[IMR] IMR87-08.TXT Westine [Page 1] ~ AUGUST 1987 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@SH.CS.NET).. BBN LABORATORIES AND BBN COMMUNICATIONS CORPORATION --------------------------------------------------- WIDEBAND NETWORK A new Wideband Butterfly Gateway was installed at ESD/Mitre in Bedford, MA on August 7. The gateway connects an ethernet at an ESD/Mitre testbed facility to the Wideband Network via a terrestrial T1 circuit terminating at the gateway and at the Lincoln Laboratory BSAT. This new connectivity has been established in support of ESD's Technology Validation Experiment (TVE). The goal of the TVE is to demonstrate that a large computer-based simulation can be run in a distributed computation environment provided by Cronus Distributed Operating System Westine [Page 1] Internet Monthly Report AUGUST 1987 "clusters" located at RADC and at BBN as well as at ESD/Mitre. Relatively high bandwidth data paths will be required for the support of the simulation's wide area intercluster communications. The Lincoln-ESD/Mitre T1 circuit, along with Wideband Network connections currently in place at RADC and BBN, will be used to meet this requirement. BSAT software Releases 5.4 and 5.5 were distributed to the Wideband Network sites during the month. The new software supports simultaneous monitoring of the network by multiple host computers, an increase in the maximum size of individual channel capacity reservation requests, and a reservation request throttling mechanism providing BSAT satellite channel module congestion control. The BSAT is being converted to run under Chrysalis Operating System Release 3.0. Chrysalis 3.0 provides a "RAMFile" system which is being investigated for potential use in a new BSAT boot procedure. SATNET The SATNET remained stable during the month of August. We did experience problems in the links between the SIMPs and the Butterfly gateways. The link between the DCEC gateway and the Roaring Creek SIMP went down on August 14th due to phone line problems. It remained down until August 17th. Although the CSS gateway link to the SATNET reported up, European traffic could not get through. The problem was not detected until the loss of the DCEC gateway. We are working to try to isolate the problem and have implemented a check to notify us if it occurs again. This will allow us to troubleshoot the problem before users are affected. Hawley Rising and John Leddy attended a meeting at COMSAT in Clarksburg Maryland on August 27th. Ken Kay discussed maintenance of the Linkabit modems. There is one Linkabit modem that remains to pass acceptance testing. When the complete shipment is accepted we will be able to replace the necessary hardware at affected sites. INTERNET R&D Our work on the SURAP implementation for the Butterfly Gateway is proceeding. We have started testing the code with a real LPR. We are completing a release (3.10) which will fix a number of outstanding bugs. These include strict source route, Satnet restart problem, and a Wideband gateway problem. This release Westine [Page 2] Internet Monthly Report AUGUST 1987 which will go in all of the Butterfly Gateways and will include support for fragmented EGP updates. We expect to begun fielding this release by the middle of September. Bob Hinden (Hinden@BBN.COM) ISI --- Internet Concepts Project Greg Finn continues to work on the vulnerability of dynamic computer network protocols. One RFC was published this month. RFC 1000: Reynolds, J.K. and J. Postel "The Request for Comments Reference Guide", August 1987. Ann Westine (Westine@ISI.EDU) Multimedia Conferencing Project Brian Hung is working on a conferencing mode of his IBM-PC scanner program, i.e., one that would allow a user to send bitmap messages to a destination during a conference. This version of Brian's program would require the incorporation of TFTP into his main Pascal program. Brian is currently attempting to do this via an assembly language routine. This routine would call the MS DOS Exec function call to invoke TFTP. Brian Hung (Hung@ISI.EDU) We are working on extending the packet video conferencing system from a point-to-point connection between two sites to be a multipoint connection among two or more sites. An interim version of the ST conferencing protocol has been implemented in the Packet Video Program; BBN has implemented the Voice Funnel side of the protocol. Testing is underway. Since there are only two copies of the prototype packet video hardware, initial testing will simulate three-site video by having one site connect back to itself while also connected to the second site. A fake video source will be used to test other aspects of the system. In the fall we plan to convert to commercial video hardware so that more than two systems can be installed. Steve Casner (Casner@ISI.EDU) Westine [Page 3] Internet Monthly Report AUGUST 1987 NSFNET Project Bob Braden organized and attended a Technical Workshop hosted by the Cornell Theory