11-Jan-82 19:18:50-PST,9695;000000000001 Mail-from: ARPANET host BRL rcvd at 11-Jan-82 1918-PST Date: 11 Jan 82 17:52:48-EST (Mon) From: Mike Muuss To: list: Subject: TCP-IP Digest, Vol 1 #11 Bcc: TCP/IP Digest Monday, 11 Jan 1981 Volume 1 : Issue 11 Today's Topics: Administrivia && The Use of Dot (.) Languages for Implementing TCP-IP? Common Law Copyrights for the Digest? Formal Copyrights for the Digest? && TCP Digest as a Public Forum COMSAT Information Update ---------------------------------------------------------------------- From: Mike Muuss Subject: Administrivia Well, the source of the leak to ComputerWorld has been found, so that we have some breathing space to mull over the implications (it was an ArpaNet recipient, so USENET sites need not worry). It is certainly clear that anything that goes out in a Digest will reach a large audience, not all of whom are involved with the ArpaNet (USENET, for example). At some time, material WILL fall into "outside" hands. The question becomes a choice between: 1) Being more careful, so that anything said is quotable, or 2) Publishing a "restricted rights" notice on the top of the digest as a deterent to re-publication. Under no circumstances do I want to restrict the membership or distribution of this Digest. I hope that we can get over these political problems, and get back to discussing technical issues once again. Thoughts? -Mike ------------------------------ From: ROODE at SRI-KL (David Roode) Subject: use of . Why not use ! instead of . for routing in network addresses. It seems a much wiser choice. ------------------------------ From: cak at PURDUE Subject: Overloading . Many other systems use . as a separator; PhoneNet for CSNET uses it as in cak.purdue@udel, the Berkeley local network can use it, as in csvax.cak, PUP uses it, as in clark.WBST@Parc-Maxc, ad nauseum.... Chris Kent [ I believe that the CSNET usage of User.Host@Forwarder is the RFC733 approved addressing format for sites which can't take User@Host@Forwarder. The choice of the dot does seem rather unfortunate. The British have adopted RF733 for their mail standard, but selected the percentage symbol "%" rather than the dot ".". -Mike ] ------------------------------ Subject: TCP-IP implementations From: AVRUNIN at USC-ECLB In implementing TCP-IP in various computer systems it would be helpful to have an implementation to start with. There seems to be severaly "C" versions. I would like to know what other languages have been used for implementations. I would especially like to know if anyone has or is using Fortran 77 for implementation. Thanks Larry Avrunin [ I believe that the Tektronix implementation for the CDC NOS system is written in RATFOR (Rational FORTRAN). -Mike ] ------------------------------ From: Walt Subject: Re: TCP-IP Digest, Vol 1 #10 To: tcp-ip at BRL Would claiming a common law copyright on this digest stop ComputerWorld from printing quotes? ------------------------------ From: Geoffrey H. Cooper Subject: Re: TCP-IP Digest, Vol 1 #10 If there is really a concern about having TCP-IP entries published on paper-based media, it would help some to put a copyright notice on each digest: (C) Copyright 1982, DARPA, all rights reserved. The material in this digest may not be copied, in whole or in part, for purposes of commercial publication without the written permission of the moderator. Individual sections of the digest may contain copyright notices which supercede this notice. This would at least make editors of computerworld and the like hesitate if given the entire digest. - geof cooper ------------------------------ From: cbosgd!mark at Berkeley To: ucbvax!tcp-ip@Berkeley Subject: TCP digest as a public forum I just want to make sure the people on this list are aware that each TCP digest is fed into USENET on newsgroup fa.tcp-ip. This is sent to (currently) about 120 machines across the US and Canada. (For those who don't know about USENET, it's a distributed bulletin board system.) USENET specifically has a policy that anything posted to net and fa newsgroups is public information that can be redistributed to whoever wants it. The point is that if you have anything you consider secret, it probably shouldn't be mailed to the list. While I am under the impression that this policy is consistent with the intent of the TCP-IP digest, if I'm wrong, it may be necessary to remove the USENET feed from the mailing list. It is possible that ComputerWorld got their information from USENET, but from the wording of the article, they seem to have gotten it from somewhere on the