From touch at ISI.EDU Wed Feb 18 18:08:52 2004 From: touch at ISI.EDU (Joe Touch) Date: Wed, 18 Feb 2004 18:08:52 -0800 Subject: [ih] change in mailing list policy Message-ID: <40341AB4.5080708@isi.edu> Hi, all, Regrettably, we have now enabled 'subscriber-only' posting on the internet-history mailing list. Although the amount of spam on this list has been kept fairly low overall, the contamination of the list archives, together with the increasingly common prevalence of this policy on other lists, has caused us to implement it here. Please let us know if this causes any complications when you try to post. Joe Touch & Bob Braden From nohzawa at w-tri.com Sat Feb 21 19:56:25 2004 From: nohzawa at w-tri.com (Toru Nohzawa ,) Date: Sun, 22 Feb 2004 12:56:25 +0900 Subject: [ih] ARPA initial IMP-IMP line speed Message-ID: Hello, all, I have a following question about the subject. If you know the fact, please let me know it Question: There are lots of documents mentioning the initial line speed of ARPAnet in 1969 was 50Kbps. As far as I know, the line speed at that time over the voice band transmission was less than 9,600bps. Most probably 4,800bps or 2,400bps. In addition, no modems for 50Kbps become available. Some document says 50Kbps over analog wideband. But this causes other question of what modem was used and who provided that. How many line speed is used with what modem at initial ARPA IMP-IMP? Regards, Toru Nohzawa nohzawa at w-tri.com From jnc at mercury.lcs.mit.edu Sun Feb 22 05:30:29 2004 From: jnc at mercury.lcs.mit.edu (Noel Chiappa) Date: Sun, 22 Feb 2004 08:30:29 -0500 (EST) Subject: [ih] ARPA initial IMP-IMP line speed Message-ID: <20040222133029.15C03872C7@mercury.lcs.mit.edu> > From: "Toru Nohzawa ," > There are lots of documents mentioning the initial line speed of > ARPAnet in 1969 was 50Kbps. As far as I know, the line speed at that > time over the voice band transmission was less than 9,600bps. ... > In addition, no modems for 50Kbps become available. Some document says > 50Kbps over analog wideband. .. > How many line speed is used with what modem at initial ARPA IMP-IMP? They didn't use over-the-counter modems - nor, IIRC, standard phone lines. Without looking I don't recall exactly how they were special (I have this vague memory that they were specially conditioned), but I'm pretty sure that among other things it was a pair of lines - i.e. 4 wires. The modems used in the first IMPs were large, complex devices made up of a number of smaller logic cards; again, without looking I don't recall the details. The 50 Kbit/second is correct. Noel From louie at TransSys.COM Sun Feb 22 07:12:31 2004 From: louie at TransSys.COM (Louis A. Mamakos) Date: Sun, 22 Feb 2004 10:12:31 -0500 Subject: [ih] ARPA initial IMP-IMP line speed In-Reply-To: Your message of "Sun, 22 Feb 2004 08:30:29 EST." <20040222133029.15C03872C7@mercury.lcs.mit.edu> References: <20040222133029.15C03872C7@mercury.lcs.mit.edu> Message-ID: <200402221512.i1MFCV0R013338@whizzo.transsys.com> > > From: "Toru Nohzawa ," > > > There are lots of documents mentioning the initial line speed of > > ARPAnet in 1969 was 50Kbps. As far as I know, the line speed at that > > time over the voice band transmission was less than 9,600bps. ... > > In addition, no modems for 50Kbps become available. Some document says > > 50Kbps over analog wideband. .. > > How many line speed is used with what modem at initial ARPA IMP-IMP? > > They didn't use over-the-counter modems - nor, IIRC, standard phone lines. > Without looking I don't recall exactly how they were special (I have this > vague memory that they were specially conditioned), but I'm pretty sure that > among other things it was a pair of lines - i.e. 4 wires. The modems used in > the first IMPs were large, complex devices made up of a number of smaller > logic cards; again, without looking I don't recall the details. The 50 > Kbit/second is correct. > > Noel The analog wideband modems that I saw consumed about half of a 23" equipment rack each. louie From nohzawa at w-tri.com Sun Feb 22 18:44:50 2004 From: nohzawa at w-tri.com (Toru Nohzawa ,) Date: Mon, 23 Feb 2004 11:44:50 +0900 Subject: [ih] ARPA initial IMP-IMP line speed In-Reply-To: <1727115981.20040223085237@brandenburg.com> Message-ID: Dave, Thank your for the information. Do you know what modem they did use ? 