Center on August 26, 27, and 28. This workshop was specifically limited to technical personnel responsible for the various middle-level networks of NSFNET; its purpose was to pass on to the attendees as much information as possible about how to run a network within NSFNET. The 40 attendees included technical personnel from BARRNET, MERIT, MIDNET, NCAR, NCSA, NORTHWESTNET, NYSERNET, PSCNET, SDSCNET, SESQUINET, SURANET, and WESTNET. We were extremely fortunate to have the two Internet notables Vint Cerf and Dave Mills among our speakers. Other major contributors to the class were Hans-Werner Braun, Mike Petry, Scott Brim, Mark Fedor, and Bob Braden. Featured in the program were detailed discussions about problem diagnosis, GATED, and routing issues. Marty Schoffstall and Jeff Case presented a demonstration of preliminary SGMP implementations. At ISI, internal testing of the background file transfer program (BFTP) continued. This resulted in the implementation of a number of improvements and several bug fixes. Work on the RFC continues. We also started work on a general tool for gathering comprehensive statistics at the Backbone and regional network nodes, for use in future management and engineering of the various components. We plan to exploit the use of Ethernets as the standard interconnection medium between NSFNET levels, by gathering statistics using Statistical Agent hosts which are monitoring these Ethernets promiscuously. The first version is being implemented on a SUN workstation, but we plan to implement a later version in PC's. To support high packet rates, these Statistical Agent hosts may have to be dedicated to the statistics function. In addition to the Statistical Agent hosts, one per interconnection node, there will be a central Statistical Collector program which queries the agents periodically and maintains a long-term database. By gathering the data outside the actual gateways, we expect to be able to provide comprehensive statistics at all protocol levels, without adding to the CPU or memory load on the gateways. The collection mechanism is being designed to be dynamically configurable and easily extensible, using an object-oriented approach. Bob Braden and Annette DeSchon (Braden@ISI.EDU and DeSchon@ISI.EDU) Westine [Page 4] Internet Monthly Report AUGUST 1987 Supercomputer and Workstation Communication Project Alan Katz continued to experiment with DEC's Xt X-Windows Toolkit. Alan was able to get a their new version of xterm which was supplied with the toolkit running on both the VAX and the Suns (there was a bug which prevented it from running on a Sun). Alan Katz has been writing a variety of small utility programs using the toolkit to improve his X-Windows user environment. Alan visited people at NASA Ames research center in San Jose, CA, Aug 3rd, about the possiblity of establishing a Wideband-Net link to the NAS Cray 2's. Alan Katz (Katz@ISI.EDU) As August drew to a close, a new version of the Wideband Net BSAT software was released that removes the limit of 2^14 channel symbols per frame for aggregation of datagrams to be transmitted. This should allow us to test NETBLT again and achieve rates above 1Mb/s. Steve Casner (Casner@ISI.EDU) MIT-LCS ------- The current NETBLT protocol implementation has now been tested over the ARPANET with success. Strict source routing was used to generate several paths across the ARPANET from MIT to USC-ISI and back again, with throughputs averaging 16,000 bits per second and peaking at 19,000 bits per second. The tests were conducted at several times during the day: 10AM EDT, 11AM EDT, 12N EDT, and 5PM EDT. Network round-trip delay and packet loss rates were quite high during some tests; NETBLT was nevertheless able to provide high throughput, demonstrating again that NETBLT can provide high performance over many network types. Lixia Zhang (LIXIA@XX.LCS.MIT.EDU) MITRE Corporation No report received. Westine [Page 5] Internet Monthly Report AUGUST 1987 NTA & NDRE ---------- No report received. SRI --- Multimedia Conferencing/Wideband Network The first demonstration of EMCE (Experimental Multimedia Conferencing Environment) over the Wideband SATNET took place between ISI and SRI on August 18, 1987. Keith Williams and Steve Casner, at ISI, talked with Ruth Lang, at SRI, for approximately one hour with no breaks in communication. All aspects of the conferencing tool were successfully exercised. End-to-end packet delays were approximately one second, mostly due to the satellite transport. Voice quality was as good as using the Ethernet alone. Ruth Lang (ruth@tsca.istc.sri.com) Keith Williams (keith@tsca.istc.sri.com) Distributed Applications Replicated Databases The TACTICS II distributed system (the system we first reported last month, which is under the ADCCP project sponsored by CECOM) supports reliable replication in spite of internetwork partitioning. In the case of partitioning, replication is accomplished automatically upon