Arpanet. It is easy to confuse private mail and public mailing lists/newsgroups, and it seems clear that the quote from the digest was written in a "I'm talking privately to friends" frame of mind. Clearly he didn't intend his words to be printed in ComputerWorld. But it is important to remember that anything which is mass-mailed is effectively published. Mark Horton ------------------------------ Return-path: Mills@COMSAT Mail-from: [29.4.6.2] received at USC-ISIE 5-Jan-82 11:46:01 From: Mills at COMSAT Subject: IP-TCP Digest info update cc: Mills at ISIE Via: Usc-Isie; 5 Jan 82 18:28-EDT Mike, Following is an extract of a recent report on the status of our IP/TCP implementation which may be of interest to your readers. The COMSAT IP/TCP implementation, which was sponsored by DARPA, was developed over the last three years and used extensively for testing, evaluation and experimentation with other implementations. This implementation runs in a sizable number of PDP11s and LSI-11s with varying configurations and applications. It consists of a package of MACRO-11 modules structured into levels corresponding to local-net, IP and TCP levels, with user interfaces at each level. It is designed to be incorporated either into a special operating system built for a multi-media internet workstation (so-called "fuzzball," based on RT-11 emulation) or into the RSX-11 operating system as part of a package to support the INTELPOST electronic-mail network. The local-net level supports several comunication devices, including synchronous and asynchronous serial lines, 16-bit parallel links and 1822 interfaces. Hosts using this package have been connected to ARPANET IMPs, Satellite IMPs, internet gateways, port expanders and to the DCNET local network. It provides optional automatic routing, time synchronization and error-reporting functions. The IP level conforms to the RFC-791 specification, including fragmentation, reassembly, extended addressing and options, but currently does not interpret options. A full set of ICMP features compatible with RFC-792 is available, including destination-unreachable, timestamp, redirect and source-quench messages. Support for an IEN-142 compatible time server and an IEN-109 (as amended) compatible internet gateway can be included on an optional basis. The internet-gateway support can be configured to include hierarchically structured gateways and subnets. Destination-unreachable and source-quench information is conveyed to the user level via the TCP and raw-datagram protocol modules. The TCP level conforms to the RFC-793 specification, including PUSH, URGENT and options, but currently does not interpret options. Its structure is based on circular buffers for reassembly and retransmission, with repacketizing on each retransmission. Retransmission timeouts are dynamically determined using measured roundtrip delays, as adjusted for backoff. Data flow into the network is controlled by measured network bandwidth, as adjusted by source-quench information. Features are included to avoid excessive packet fragmentation and retransmission into zero windows. The user interface level provides error and URGENT notification, as well as a means to set outgoing IP/TCP options. A raw-datagram interface is available for XNET (IEN-158), UDP (RFC-768) and similar protocols. It includes internal congestion and fairness controls, multiple-connection management and timestamping. A number of user-level protocol modules have been built and tested with other internet hosts, including user/server TELNET (RFC-764) user/server FTP (RFC-765), user/server MTP (RFC-780) and various utility file-transfer, debugging and control/monitoring protocols. Code sizes and speeds depend greatly on the system configuration and features selected. A typical 30K-word LSI-11/2 single-user configuration with all features selected and including the operating system, device drivers and all buffers and control blocks, leaves about 16K words for user-level application programs and protocol modules. A typical 124K-word LSI-11/23 configuration provides the same service to a half-dozen individually relocated users. Disk-to-disk FTP transfers across a DMA interprocessor link between LSI-11/23s operate in the range 20-30 Kbps with 576-octet packets. The 124K-word PDP11/34 INTELPOST adaptation supports two 56-Kbps lines and a number of lower-speed lines. Further information is available from D. Mills (Mills@ISIE) on the IP/TCP implementation and D. Jupin (Jupin@ISIE) on the RSX-11 adaptation. Regards, Dave Mills END OF TCP-IP DIGEST ********************