50Kbps or 56Kbps whichever at that time, seems 10 or more times faster than Chronologically told modem technology. Who could provide it with what modulation method ? Toru, -----Original Message----- From: Dave Crocker [mailto:dhc at dcrocker.net] Sent: Monday, February 23, 2004 9:53 AM To: Toru Nohzawa , Cc: internet-history at postel.org Subject: Re: [ih] ARPA initial IMP-IMP line speed Toru, For simplification, all initial Arpanet links were 56Kbps. This avoided having to deal with differential performace stemming from different line speeds. d/ TN> Hello, all, TN> I have a following question about the subject. TN> If you know the fact, please let me know it TN> Question: TN> There are lots of documents mentioning the initial line speed of ARPAnet in TN> 1969 was 50Kbps. TN> As far as I know, the line speed at that time over the voice band TN> transmission was TN> less than 9,600bps. Most probably 4,800bps or 2,400bps. TN> In addition, no modems for 50Kbps become available. TN> Some document says 50Kbps over analog wideband. But this causes other TN> question of what modem was used and who provided that. TN> How many line speed is used with what modem at initial ARPA IMP-IMP? TN> Regards, TN> Toru Nohzawa TN> nohzawa at w-tri.com d/ -- Dave Crocker Brandenburg InternetWorking Sunnyvale, CA USA From info at w-tri.com Sat Feb 21 14:52:40 2004 From: info at w-tri.com (Info TechnoReview) Date: Sun, 22 Feb 2004 07:52:40 +0900 Subject: [ih] ARPA initial IMP-IMP line speed Message-ID: Hello, all, I have a following question about the subject. If you know the fact, please let me know it Question: There are lots of documents mentioning the initial line speed of ARPAnet in 1969 was 50Kbps. As far as I know, the line speed at that time over the voice band transmission was less than 9,600bps. Most probably 4,800bps or 2,400bps. In addition, no modems for 50Kbps become available. Some document says 50Kbps over analog wideband. But this causes other question of what modem was used and who provided that. How many line speed is used with what modem at initial ARPA IMP-IMP? Regards, Toru Nohzawa nohzawa at w-tri.com From map at snap.org Sun Feb 22 15:49:50 2004 From: map at snap.org (Mike Padlipsky) Date: Sun, 22 Feb 2004 15:49:50 -0800 Subject: [ih] ARPA initial IMP-IMP line speed In-Reply-To: <20040222133029.15C03872C7@mercury.lcs.mit.edu> Message-ID: <5.2.1.1.1.20040222150633.00aa46b8@mail.lafn.org> At 05:30 AM 2/22/2004, Noel Chiappa wrote: >(I have this >vague memory that they were specially conditioned) well, i have a clear memory that we called 'em 'leased lines', but since i never did 'do' hardware i had no idea how they were specially conditioned, tho that they were such was clearly implied. indeed, i strongly suspect that even the dial-up lines to multics, for those lucky enough to have access from home, had to be specially conditioned in some fashion in those times -- and weren't anywhere near 50kb/s. [memory does serve a fault as to whether The Phone Company still had to be argued with, and paid thru the nose, for any sort of modem by then, but i wouldn't be at all surprised if it were still the case.] [where then = '69 thru, say, '72 or '73 ... or for all i know '74 or later.] presumably, somebody who did 'do' hardware at the time will notice that something non-spammish is happening on [ih] again and will set us all straight on just what TPC did to the lines to make 'em carry that many bits, but since i do remember how pleased we were long about '72 when we were able to get rates of some 40 kb/s on ftp's going between a couple of multicses one hop away from each other, at least i can make it unanimous that 50kb/s lines were in play. for the 'main' hosts, anyway; according to another fairly clear memory, it wasn't too long before some hosts [well, ok, some hosts' imps] were attached over slower lines, maybe as slow as 9.6 in at least one case, whether for economic reasons or because of local unavailability of leasable lines suitable for the necessary conditioning ... or whatever. cheers, map [whose shoulder problems caused him to break down some time ago and create a 'signature' file to apologize for the lack of his formerly customary e-volubility -- and who's been employing shiftless typing for a long time now to spare his wristsnfingers, in case you didn't know ... and who's further broken down and done http://www.lafn.org/~ba213/mapstuff.html , rather grudgingly] From dhc at dcrocker.net Sun Feb 22 16:52:37 2004 From: dhc at dcrocker.net (Dave Crocker) Date: Mon, 23 Feb 2004 08:52:37 +0800 Subject: [ih] ARPA initial IMP-IMP line speed In-Reply-To: References: Message-ID: <1727115981.20040223085237@brandenburg.com> Toru, For simplification, all initial Arpanet links were 56Kbps. This avoided having to deal with differential performace stemming from different line speeds. d/ TN> Hello, all, TN> I have a following question about the subject. TN> If you know the fact, please let me know it TN> Question: TN> There are lots of documents mentioning the initial line speed of ARPAnet in TN> 1969 was 50Kbps. TN> As far as I know, the line speed at that time over the voice band TN> transmission was TN> less than 9,600bps. Most probably 4,800bps or 2,400bps. TN> In addition, no modems for 50Kbps become available. TN> Some document says 50Kbps over analog wideband. But this causes other TN> question of what modem was used and who provided that. TN> How many line speed is used with what modem at initial ARPA IMP-IMP? TN> Regards, TN> Toru Nohzawa TN> nohzawa at w-tri.com d/ -- Dave Crocker Brandenburg InternetWorking Sunnyvale, CA USA From m.brescia at comcast.net Sun Feb 22 20:19:39 2004 From: m.brescia at comcast.net (Mike Brescia) Date: Sun, 22 Feb 2004 23:19:39 -0500 Subject: [ih] ARPA initial IMP-IMP line speed Message-ID: >There are lots of documents mentioning the initial line speed of ARPAnet in >1969 was 50Kbps. That was the speed of the modem. >As far as I know, the line speed at that time over the voice band >transmission was >less than 9,600bps. Most probably 4,800bps or 2,400bps. The best voice band modem available then coded 9600 bps on 1200 baud half-duplex; for full duplex you would need 4 wires (two "phone lines") which were "conditioned" (at least tested for quality, maybe more). >Some document says 50Kbps over analog wideband. But this causes other >question of what modem was used and who provided that. The 50kb was called "wideband" because it did baseband signalling at a rate higher than the 3kHz voice bandwidth, in other words, the data signalling was actually 50kbaud, not like today's 50+kbps on top of 1200 baud (remember Nyquist) I'm really out of my depth here, but recollection is that the lines (4-wire again) were bare copper, no filters etc, and had to be specially patched through each level of the telco office wiring. The wiring had to maintain DC polarity correctly, otherwise the data came out inverted. For "long" distances, there must have been some sort of repeater. You really need to find someone familiar with the circuitry and signalling. Regards, -- Mike From mills at udel.edu Sun Feb 22 20:28:45 2004 From: mills at udel.edu (David L. Mills) Date: Mon, 23 Feb 2004 04:28:45 +0000 Subject: [ih] ARPA initial IMP-IMP line speed References: Message-ID: <4039817D.BED9A6D1@udel.edu> Dave & Co., Little bit of history here. In 1969 AT&T (Western Electric) was the only source for analog and digital lines and most high-speed modems used a 48-kHz analog group circuit normally used to multiplex 12 4-kHz voice channels. There were several modems used by AT&T at 40, 48 and 56 kbps. In theory, AT&T could provide a supergroup at 6.3 MHz or jumbogroup at 44 MHz with digitronics to match and I bet they did on special order. Before 1968 the higest speed you could get on an ordinary telephone line was 300 bps full-duplex (Bell 103 FSK), 1200 bps half-duplex (Bell 202 FSK) and 2000 bps half-duplex (Bell 201 DQPSK). The last could go at a blinding 2400 bps full-duplex on a 3002 leased circuit. You couldn't get anything higher on an ordinary phone or leased line until the middle eighties. Starting in 1986, I had a full-duplext 3002 at 9600 bps between my home and campus, so that may calibrate the curve. Until 1968 you couldn't put anything on an ordinary phone line except AT&T devices. But, the FCC told AT&T to provide a coupling device called a Data Access Arrangement (DAA) so the customer could use non-AT&T modems. I had the first one installed in Michigan. They wanted to get it right, so they sent a fleet of trucks to install it. But, there were lots of special order things AT&T could provide, like analog program radio and TV lines and on these you didn't have to use AT&T modems. That's how I tested early spread-spectrum digital