reconstitution. The current mechanism for replication is based on queueing of database updates with a primary site responsible for replicating information to other databases. This design clearly suffers from being vulnerable to single points of failure, namely the primary site and its queues. We are currently investigating alternative replication mechanisms less vulnerable to single points of failure. The investigation focuses on the use of optimistic replication strategies for database updates and semantic-based consistency techniques. This alternative design aims at achieving real-world consistency, vesus time-sequence consistency, of database information. Resource management The TACTICS II resource monitor is used as an integral part of a testing facility for experimenting the TACTICS II system Westine [Page 6] Internet Monthly Report AUGUST 1987 in several internet configurations. An experiment was conducted over packet radio and local area networks. The goal of the experiment was to demonstrate the feasibility and capabilities of testing distributed system designs and implementations in an internetwork environment. The resource monitor was used to collect traffic, usage pattern and the like real-time information of the networks, the experimental hosts, and the services on those hosts. The data collected by the resource monitor was stored and displayed along with other system information over the course of the experiment. Joan Wrabetz (wrabetz@spam.istc.sri.com) UCL --- Work on a learning system for protocol performance tuning is well under way, with a simple learning shell now being tried out on some TCP parameters and a simple model of the internet. The intention is to glue this to the 4.3BSD TCP and compare empirical with simulated results. John Crowcroft (jon@NSS.CS.UCL.AC.UK) UNIVERSITY OF DELAWARE ---------------------- 1. Development continues on the Dissimilar Gateway Protocol (DGP). Mike Little and I worked out a DGP model based on the constraints of existing host and gateway implementations and another model based on selected changes to these implementations. 2. A driver for the ACC 5250 X.25 interface was constructed for the fuzzball and installed on the swamp.arpa gateway. The intent of this implementation is to measure the overheads and loss rates of the DDN Standard X.25 protocol in actual service, as well as assess the performance of the ACC hardware itself. 3. I helped Steve Casner of ISI bring up a new WWVB clock, which is now ticking at 128.9.2.129. There are now two WWVB reference clocks in the western US (NCAR and ISI) and two in the eastern US (M/A-COM and UMD), as well as a GOES reference clock (Ford) and assorted WWV clocks (UMich, UDel). The San Diego Supercomputer Center (SDSC) threatens to bring up another WWVB clock as well. Westine [Page 7] Internet Monthly Report AUGUST 1987 4. Some digging into the time-synchronization mechanisms of AT&T and other common carriers disclosed interesting parallels between their network- synchronization systems and the ad-hoc systems now budding in the Internet. I am working on a report discussing these issues and how Internet clocks can be further improved. 5. I continued to watch the NSFNET Backbone for throughput and loss data. Problems with primary/fallback routes continue to bother the system. 6. Mike Minnich, having attacked several megabytes of performance data harvested from the NSFNET Backbone, captured further megabytes of performance date from the BBN LSI-11 gateway weekly traffic reports. Mike is developing flow equations in order to measure overheads, mean flight paths and other statistics useful to assess the performance of the routing and congestion-control algorithms. 7. I presented a briefing for the NSFNET Road Show at Cornell on 26-27 August and prepared a paper on the NSFNET Backbone network for the ACM SIGCOM Symposium on 11-13 August. Dave Mills (Mills@UDEL.EDU) Westine [Page 8] Internet Monthly Report AUGUST 1987 NSF NETWORKING -------------- NSF NETWORKING UCAR/BBN LABS NSF NETWORK SERVICE CENTER (NNSC) Craig Partridge presented a paper on round-trip time estimation and spoke in a session on NSFNET at the SIGCOMM conference. The implementation of the HEMS agent software (for network management) is currently being debugged. Approximately 2500 copies of the NSF Network News have been distributed. By Karen Roubicek (roubicek@nnsc.nsf.net) NSFNET BACKBONE SITES CORNELL UNIVERSITY THEORY CENTER Backbone Operations Sixty-four million packets were delivered by the backbone in August -- double what we delivered in May. We couldn't have done this well without heroic and ingenious software efforts by Dave Mills to deal with congestion (as described in the paper he and Hans-Werner Braun gave at SIGCOMM). Also, the software in the fuzzballs now allows for 120 networks in the routing table and multi-packet routing updates. Westine [Page 9] Internet Monthly Report AUGUST 1987 Statistics (from Doug Elias): Total Traffic