18-Jan-82 21:49:32-PST,15354;000000000001 Mail-from: ARPANET host BRL rcvd at 18-Jan-82 2148-PST Sender: Mike Muuss From: TCP-IP at BRL To: TCP-IP at BRL Reply-To: TCP-IP at BRL Date: 18 Jan 1982 Subject: TCP-IP Digest, Vol 1 #13 Via: Brl; 18 Jan 82 19:38-EDT TCP/IP Digest Monday, 18 Jan 1982 Volume 1 : Issue 13 Today's Topics: New Digest Headers Public Distribution of Digests Restricted Distribution of Mailing Lists/Digests/Etc? What is PUBLIC and What is PRIVATE Distribution? Precedent for Privacy of Electronic Mail Related Discussion Group Information To TCP or not to TCP? InterNet Addresses && Overloading the Dot TCP/IP on a Cyber ---------------------------------------------------------------------- LIMITED DISTRIBUTION For Research Use Only --- Not for Public Distribution ---------------------------------------------------------------------- From: Mike at BRL Subject: New Digest Headers This digest marks the first issue with the new headers people have been asking for. If this is not completely delightful to everybody, please let me know. Cheers, -Mike ------------------------------ From: Mike Peeler Subject: Public Distribution One way we could tell people that digest material is not public would be first to announce the fact as an administrative note in the digest and then to mention it whenever introducing a new member to the list (by sending them a "welcome" note). This approach is much less distasteful than that of putting a warning on the label. The problem with it is that not all subscribers are added by one list maintainer. In fact, many of them get the digests by mechanisms essentially equivalent to a bulletin board. This means that a large fraction of our new readers will never get a welcome note. If we regard the above as ineffective, then we want to make the warning as unobtrusive as possible. Two possibilities present them- selves. The first possibility would be to tuck it away in the header instead of the banner; for example Date: Friday, 15 January 1982 From: PCP-IV at BAL Private: Not for public distribution Subject: PCP-IV Digest V9 #23 To: PCP-IV at BAL The second possibility would be to place the warning at the bottom of the digest instead of the top; it might look like ------------------------------ ************** End of PCP-IV Digest ************** ********* (Not For Public Distribution) ********** ****************************************************** or perhaps something less ostentatious to the same effect. The latter, like the warning in the banner, will break some undigestifying programs, but probably in a milder and more easily reparable way. The former will not break most software, but may not be as aesthetically acceptable. Yours very sincerely, Mike ------------------------------ From: Paul A. Karger Subject: restricted distribution of mailing lists/digests/etc While putting a restricted distribution statement on a digest may be a psychological limitation on distribution, there a couple of problems. First, since ARPA and DCA are part of the DoD, there are specific regulations on what may or may not be marked as FOR OFFICIAL USE ONLY. The regulations are in part designed to not let people invent other kinds of markings. This dates back to the Ellsberg case and the desire to limit the ability of govt people to conceal information from the "public" (whoever that is) What happens if someone submits a Freedom of information Act request for Volume 2, Issue 6 of TCP-IP Digest? Worse still, what if someone submits a FOIA request for SFL?? My familliarity with the applicable regs is a little stale, because I left the Air Force over 2 years ago, but I would be very careful about developing new ways to restrict distribution of government information. Paul Karger ------------------------------ From: V. Ellen Golden Subject: [Paul A. Karger: restricted distribution of mailing lists/digests] THE ELLSBERG CASE? Gosh... I am ancient. Daniel Ellsberg made copies of the Pentagon Papers, which were secret discussions about the Vietnam War. He worked for a Harvard/MIT institute for Government (Political Science) sort of thing. He thought "the people" should know what was going on, and so made copies of the reports. The CIA etc was not enthusiased. The group who were involved in stopping him were the same group which became familiar later: The Plumbers (they invaded his psychiatrist's office, in California, yet). The Pentagon Papers were published in the New York Times. I admit I am not sure if this was the event which triggered the Plumbers (of later Watergate fame) to go to the psychiatrist's office, or if after all was said and done, the stuff finally made it to the