telephone gadgets as a grad student. Dave "Toru Nohzawa ," wrote: > > Dave, > > Thank your for the information. > > Do you know what modem they did use ? > 50Kbps or 56Kbps whichever at that time, seems 10 or more times > faster than Chronologically told modem technology. > Who could provide it with what modulation method ? > > Toru, > > -----Original Message----- > From: Dave Crocker [mailto:dhc at dcrocker.net] > Sent: Monday, February 23, 2004 9:53 AM > To: Toru Nohzawa , > Cc: internet-history at postel.org > Subject: Re: [ih] ARPA initial IMP-IMP line speed > > Toru, > > For simplification, all initial Arpanet links were 56Kbps. > > This avoided having to deal with differential performace stemming from > different line speeds. > > d/ > > TN> Hello, all, > TN> I have a following question about the subject. > TN> If you know the fact, please let me know it > > TN> Question: > TN> There are lots of documents mentioning the initial line speed of ARPAnet > in > TN> 1969 was 50Kbps. > TN> As far as I know, the line speed at that time over the voice band > TN> transmission was > TN> less than 9,600bps. Most probably 4,800bps or 2,400bps. > TN> In addition, no modems for 50Kbps become available. > TN> Some document says 50Kbps over analog wideband. But this causes other > TN> question of what modem was used and who provided that. > > TN> How many line speed is used with what modem at initial ARPA IMP-IMP? > > TN> Regards, > > TN> Toru Nohzawa > TN> nohzawa at w-tri.com > > d/ > -- > Dave Crocker > Brandenburg InternetWorking > Sunnyvale, CA USA From rogers at ISI.EDU Sun Feb 22 23:02:43 2004 From: rogers at ISI.EDU (Craig Milo Rogers) Date: Sun, 22 Feb 2004 23:02:43 -0800 Subject: [ih] ARPA initial IMP-IMP line speed In-Reply-To: <4039817D.BED9A6D1@udel.edu> References: <4039817D.BED9A6D1@udel.edu> Message-ID: <20040223070243.GA5981@isi.edu> On 04.02.23, David L. Mills wrote: > Before 1968 the higest speed you could get on an ordinary telephone line > was 300 bps full-duplex (Bell 103 FSK), 1200 bps half-duplex (Bell 202 > FSK) and 2000 bps half-duplex (Bell 201 DQPSK). The last could go at a > blinding 2400 bps full-duplex on a 3002 leased circuit. You couldn't get > anything higher on an ordinary phone or leased line until the middle > eighties. By the mid-1970's you could get a 1200 bps modem with a 150 bps back channel; the combination provided asymmetric full-duplex asynchronous communication, which was much more desirable than 1200 bps half-duplex operation for timesharing terminal usage (but not, perhaps, for IMP-to-IMP connections!). See, for example, the Stanford Artificial Intelligence Laboratory (SAIL) history Web site: http://www-db.stanford.edu/pub/voy/museum/pictures/display/3-2-Modem.htm 1200 bps full-duplex modems for ordinary phone lines were commercially available at a "reasonable" price (under $1000) in 1976, according to the above-cited SAIL Web site. I used 1200/150 full-duplex asynchronous modems to dialin to UCLA's Center for Computer-based Behavioral Studies (UCLA/CCBS, an ARPAnet node) from home in 1974 (or earlier?). I used both a TI Silent 700 terminal and one of CCBS's special storage tube display terminals. Switching from a 300 bps full-duplex modem to a 1200/150 full-duplex modem, I immediately noticed the reduced uplink bandwidth when using the TI Silent 700: my Silent 700 didn't buffer keystrokes, and I could type alternate-hand character pairs faster than they could be transmitted, causing the second keystroke to be lost. I had to slow my typing rate to adapt to the "higher speed" modem. I don't think I had the same problem with the storage tube terminal; it buffered keystrokes, I believe. I can almost recall the schematic. Somewhere in storage, I have a manual... :-) Craig Milo Rogers From rogers at ISI.EDU Sun Feb 22 23:50:11 2004 From: rogers at ISI.EDU (Craig Milo Rogers) Date: Sun, 22 Feb 2004 23:50:11 -0800 Subject: [ih] ARPA initial IMP-IMP line speed In-Reply-To: <5.2.1.1.1.20040222150633.00aa46b8@mail.lafn.org> References: <20040222133029.15C03872C7@mercury.lcs.mit.edu> <5.2.1.1.1.20040222150633.00aa46b8@mail.lafn.org> Message-ID: <20040223075011.GB5981@isi.edu> On 04.02.22, Mike Padlipsky wrote: > At 05:30 AM 2/22/2004, Noel Chiappa wrote: > >(I have this > >vague memory that they were specially conditioned) > > well, i have a clear memory that we called 'em 