Figures Between Routers Ethernet Input 100067210 60405859 Output 98589040 64139840 In+Out 198656250 124545699 Grand 323201949 Traffic Delivered to Ethernets min mean max total % PSC 1452 24276.22 63819 16969080 25.44 * Cornell 282 11321.60 197799 7913799 11.87 * JvNC 1364 29185.45 93200 21101079 31.64 * NCAR 82 12647.67 53718 8840724 13.26 * SDSC 4 3645.39 22760 2548131 3.82 SURA 0 3657.52 29752 2552946 3.83 * NCSA 682 9681.01 39906 6767027 10.15 Overall 66692786 100.00 The "*" indicates statistics based on data containing "missing observations"; these cause the mean to be artifically elevated, but by no more than ~2%. By Scott Brim (swb@tcgould.tn.cornell.edu) UNIVERSITY OF ILLINOIS AT URBANA-CHAMPAIGN We lost a delni into which the fuzzball and vitalink stuff were plugged. For some amount of time (ending at 1627Z) all connections to campus from NSFNET and Uchicago, Indiana University, and Northwestern and between those schools and NSFNET was unavailable. By Ed Krol (krol@uxc.cso.uiuc.edu) JOHN VON NEUMANN NATIONAL SUPERCOMPUTER CENTER This report is designed to inform the JVNC Consortium and JVNCNET network members as well as the Internet community of monthly status of the JVNCNET network. The data used on this report is collected using a number of techniques developed at JVNC, together with data from the JVNC operations group. The NSFNET data is obtained from the NSFNET NOC (Network Operations Center). Westine [Page 10] Internet Monthly Report AUGUST 1987 NETWORK BRIEF: JVNCNET has the configuration of a "tree". The 13 Consortium Universities*, together with the 3 NRAC** schools form the 16 node network. The gateways are 15 VAX's and 6 Ungermann Bass (UB) routers, connected by 9 T1 lines, 5 56kbps lines and 2 satellite links. Locally at JVNC (center of the JVNCNET network), JVNCA (a VAX8600 running ULTRIX) serves as major router, primary name server for domain "csc.org" and primary network monitoring system, JVNCB (a VAX750, ULTRIX)is also a gateway, finally we utilize a dedicated UB router to connect to the NRAC group. NETWORK STATISTICS DATA: The network monitoring statistics are performed from "jvnca", and are affected by its down time. In order to compensate for this we multiply (in our analysis below), the numbers with the percentage of down time of jvnca thus representing the worst case. We plan to overcome this situation with the next release of the JVNC monitoring package (see "Network Monitoring Project" below). The network is polled every 10 minutes with "icmp-echo" packets, and the analysis of the data is performed on this data. The data is processed using information on the "scheduled down time table" for the systems. Therefore, if a gateway is scheduled to be down, its information is not computed during that time, whether the gateway is operational or not. The special configuration of this network together with the fact that we use different subnets for the point to point links allows us to determine very accurately the reachability for the gateways. When we receive an alarm that one node/gateway is unreachable, we proceed to determine whether the node is down or any other component of the the link is non functional. This information is not affected much by the routing. Westine [Page 11] Internet Monthly Report AUGUST 1987 TABLE I, Gateways Reachability/Link Status ------------------------------------------------------------------------- Created: Sun Aug 30 21:38:23 1987 ------------------------------------------------------------------------- node meanttr maxttr meantbf sched Up Dwn avail perf min min min % % % % ------------------------------------------------------------------------- ias 197 467 1030 2 93 6 93 95 mit 43 43 1446 0 99 0 99 96 nyu 43 43 1446 0 99 0 99 98 njit 71 310 1257 1 98 1 98 96 brown 24 43 1349 0 99 0 99 99 umdnj 78 280 1204 1 98 1 98 90 arizona 80 570 1009 1 97 2 97 97 harvard 43 43 1446 0 99 0 99 90 rutgers 19 20 1092 0 99 0 99 99 stevens 71 310 1257 2 98 1 98 96 colorado 26 43 1383 2 99 0 99 93 columbia 244 1429 1006 0 88 11 88 88 princeton 121 650 1105 0 97 2 97 97 rochester 30 59 1269 0 99 0 99 95 u_of_penn 41 40 1381 0 99 0 99 99 penn_state 275 539 1337 0 96 3 96 96 jvncb.csc.org 19 43 1299 3 99 0 99 99 Total test time (min): 33303 ------------------------------------------------------------------------ The Header definitions of table I are: nodename: the gateway or host on JVNC-NET network (128.121) meanttr: in minutes, the mean time to recover from the "down" state to the "up" state. Where "down" state is when the result of sending "icmp-echo" packets is no packets return, and the "up" state is when we receive packets back. Each test is performed 10 times, every 10 minutes, and averaged each time. maxttr: in minutes, the maximum time to recover, from the "down" state to the "up" state (see above). meantbf: in minutes, the mean time between failures. sched_down time: in percent, is the time the gateways/hosts were "scheduled" to be down, respect to the total time of test. avail: in percent, is the time for which the gateways were available ("up" state) respect to the total time of test minus the scheduled down time). perform: in percent, is the figure of merit that considers the Westine [Page 12] Internet Monthly Report AUGUST 1987 number of packets lost and the available time. ANALYSIS: The data of Table I is a product of the gateways availability together with lines/satellite status (this last including the communications equipment such as T1 muxes, CSU/DSU, etc). The following is broken down in gateways and links ( lines and satellite). 