Internet Monthly Report November 1989 routing equipment consisted of a DECSA Communications Server placed at Edinburg and a VAXstation II placed at Brownsville, along with the necessary Data Service Units and Ethernet transceivers required to make all the appropriate connections. Both machines were installed with no problems. The BrownsvilleEdinburg circuit was already installed by the phone company and was ready for the routing equipment, but due to a faulty DSU at Edinburg and remodeling operations at Brownsville the circuit was not made operational until November 30. The Edinburg-San Antonio circuit has not yet been installed by the phone company. It is expected to be operational the first of next year. by Don Nash (don@thenic.the.net) UCL ---- We have now successfully ported the X11 Windows protocol over the ISODE ISO Transport Service. This works over both TCP (treated as network service) and X.25. We are working on aligning the mapping with ANSI proposals for such a stack. You may have heard that UCL is developing a Message Transfer Agent called PP. We have reached the stage where the system is ready to let out of UCL on a limited basis. This message is a request for beta test sites, and also announces a mailing list. MAILING LIST For anyone interested in PP, there is a mailing list: pp- people@cs.ucl.ac.uk. To join, send to pp-people- request@cs.ucl.ac.uk (S=pp-people-request; OU=CS; O=UCL; PRMD=UK.AC; ADMD=Gold 400; C=GB;). WHAT IS PP PP is a Message Transfer Agent, developed primarily at the Departments of Computer Science at UCL and Nottingham. This develpment started in early 1986, aimed as a follow on to MMDF, getting things "right", and adding in X.400. This took much longer than we'd have imagined, and has also changed significantly since the original ideas. The main aims of PP are: - robust operation for switching large numbers of messages - support for other groups to build UAs - managability Westine [Page 27]  Internet Monthly Report November 1989 - support for multiple message transfer protocols and conversion between them - message content conversion The major components of the first release of PP will be: - (UA) service interfaces giving a superset of RFC 822 and X.400(88) funtionality, including some aspects beyond both. Thies, who will be planning to replace MMDF or Sendmail with PP. I will look at commercial test sites, but will need persuading. Any beta test site should be running ISODE now. PP will link with ISODE 5.0 (recommended) or the version on ISODE development sites. Beta test sites shoudl be familiar with large mail systems used in complex configurations. We are interested in any sort of feedback, including: - bug fixes - portability issues, and ports to new UNIX flavours (currently SUNOS and Ultrix only). - functionality issues (particularly if fixes offerred) - new UUCP channel (the current channel and rmail are not very satisfactory) - new Channels We are interested in sites which plan to work on (openly available) UAs. Whilst PP can be used with MH, we are hoping that state of the are UAs will be contributed to go with PP. In general, I'd like sites which are able to put in some real effort, and want to make the whole thing fly. If you just want to look, please wait for 5.0. John Crowcroft (jon@CS.UCL.AC.UK) UNIVERSITY OF DELAWARE ---------------------- 1. Paul Schragger continues work on the gigabit reservation algorithm, which now generates collision-free timetables for reservation requests. He is also developing a formal graph- theoretic description based on a series of open, directed- trail, covering-enumeration algorithms for constrained bipartite lattices, hereafter known as trailblazer algorithms. Westine [Page 28]  Internet Monthly Report November 1989 2. Mike Davis has snarfed a third of a gigabyte of traffic- monitoring data on the NSFNET backbone and begun design of parsing programs to extract interesting data. Mike, Erik Perkins and Ken Monington are finishing up the design of a VLSI switch for our gigabit project. Erik is also working on experiment analysis and design for RIG related activities. 3. Dave Mills presented briefings on our high-speed network activities at a UNIXEXPO panel and attended a telemeeting of the E2E task force. He is also working on papers for CCR, including a report on the last INARC workshop. 4. Analysis and simulation have been completed on improved algorithms for use in Network Time Protocol (NTP) implementations. The new design improves frequency stability to the order of a millisecond per day, while reducing the polling overhead by a factor of 16. New algorithms for combining the timescales of multiple peers have been adapted from those used at NIST and other national standards labs. 5. With the kind cooperation of the US Coast Guard, we have installed a LORAN-C receiver to keep our cesium clock company and insure it keeps accurate time to the microsecond. However, it seems we have to know a bunch of other things, including precise geographic coordinates, receiver delays and whatnot, so we plan to simply haul it to US Naval Observatory for rewind of spring. Of course, since the clock can never be turned off, transporting it any distance means heavy-duty batteries, power inverters, etc. Dave Mills (Mills@UDEL.EDU) UNIVERSITY OF ILLINOIS AT URBANA-CHAMPAIGN/NCSANET -------------------------------------------------- No report received. WESTNET -------- 1. Westnet held its annual Technical Workshop in Boulder, CO, November 15-17. There were excellent presentations on PC networking and Supercomputing applications by Elizabeth Walden, Dean Wallace, and Mark Christon. We are also very appreciative of Craig Westine [Page 29]  Internet Monthly Report November 1989 Partridge of BBN, Sue Hares of MERIT, Eric Decker of cisco, Roy Perry of US West, and MCI for their participation in the workshop. 2. We continue to try to get a working T1 circuit between NCAR (in Boulder) and New Mexico Technet (in Albuquerque). 3. We have tried running cisco's 8.0(6) firmware on the Westnet router at NCAR, but dropped back to 7.0 because of difficulties with propagating Westnet routes via EGP to the NSS. 4. We have made a minor reconfiguration between Idaho and Utah. The connections now runs from Univ of Utah to Idaho State Univ., and from there to the Idaho National Engineering Laboratory to Boise State University. 5. Cray Computer Corporation in Colorado Springs is now connected. by Pat Burns (pburns@csupwb.colostate.edu) David C. M. Wood (dcmwood@spot.colorado.edu)