[IMR] IMR86-08.TXT AUGUST 1986 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 task forces and contractors in the ARPA Internet Research Program. This report is for research use only, and is not for public distribution. Each task force and contractor 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 ARPANET mail to Westine@B.ISI.EDU. Reports are requested from BBN, ISI, LL, MIT-LCS, NTA, SRI, UCL, and UDEL. Other groups are invited to report newsworthy events or issues. BBN LABORATORIES AND BBN COMMUNICATIONS CORPORATION --------------------------------------------------- WIDEBAND NETWORK On August 4, a new BSAT/ESI-B installation was performed at DCEC in Reston, VA. With that installation the BSAT-based Wideband Network has grown to a total of nine operational sites. The eighth operational Wideband Butterfly Gateway was installed the following day, connecting a local ethernet at DARPA in Arlington, VA to the Wideband Network via the DCEC BSAT. The gateway's Wideband Network connection is supported by a 1.5 Mb/s T1 phone link running between DARPA and DCEC. Multiplexers at either end of the T1 line will also 1 provide a connection between the DCEC BSAT and a Voice Funnel machine to be installed at DARPA at a future date. On August 14, the Wideband Network provided the high-bandwidth data service required to conduct an all-day, cross-country Internet Activities Board video conference/meeting. The conference, conducted at BBN in Cambridge, MA, and at ISI in Marina del Rey, CA, represented the longest such continuous use of the Wideband Network to date. The conference was supported by ISI's packet video system, BBN's Voice Funnel system, Butterfly Internet Gateway, and the Diamond Multimedia Conferencing system in addition to the Wideband Network. The BSAT machines at both CMU and M/A-COM Linkabit were upgraded to consist exclusively of 1-Mbyte Butterfly processor nodes. All of the BSATs have now been so upgraded. Two new features that should prove useful for Wideband Network operations and maintenance have been implemented in the BSAT software. The first feature permits control of any arbitrary set of remote BSATs to be exercised from any local BSAT console. The second feature provides a built-in BSAT satellite channel bit error rate testing facility. These features will be included in the next BSAT software release. GATEWAYS The major event of the month was the installation of a new software release (Rel. 3.2) into the Satnet Butterfly Gateways. The new software is now running at NTA, CNUCE, CSS, and UCL. First reports are that it is much more reliable and allows mail, ftp, and telnet to flow. The next step, after it is stable at UCL for a few weeks, is to reconnect the UCL Butterfly Gateway back to Satnet. SATNET The performance of the SIMPs remained stable. There appear to be some problems with channel 1 (errors and some lost hellos). However until we obtain spare modems from Linkabit, we have decided to refrain from attempting any repairs. Since channel 0 performance was stable, SATNET service continued uninterrupted. This month, a new release of Butterfly gateway software was installed at NTARE and CNUCE. As a result, connectivity between the SATNET and the local networks at these sites was greatly improved. The SIMP configuration at Goonhilly was changed in preparation for the re-installation of the Butterfly gateway at UCL next month. Bob Hinden 2 ISI --- Internet Concepts Project Jon Postel, Bob Braden, and Paul Mockapetris participated in the TCP/IP Workshop in Monterey, CA, 25-26 Aug. This workshop was very successful in brining together the technical people from commercial vendors of products that incorporate IP/TCP protocols and experts from the research community. In theory most of the participants already knew about the protocols, so this was not a tutorial rather a discussion of the issues and pitfalls not covered in the documents, a passing on of the folklore. About 275 people participated, with about 100 companies represented. One result of the workshop is the formation of a IP/TCP vendors coalition, currently being coordinated by Dan Lynch (LYNCH@B.ISI.EDU). Annette DeSchon continues to work on developing a domain name resolver which will be accessed via the standard software interfaces in the XEROX ARPA protocols package, as well as by a special purpose test tool (NameTool) running in the XDE. We welcome two new ISIers to DIV-2, Francisco Chaparro, and Rivi Sherman, who are working on the FAST-PARTS project. Jon Postel and Ann Westine Multimedia Conferencing Project The prototype multimedia teleconferencing facility passed its toughest test to date on August 14 when the first bi-coastal Internet Activities Board meeting was held between meeting rooms at ISI and BBN. During the 7-hour meeting there was continuous packet video transmission in both directions except for a few-minute outage caused by the accidental reactivation of a Wideband Network site having some failed equipment. Service resumed automatically without manual intervention. The Wideband Network carried both the video channel from ISI's packet video system and the shared-workspace display of BBN's Diamond multimedia conferencing system. Video traffic uses the ST protocol through the Voice Funnel gateway and Diamond uses IP through the Butterfly gateway. From all appearances, the communication path provided by the teleconference facility allowed the meeting to be essentially as effective as a face-to-face meeting. For the IBM-PC based image scanner system, Brian Hung has added an additional write_to_file function to the current set of functions. 