1.- Gateways Availability: Considering that "jvnca" was up 97.87% of the time, and the node that was unavailable for the longest time (while jvnca was up) was Columbia, with 88%, then we can safely say that Columbia was up at least 86.13% of the time, representing the worst case on the network. In the average, the gateways were up and available at least 95.24% of the time (worst case). Table II, Gateways Availability ------------------------------------------------------------------------- gateway node problems ------------------------------------------------------------------------- jvnca JVNC buffer starvation, power hit, gated jvncb JVNC power hit, gated fuzzball JVNC power hit, upgrade of fuzzware jvax ARIZONA power problems in Arizona jvnc-njit JVNC software upgrade njit-jvnc NJIT software upgrade njit-umdnj NJIT software upgrade njit-stevens NJIT software upgrade umdnj-njit UMDNJ software upgrade stevens-njit STEVENS software upgrade, routing-rip iasvax IAS construction work in IAS psugate PENN STATE routing-gated hucsc HARVARD routing-routed super-fs RUTGERS storm related power failure ------------------------------------------------------------------------ 2.- Lines Status: Table III, Line Status ------------------------------------------------------------------------- from to type problem ------------------------------------------------------------------------- JVNC PENN STATE T1 T1 mux hang JVNC ROCHESTER 56kbps AT&T problem upstate NY JVNC PRINCETON T1 T1 mux lost power unit JVNC COLUMBIA 56kbps flaky line, AT&T JVNC U. OF PENN T1 glitch on line took route down ------------------------------------------------------------------------- Westine [Page 13] Internet Monthly Report AUGUST 1987 3.- Satellite Nodes Status: Table IV, Satellite nodes Status -------------------------------------------------------------------------- from to type problem -------------------------------------------------------------------------- JVNC ARIZONA 56k sat. bad weather JVNC COLORADO 56k sat. bad weather -------------------------------------------------------------------------- 4.- Traffic: The traffic that transits JVNCNET is a combination of JVNC supercomputer traffic and traffic "in transit". The JVNC supercomputer traffic is the one that has the JVNC front ends (jvncc, jvncd or jvncf) as source/ destination, while the traffic in transit is the rest. The total traffic is: T = traffic on subnet 50 (T1) + traffic on subnet 51 (T2) + + traffic between the MIT node and the Harvard node (T3) * + + traffic between the MIT node and the Brown node (T4) * + + traffic between the NJIT node and the Stevens node (T5)* + + traffic between the NJIT node and the UMDNJ node (T6) * where: T1 = ( jvnca + jvncb + jvncc + jvnc-njit + colo + jvax + fuzzball + + term-serv)|subnet 50 T2 = ( jvnca + jvncd + jvncf + picasso + monet + iris + term-serv1 + + term-serv2 + term-serv3)|subnet 51 T3 = |coventry|subnet_3 - coventry|subnet_9| T4 = |coventry|subnet_3 - coventry|subnet_10| T5 = |njit-jvnc - njit-stevens| T6 = |njit-jvnc - njit-umdnj| * traffic not seen on subnet 50 Until now, we have been reporting |T1|max (see above) which is the maximum value of T1_input and T1_output (the reason for the difference in values is simply that we are not gathering information for all the hosts/gateways). This month's count is: number of packets on subnet 50 (one direction, T1) > 105,948,605 =========== Note that this is a 60% increase respect of last month. --- The data is collected from jvnca, jvncb and JVNC_fuzzball every 10 minutes automatically and compiled each month. The data from Colo, JVAX, jvnc-njit is only partial. The data for jvncc and the terminal server on subnet 50 is not present. Note that this figure would be Westine [Page 14] Internet Monthly Report AUGUST 1987 largely incremented if we would consider jvncc's statistics as well. Table V, Traffic on JVNCNET (subnet 50 only) ------------------------------------------------------------------------- local gateway packets in packets out comments ------------------------------------------------------------------------- JVNCA 40,312,393 48,900,159 accurate JVNCB 28,275,959 33,306,302 accurate COLO 1,117,953 5,691,597 only last 18 days JVAX 111,463 231,809 only 7 days JVNC FUZZ 18,536,320 15,812,886 accurate JVNCC - - not available JVNC-NJIT 3,850,500 2,005,852 only 23 days TERM SERV - - not available total ........ 