'leased lines', but since i > never did 'do' hardware i had no idea how they were specially conditioned, > tho that they were such was clearly implied. indeed, i strongly suspect > that even the dial-up lines to multics, for those lucky enough to have > access from home, had to be specially conditioned in some fashion in those > times -- and weren't anywhere near 50kb/s. My memory is that there were three leased line services that computer folk would commonly request from the phone company at the time: 1) A "conditioned" line was usually a line from the telco's central office (CO) to the user's end-point that had been tested, and if necessary, adjusted, to provide a relatively flat gain over a certain frequency band. 300 bps FSK modems usually worked over ordinary, unconditioned phone lines, but a 1200 bps modem was more questionable over an unconditioned line. Conditioned lines would have a uniform impedence, eliminating nasty echoes, etc. Conditioning might be a one-time affair, or a continuous "guarantee" by the telco. They worked well for some people, but were an endless nightmare for others. For example, some lines would lose their conditioning whenever it rained (an infrequent occurance in Los Angeles), and regain their desired characteristics by themselves when they dried (perhaps before the telco serviceman arrived). Conditioned lines might be part of the standard dial phone network, or they might be provided as part of a private dial network or a dedicated circuit between two (or more, "multidrop") end-points. 2) An "unloaded bare copper" circuit was a leased circuit between two user end-points with no filters, no frequency-multiplexed shared circuits, etc. You could use it for simple burglar alarm circuits (which used DC signals to indicate a breakin), you could attach teletypes (current loop at 75 or 110 bps, perhaps), or you could attach computer modems to it (so long as your signal didn't exceed certain frequency and energy limits, designed to prevent signals on adjacent pairs of wires from coupling to and interfering with each other). Unloaded bare copper circuits in the same CO were relatively easy for the telco to provide ('till the neighborhood pairs ran out), and relatively inexpensive. You might use an unloaded bare copper circuit to connect a Host to an IMP in a nearby location. Some organizations (such as, say, UCLA) would lay their own bare copper wires to avoid telco hassles or provide direct, shorter circuits by bypassing the telco's CO -- and lose track of them, later. 3) A "dedicated long-distance circuit", with various caveats about the type of modulation to be used on the line and the degree of isolation needed between it and other lines. This was the type of line used between most (but not all!) adjacent ARPAnet IMPs, and required substantial one-time engineering charges and hefty monthly lease fees. Craig Milo Rogers From nohzawa at w-tri.com Mon Feb 23 06:30:59 2004 From: nohzawa at w-tri.com (Toru Nohzawa ,) Date: Mon, 23 Feb 2004 23:30:59 +0900 Subject: [ih] ARPA initial IMP-IMP line speed In-Reply-To: <4039817D.BED9A6D1@udel.edu> Message-ID: Dave, For IMP-IMP communication, theoretically you are suggesting that at 1969, they could used Frequency Divisional Modulation modem of 48KHz, 6.3Mhz or 44Mhz provided by AT&T. Neglecting the physical cross-continental line configuration within AT&T, logically, ARPA could reserve over 12 cross-continental voice-band leased lines as a single 50Kbps leased line. Do you know the way we can confirm your suggetion ? Or, do you know somebody who knows the actual IMP spec. ? Toru -----Original Message----- From: David L. Mills [mailto:mills at udel.edu] Sent: Monday, February 23, 2004 1:29 PM To: Toru Nohzawa , Cc: Dave Crocker; internet-history at postel.org Subject: Re: [ih] ARPA initial IMP-IMP line speed Dave & Co., Little bit of history here. In 1969 AT&T (Western Electric) was the only source for analog and digital lines and most high-speed modems used a 48-kHz analog group circuit normally used to multiplex 12 4-kHz voice channels. There were several modems used by AT&T at 40, 48 and 56 kbps. In theory, AT&T could provide a supergroup at 6.3 MHz or jumbogroup at 44 MHz with digitronics to match and I bet they did on special order. Before 1968 the higest speed you could get on an ordinary telephone line was 300 bps full-duplex (Bell 