public eye. (Other ancients like myself may recall better than I, and I beg them to correct my somewhat sketchy remembrance of the events.) anyway... in some way the Ellsberg case does have a bearing here in the Digest-people discussion. Do "the people" have a right to "know" (what?). I would say myself that what is in digests is not classified in any sense of the word, but I have no idea how that relates to our (ARPANET) status. ------------------------------ Subject: What is Public and what is Private distribution? From: TMPL at BBNG How my name got on the Digest list is a longish story, so I won't go into it, except to say that it had to do with my exploring some form of internetworking as a better way of accessing the net than I do currently (corporate phone lines to Boston or Telenet). My main use of the net is as part of the computer security community and is fully and properly justified by the right set of letters from the Pentagon to BBN, DCA, etc. ANYWAY, I have found the Digest most interesting and have been circulating it to our communications and systems research people, chiefly because it is quite clear that at some point our products will have to support TCP etc. for our government customers at least. I view such a redistribution as an acceptable way of communicating research results to those who can and ought to use them. It certainly isn't a public distribution in the sense of the mass media, but I would like to be reassured that it is an acceptably private one. Ted Lee Manager, Systems Security Sperry Univac Roseville, Mn. ------------------------------ From: Mike Muuss Subject: Re: What is Public and what is Private distribution? Ted - I am pleased that a major manufacturer like Sperry Univac is looking at the TCP/IP protocols (and the Digest). It is groups just like yours that everyone expects the Digests to go to. Frequently, the Digest will contain status reports about on-going projects, or discussions of evolving featues/bugs/whatever. This type of "working notes" material can often be confusing or misleading to the casual onlooker, and hence the concern that information which is going to the general public be obtained through official channels (Jon Postel or Vint Cerf, DCA/DARPA/ISI). Please do not re-publish materials or statements from the Digest, but please DO feel free to distribute it to any interested people within your organization. Sleepily, -Mike ------------------------------ Sender: TMPL at BBNG Subject: Precedent for Privacy of Electronic Mail There does seem to be some sort of legal basis for claiming that material such as the Digest is covered by existing privacy laws. The report in Computerworld about the US Postal Service's new Electronic Computer-Originated Mail (Ecom) contains the following interesting observation: "...because at least 50% of Ecom is electronically based, a part of the operation is covered under the rules of the Communications Act of 1934. A sign is conspicuously posted in Ecom's Boston center stating that anyone who violates a message's confidence or removes a scrap of paper from the Ecom computer room will be subject to a $10,000 fine, two years in jail or both. "Once the messages are put in the bright blue and white Ecom envelope, they are considered as mail and covered under the standard postal regulations governing mail handling, an Ecom spokesman said." (from Computerworld, 11 Jan 1982, p. 12) Ted Lee Sperry Univac ------------------------------ From: Zellich at OFFICE-3 (Rich Zellich) Subject: Pointers to mail header discussions Mike - Hers is what I have on Header-People and MsgGroup lists and their archives. This information is from OFFICE-3 publicly-accessible file INTEREST-GROUPS.TXT OFFICE-3 supports the net "standard" ANONYMOUS login with password ARPA. -Rich Zellich ALMSA --- HEADER-PEOPLE at MIT-MC Interest specifically in the format of message headers and related issues such as inter-network mail formats/standards, etc. Header-People messages are filed on MIT-MC:KSC;HEADER MINS [and MINS01, MINS02, etc.]