92,204,588 105,948,605 ------------------------------------------------------------------------ 5.- PSN Status: Still no news from DCA when they will connect. Everything is ready. 6.- Routing: UB routers: new version of the software, RIP between UB routers and "external" gateways replace previous routing reachability via EGP between JVNCB and the JVNC-NJIT router. VAXs: new version of gated, hello protocol running in broadcast form between local VAXs and the fuzzball has SOLVED the route flapping due to the ptp helloers and RIP broadcasts. 7.- Comments: The network routing has behaved better this month, after the upgrade of gated. The routing on the dedicated routers has improved too with the new code allowing for more dynamic routing as well as more control. The traffic has increased at least 60%. And the uptime for the systems has remained high. The projects are moving ahead. PROJECTS: 1.- Network Monitoring: Status: On going The network monitoring package "netmon" has been designed to provide JVNCNET with the tools to quickly determine a network problem in our Westine [Page 15] Internet Monthly Report AUGUST 1987 environment. This package has been in constant evolution as we have learned more on its limitations and the needs for monitoring. Today the so called "phase I" is completed and the program is fully operational. Still, "phase II" is on its way, to provide for new functions like the concept of "primary" and "secondary" monitoring center, at the same time "phase II" will allow for portability of the code, this will satisfy some of the JVNC Consortium's requests for copies of the program for their local use. Unfortunately we can only support the code for local use due to staff limitations. We are also looking very attentively at the efforts of the IETF's gateway monitoring group (chaired by Craig Partridge), and we'll follow their recommendations to enhance our monitoring facilities. 2.- Network Characterization: Status: On going The Network Characterization program is directed towards determining the parameters that characterize the JVNCNET network's diverse type of services. This effort will be utilized not only as a research subject but to find/predict network bottlenecks and problems before they are obvious to the end users. This task started two weeks ago with the collection of data and the automation of the collection, and will continue with the study of the characteristics and patterns that distinguish each point to point lines whether T1, 56kpbs or satellite. The results will be available to the community. 3.- Traffic Analysis: Status: On going The traffic data is currently being collected from JVNCA, JVNCB, and the fuzzball automatically. This will be extended to the other VAXs on the same ethernet and the other routers on the network. At the same time the data will be compiled per campus. INFORMATION: JVNCNET NOC: "net@jvnca.csc.org" (JVNCNET Network Operations Center) JVNCNET manager: "heker@jvnca.csc.org" * CSC Consortium: Princeton University, MIT, Harvard, Brown, University of Pennsylvania, Rutgers, IAS, Columbia, University of Rochester, NYU, Penn State, University of Arizona, University of Colorado. Westine [Page 16] Internet Monthly Report AUGUST 1987 ** NRAC (Newark Remote Access): New Jersey Institute of Technology (NJIT), Stevens Institute of Technology, University of Medicine and Dentistry of New Jersey (UMDNJ). By Sergio Heker (heker@jvnca.csc.org) NATIONAL CENTER FOR ATMOSPHERIC RESEARCH AND UNIVERSITY SATELLITE NETWORK PROJECT Preliminary statistics show that the pass-through traffic on USAN (i.e., traffic not destined for NCAR) is about 60%. Further breakdowns of the traffic flows will be made with the recently acquired Excelan Lanalyser. With the exception of the one host (the only host) on USAN, windom.ucar.edu (which runs gated), Wisconsin, and the NSFNET fuzzball, the USAN/UCAR nets use static routing. We plan to phase out static routing by bringing up each gateway it turn with dynamic routing. The planned USAN/Institute for Naval Oceanography connection has been postponed until at least FY 88. The USAN/Naval Research Lab connection is scheduled for September 20. By Don Morris (morris@scdsw1.ucar.edu) PITTSBURGH SUPERCOMPUTER CENTER Our Fuzzball has had no downtime except for one reboot for a new version on August 5 and one short line outage. The fuzzball has been passing on the order of half a million packets daily to the local net. PSC Gateway was also up the entire month, forwarding nearly 1.25 million packets daily. The newest configuration and newest version of gated have been running since August 20. A new PSN arrived here on August 3, and its installation was completed by BBN on August 5. A satellite dish and associated equipment have also been delivered but are not yet installed. Lines have been ordered, but these too have not yet been installed. By Dave O'Leary (oleary@morgul.psc.ed) Westine [Page 17] Internet