3 This function has been incorporated into a new application which scans, clips and displays documents and also writes the scanned data into a file simultaneously. Joyce Reynolds presented a Multimedia Mail demo to a group of people from the University of Davis, and Dongman Lee from KAIST in Korea. Steve Casner and Brian Hung Supercomputer and Workstation Communication Project Alan Katz attended the Cornell Summer Supercomputer Institute in Ithaca, N.Y., and worked on a paper on issues concerning an Equations representation standard. Ann Westine Computer Center No internet-related progress to report. MIT-LCS ------- No report received. NTA & NDRE ---------- No report received. SRI --- No report received. UCL --- 1. Further tests using the Network Time Protocol over SATNET and ARPANET have been conducted against various remote NTP servers. Because our own radio signal receiver has broken, we have been taking the time from the remote server and then calculating the offset between its clock and ours, using the round trip delay. Analysing the data was complicated by the fact that, over the last few weeks, we have been regularly recording round trip delays of over 50 seconds, and occasionally some reaching 100 seconds. However, after making the necessary adjustments, the offset variance remained at less than half a second. 4 2. Jon Crowcroft attended the TCP-IP Vendors conference at Monterey, putting the case that internetworking from our side of the Atlantic is a little more difficult. Jon also called in at BBN on the way and talked with Jill Westcott about network management. Peter Lloyd UDEL ---- 1. I have been appointed Professor in the Department of Electrical Engineering at the University of Delaware, effective 1 September, so this and future reports will originate from there. I will continue as a consultant to Linkabit and assist in their Dissimilar Gateway Protocol program for RADC and Ford Aerospace, as well as in their Multiple Satellite System program for DARPA, RADC and a cast of thousands. 2. I began the intricate process of moving myself and DCnet swamp creatures, creeks and bogs to the University of Delaware. The present plan is to rehome the DCnet swamp via the UDEL IMP (96), but leave some fuzzcreatures behind to maintain ARPANET connectivity for the NSFNET Backbone clients, as well as the University of Michigan, University of Maryland, Ford Scientific Research and Ford Aerospace. 3. Much of my effort this month was in the design and simulation of routing algorithms for MSS. I distributed a preliminary summary of issues and suggestions to the INARC for comment early in the month and prepared more substantive simulation data and results for a presentation at the SURAN Implementors meeting early next month at BBN. A new algorithm I have called Shortest Total Path First (STPF) works very well for metric routing in the MSS topology, but the old standby Dijkstra (SPF) works better for mean-square (minimum-energy - not a metric) routing. MSS has turned up some new and interesting issues in this area. 4. The NSFNET Backbone is picking up real user traffic and seems to be surviving various infant-mortality problems and training exercises. The officially blessed ARPANET gateway at CMU is not yet up, so the temporary gateways at Cornell and Linkabit are horsing the load, currently about 77000 packets per day through Linkabit alone. Severe problems were occasionally encountered during the month, apparently due to corequakes when the core tables filled up again. Some issues concerning EGP and the dual-connected nature of the system have not yet been resolved. 5. With help from Hans-Werner Braun at the University of Michigan, I developed, tested and deployed, a major change to the fuzzball routing algorithm for use in the NSFNET Backbone. The change 5 involves the distribution of network numbers, as well as delay/offset data, in the hello messages. The result is the elimination of frequently updated, handcrafted tables and the automatic distribution of net-reachability information throughout the amalgamated twenty-network swamp. 6. Bill Nesheim at Cornell, Mike Petry at the University of Maryland and I, cooked up a scheme to splice the fuzzball routing data and EGP. Bill and Mike implemented it under 4.3 bsd and I implemented it in the fuzzball. The implementations are now cooking on cu-arpa.cs.cornell.edu, trantor.umd.edu and dcn-gateway.arpa respectively. We are still investigating the stability issues involved, including such things as the hold-down period, hop-count/delay interactions and so forth. 7. I attended the IAB meeting on 14 August at BBN and the TCP-IP Implementors Workshop on 25-27 August in Monterey, CA. Dave Mills 6 TASK FORCE REPORTS ------------------ APPLICATIONS No report received. END-TO-END SERVICES End2end TF: No report this month. Bob Braden INTERNET ARCHITECTURE 1. I have been appointed Professor in the Department of Electrical Engineering at the University of Delaware, effective 1 September, so this and future reports will originate from there. 2. Issues involved with the recently installed NSFNET Backbone and Multiple Satellite System were submitted for discussion, but activity in the mail group continues at a low ebb. 3. I have invited Greg Lauer to discuss recent thinking in the SURAN routing community at the next INARC meeting in October. I have also invited Len Bosack to discuss issues in advanced routing. It looks like we might lose our eagerly anticipated ANSI/ISO participants as the result of an ISO meeting in Tokyo. Since this is the second time this has happened, I may suggest holding the following INARC meeting in conjunction with the next ANSI or ISO meeting. Dave Mills INTERNET ENGINEERING No report received. INTEROPERABILITY No report received. 7 PRIVACY John Linn distributed copies of a draft RFC, entitled "Privacy Enhancement for Internet Electronic Mail, to the Privacy Task Force Membership", to be reviewed at the September task force meeting at UCL. Guest speakers were recruited for that meeting. John Linn ROBUSTNESS AND SURVIVABILITY No report received. SCIENTIFIC COMPUTING No report received. SECURITY No report received. TACTICAL INTERNET No report received. TESTING AND EVALUATION No report received. 8 -------