. The ones more than 3 years old have been "reaped" but could be retrieved if anyone wants to see them. Coordinator: David A. Moon MsgGroup at MIT-ML Interest in electronic mail, message formats, message systems, and the sociological implications of the above. Coordinator: EStefferud at USC-ECL/MsgGroup at USC-ECL [ In future discussions of these issues, please CC these other lists. I would like to see the meta-discussion migrate to a more appropriate place. This does not remove it's vital importance to this list, however. -Mike ] ------------------------------ From: Christopher C. Stacy Subject: To TCP or not to TCP? Date: 14 January 1982 0206-PST (Thursday) From: lauren at UCLA-Security (Lauren Weinstein) Is there some good reason that the ITS machines cannot be gatewayed through a supported machine? Even little 11's like the 24 should be able to run some sort of existing TCP/IP implementation. Rand-Unix currently talks to the ARPANET over a 9600 baud tty line via an 11/34 running the NCP. --Lauren-- No, there is not any real reason why we cannot set up some limited gateway. However, the design of really complete gateways with protocol translation like one would want is an unsolved research question. In fact we will probably implement some limited functionality to connect our local network to the Internet, but we will not do any development which requires a major software effort. To implement a real TCP for an ITS machine would require about two years of heavy duty full time system programming, and since we are rapidly phasing out our timesharing machines we are not going to undertake such work. We are not going to do anything until we wake up one day and discover our plug being pulled out. ------------------------------ From: Hal.Cornell at UDel Subject: Internet addresses I was bothered by the "user.host @ net" internetwork address for another reason: the meaning of "@ foo" is ambiguous. Foo could be either a network name or just the name of a regular ARPANET host. How about using "user @ host @@ net" for internet addresses? The "@@ net" could be omitted if the user is on the same network, and if the user is on the same host, "@ host" could be eliminated also. Hal Perkins (Hal.Cornell @ UDel) ------------------------------ From: POSTEL at USC-ISIF Subject: Overloading the Dot In the transition plan for converting from NCP to TCP (RFC801), the plan for mail includes a provision for forwarding mail through a special forwarder program from NCP/FTP mail hosts to TCP/SMTP hosts by using the special forwarding address "user.host@fwdr". After considering the comments in this digest and in other inputs I have received, I do agree that the syntax "user.host@fwdr" will be awkward for some hosts to handle. We need to select a different character for the separator between the "user" and "host" parts in this special forwarding address. My current candidate is per cent (%). Are their any problems with this choice? I have discussed this with the people developing the JNT mail system in the UK, and this use of per cent fits in well with their syntax. --jon. ------------------------------ From: Crimmins at BRL-BMD Subject: TCP/IP on a CYBER I spoke with Tim Fallon of Tektronix about implementing TCP/IP on a CYBER 175 running NOS. The results of my conversation are as follows. DOCUMENTATION: All documentation is in the code and the Tektronix legal dept. is not ready to release the code. Therefore, no documentation is available. IMPLEMENTATION: IP implemented in a PP, ~4000 lines of PP assembler code. TCP runs at a control point in NOS, written in SYMPL. SYMPL is a CDC product. TomC ------------------------------ [ The following letter arrived Via UUCP, and cannot be answered through the BRL-BMD machine. -Mike ] From: steveg.Azure at BRL-BMD Cc: rickk.Teklabs at BRL-BMD, timf.Teklabs at BRL-BMD, clemc.Teklabs at BRL-BMD Subject: Tek's implementation of tcp-ip for Cyber The statement in the digest was somewhat wrong. [Volume 1 : Issue 12] The TCP implementation is done in SYMPL (an implementation language for the Cyber). The IP implementation is in Cyber Peripheral Processor assembler (and possibly some SYMPL code). The upper level stuff (FTP et. al) is being done in Ratfor. There are no plans to do a telnet server as Cyber timesharing is too heavily bound into the front-end hardware (psuedo terminals are very difficult to implement). We are using Hyperchannel hardware and have not done any work with any other hardware. Throughput looping back in the hyperchannel adapter on an idle Cyber 175 is in the 1Mbit range. This is all under NOS. An implementation fo VMS is also underway here, but I'm not very up on that one. I think it's being done in Ratfor (so they can use the same code for FTP as the Cyber). We did write a hyperchannel driver for 3Com's UNET on both the 11/70 and Vax Unix systems. The actual guys that did the work can be reached at: ...!teklabs!rickk (Rick Krull - FTP, Unix stuff) ...!teklabs!timf (Tim Fallon - Cyber IP, TCP, project mgr) [ snail mail: Box 500, M/S 50-454, Beaverton OR 97077 (503) 644-0161 ] Currently, they're working with 3Com to get the Unix-TCP flow control code working fully as the Cyber can easily overdrive the Unix machines. Steve Glaser END OF TCP-IP DIGEST ********************