Monthly Report AUGUST 1987 SAN DIEGO SUPERCOMPUTER CENTER Our PSN (IMP #26) was installed during the month. At this time it has no trunk connections; neither VSAT (U Wisc) nor terrestrial (ISI, UCLA, NASA/AMES). The T1 microwave link to the Salk Inst. (p4200 to p4200) is now operational. It is running version 7.3 and supports both TCP/IP and DECNET. While we are looking forward to 7.4 to correct a few "hitches" in the DECNET support, by and large we are pleased with the link. The link with BARRNET (at UC Berkley) is in place and tested. It will be turned up on Tuesday, 8 Sept. This is also a p4200 to p4200 connection. JPL has elected to use this as their path to SDSC (and NSFNET). The use of an existing 9.6 line between JPL and SDSC, mentioned 2 months ago will not be used. We have received our GATED configuration file from Cornell and it is now running on a SUN 3/50. It will be transferred to a VAX (SRI Multinet) with the next release from SRI - this will bring the SRI version of GATED inline with Cornell's. Our Pacer/Kinetics-SC combination is now stable - ~60 MAC's will be coming online via Ethernet over the next 2 weeks. Locally, work is progressing on IP traffic over SDSCNET links. We also have UDP services working to our Cray. DEC has delivered the VAX 8250 and DMB32C for our VAX-BI/Cray-CTSS development. By Paul Love (loveep@sds.sdsc.edu) NSFNET REGIONAL AFFILIATED & CONSORTIUM NETWORKS BARRNET Missed report for July as Tom Ferrin was too busy to take his turn at writing it. Little or no new activity in August, lots of people on vacation. UC Berkeley split its 112kb link to SDSC in preparation for activation of BARRNET-NSFNET link into Berkeley using one of the two 56kb paths. The connection will be tested in early September. It will be assigned (tentatively) class C address 192.31.52.xx. We continue to operate all BARRNET links at 1.5+Mb rate on T-1 facilities but are experiencing high error rates on one (UC owned) microwave link (UC Berkeley to UC San Francisco) and began near month end to experience high error rates on the telco supplied UC Santa Cruz - Stanford link which had been very clean for previous Westine [Page 18] Internet Monthly Report AUGUST 1987 month. UC plans replacement of the current microwave equipmentin the San Francisco link which is a T1C channel with a mux which takes special diagnostic equipment for trouble shooting more difficult. The Santa Cruz link problem is being studied. A T-1 BERT tester is a necessity, but even that is of limited help with the "high capacity" telco facilities in the path (44+Mb/s fiber). We interfaced Grafpoint Tektronix 4xxx graphics terminal emulator package to the Stanford version of PC/IP for direct graphic output from the Cray at NASA via telnet. The Stanford Mac/IP package v.2.0p is now in production in many sites and runs on Mac2's as well as 512s, Plus and SE. BARRNET Consortium and Technical Subcommittee meetings are scheduled for September. A key subject of discussion will be network expansion and new members. There are about half a dozen sites which have expressed a desire to connect and the network operation is nearing sufficient stability to entertain such expansion. By Bill Yundt (gd.why@forsythe.stanford.edu) JVNCNET (Refer to JVNNSC backbone report) MERIT (No report received) MIDNET As of the end of August nine of the twelve MIDNET links were up and running. The other three links are waiting for delivery of the CSU/DSU's that connect the phone lines to the routers. Unfortunately, two of the remaining links are the ones into our backbone connection, NCSA. By Doug Gale (doug@unlcdc3.bitnet) NORTHWESTNET (No report received) Westine [Page 19] Internet Monthly Report AUGUST 1987 NYSERNET As of 1 August 1987, NYSERNET had the following topology: Clarkson Syracuse--+ | | | Rochester--------Cornell---------RPI---Albany | | | Buffalo...Fredonia | | | | | Binghamton | +-------- | ------StonyBrook | | | | | | | | CUNY--NYTEL/NSMAC--Columbia------NYU-+ | |\ | | | /| | | | \ NYTEL/GC | NYNEX/S&T / | | | | \ BNL / | | | | \ / | | | | +-------------Rockefeller | | | | | | | +------------------------------+ | | | +-------------POLY-------------------+ C.NYSER.NET was enabled as a root domain server as part of the "2nd-root" group with Maryland and NASA. After a testing period these machines should be moved into the set of official root domain servers. A number of NYSERNET people participated in the NSFNET workshop at Cornell. Live demo's of SGMP were show for UNIX, MSDOS and Proteon's gateway. By Marty Schoffstall (schoff@nic.nyser.net) SDSCNET (Refer to SDSC backbone report) Westine [Page 20] Internet Monthly Report AUGUST 1987 SESQUINET The complete initially proposed configuration of SesquiNet has been installed and in operation for a full month now. As a consequence, the following campus networks are being served and being advertised via EGP to the core: Baylor College of Medicine 128.249 Houston Area Research Center 192.31.87 Rice University 128.42 Texas A&M University 128.194 Texas Southern University 192.31.101 and the University of Houston 129.7 Although the network is lightly loaded, we have still experienced no hardware or software failures of our gateways. Work is continuing on the connection from NSFNET/NCAR to SesquiNet/Rice via fuzzballs at the two sites. We are now using recent cisco support for the Hello routing protocol. This allows SesquiNet's cisco gateway at Rice to dynamically advertise its campus networks to our ARPANET/NSFNET fuzzball gateway. We have monitored network operation during the month of August, and will soon be analyzing it to compute our MTBF and MTTR. By Guy Almes (almes@rice.edu) SURANET The dates for the installation of the new SURANET data lines as presently projected, by the carriers, are as follows: U of Virginia to CEBAF- Sept 11 U of Maryland to U of Virginia- Sept 15 NSF to Georgetown U- Sept 16 U of Alabama at Birmingham to U of Alabama at Tuscaloosa- Sept 16 Georgetown U to Catholic U- Sept 16 Catholic U to Gallaudet- Sept 16 Gallaudet to George Washington U- Sept 16 George Washington U to George Mason U- Sept 16 Westine [Page 21] Internet Monthly Report AUGUST 1987 CEBAF to William and Mary- Sept 16 U of Maryland to Johns Hopkins U- Sept 16 Vanderbilt U to U of Tennessee- Sept 16 U of Maryland to U of Florida- Sept 21 Florida State U to Louisiana State U- Sept 25 University of Alabama at Birmingham to Louisiana State University- Sept 25 TUCC to CEBAF- Sept 25 Three additional Federal Agencies are planning to connect to SURANET through the University of Maryland. They are: the National Institutes of Health, the National Bureau of Standards and the Naval Research Laboratory. By Jack Hahn (hahn%umdc.bitnet@wiscvm.wisc.edu) WESTNET 1. cisco has been finalized as the vendor for IP Routers. 2. We are soliciting bids for CSU/DSUs. 3. NCAR is advertising Supernet (the Colorado network) to NSFNET. 4. Four Westnet technical representatives attended the Cornell TCP/IP workshop, and found it to be excellent. 5. We are still waiting for our FY'87 funding from NSF, which is expected about 9/15/87. After the funding is in, we will order equipment for the three New Mexico schools, and the University of Wyoming, as these sites will have circuits provided as cost sharing. By Pat Burns (pburns%csugreen.bitnet) Westine [Page 22] Internet Monthly Report AUGUST 1987 TASK FORCE REPORTS ------------------ APPLICATIONS - USER INTERFACE Nothing much to report. Chris Schmandt has prepared a (strawman of the) strawman voice server specification for submission to the 2nd Conference on Computer Workstations; it will be available soon to interested parties. The next meeting of the task force has been rescheduled to October 20- 21, still in Cambridge. Charlotte I. Tubis (tubis@purdue.edu) AUTONOMOUS NETWORKS The Autonomous Networks Task Force will meet in Boston November 5-6. Representatives from the ICCB and ANSI Working Group on Routing Architectures will attend to discuss research issues of mutual interest. Deborah Estrin (estrin@OBERON.USC.EDU) END-TO-END SERVICES No progress to report this month. Bob Braden (Braden@ISI.EDU) INTERNET ARCHITECTURE An announcement for the next INARC meeting workshop is in the works. The current plan is to hold this meeting in the Washington, DC, area in early November. Dave Mills (Mills@HUEY.UDEL.EDU) INTERNET ENGINEERING No report received. Westine [Page 23] Internet Monthly Report AUGUST 1987 INTERNET MANAGEMENT The chairman of this task force met with and briefed the results of the May 87 meeting to Dr. Gordon Bell, Director of the Computing, Information Science and Engineering Division of NSF. A subset of this material is to be provided to Dr. Jack Schwartz, the new director of DARPA, Information Systems Technology Office on 9 Sept. Vint Cerf (Cerf@A.ISI.EDU) PRIVACY No reportable Privacy Task Force progress occurred during August. John Linn (Linn@CCY.BBN.COM) ROBUSTNESS AND SURVIVABILITY No report received. SCIENTIFIC COMPUTING 1. Three white papers are currently in preparation and are hoped to be completed in draft form by the end of September. These will cover high bandwidth requirements, user interface, and equation standards. 2. NASA has initiated a program, called the Telescience Testbed Pilot Program, to use a rapid-prototyping user-oriented testbed approach to develop requirements and explore potential design concepts in support of telescience. It is expected that this program will provide considerable input to the task force. For further information, contact Maria Gallagher (Testbed Coordinator) at Maria@riacs.edu. 3. The task force meeting tentatively scheduled for September has been postponed to allow completion of the draft white papers. Barry Leiner (leiner@ICARUS.RIACS.EDU) Westine [Page 24] Internet Monthly Report AUGUST 1987 SECURITY No report received. TACTICAL INTERNET No report received. TESTING AND EVALUATION No report received. Westine [Page 25]