103 FSK), 1200 bps half-duplex (Bell 202 FSK) and 2000 bps half-duplex (Bell 201 DQPSK). The last could go at a blinding 2400 bps full-duplex on a 3002 leased circuit. You couldn't get anything higher on an ordinary phone or leased line until the middle eighties. Starting in 1986, I had a full-duplext 3002 at 9600 bps between my home and campus, so that may calibrate the curve. Until 1968 you couldn't put anything on an ordinary phone line except AT&T devices. But, the FCC told AT&T to provide a coupling device called a Data Access Arrangement (DAA) so the customer could use non-AT&T modems. I had the first one installed in Michigan. They wanted to get it right, so they sent a fleet of trucks to install it. But, there were lots of special order things AT&T could provide, like analog program radio and TV lines and on these you didn't have to use AT&T modems. That's how I tested early spread-spectrum digital telephone gadgets as a grad student. Dave "Toru Nohzawa ," wrote: > > Dave, > > Thank your for the information. > > Do you know what modem they did use ? > 50Kbps or 56Kbps whichever at that time, seems 10 or more times > faster than Chronologically told modem technology. > Who could provide it with what modulation method ? > > Toru, > > -----Original Message----- > From: Dave Crocker [mailto:dhc at dcrocker.net] > Sent: Monday, February 23, 2004 9:53 AM > To: Toru Nohzawa , > Cc: internet-history at postel.org > Subject: Re: [ih] ARPA initial IMP-IMP line speed > > Toru, > > For simplification, all initial Arpanet links were 56Kbps. > > This avoided having to deal with differential performace stemming from > different line speeds. > > d/ > > TN> Hello, all, > TN> I have a following question about the subject. > TN> If you know the fact, please let me know it > > TN> Question: > TN> There are lots of documents mentioning the initial line speed of ARPAnet > in > TN> 1969 was 50Kbps. > TN> As far as I know, the line speed at that time over the voice band > TN> transmission was > TN> less than 9,600bps. Most probably 4,800bps or 2,400bps. > TN> In addition, no modems for 50Kbps become available. > TN> Some document says 50Kbps over analog wideband. But this causes other > TN> question of what modem was used and who provided that. > > TN> How many line speed is used with what modem at initial ARPA IMP-IMP? > > TN> Regards, > > TN> Toru Nohzawa > TN> nohzawa at w-tri.com > > d/ > -- > Dave Crocker > Brandenburg InternetWorking > Sunnyvale, CA USA From touch at ISI.EDU Tue Feb 24 11:07:12 2004 From: touch at ISI.EDU (Joe Touch) Date: Tue, 24 Feb 2004 11:07:12 -0800 Subject: [ih] updated archives Message-ID: <403BA0E0.8080402@isi.edu> Hi, all, In an effort to make the archives more useful, we've filtered most of the spam out and reposted the updated version, which are now available at the usual place (www.postel.org/pipermail/internet-history) NOTE: there may be a few spam messages still left; this was not an attempt to be exhaustive. Thanks, Joe From louie at TransSys.COM Tue Feb 24 11:25:32 2004 From: louie at TransSys.COM (Louis A. Mamakos) Date: Tue, 24 Feb 2004 14:25:32 -0500 Subject: [ih] ARPA initial IMP-IMP line speed In-Reply-To: Your message of "Tue, 24 Feb 2004 14:54:02 +0900." References: Message-ID: <200402241925.i1OJPW0R072264@whizzo.transsys.com> > Dave, > > My point is IMP-IMP fact. > I thought there must be some people who handled the IMPs > and they must know the fact. > > If early ARPAnet used the multiplexed or wideband modem for > IMP-IMP communication, it might be new news to many people > and it should be described on the internet history. > > Do you know somebody who handled the communication portion of the IMP ? I think that multiple people already confirmed that the ARPANET IMP/PSNs were interconnected at 50kb/s. I think you can consider that a fact. I was there, though later on with DDS circuit provisioning on ARPANET IMP (then called "PSN") number 17. I did, however, see a couple of early ARPANET IMP sites with the old analog wideband modems. Most of the discussion subsequent to that has been to explain how "wideband" telecom circuits might have been provisioned by AT&T at the time. That's interesting, though there are likely many other venues that might have move detailed history of AT&T and the stuff they deployed in the field. louie From chris at cs.utexas.edu Tue Feb 24 12:23:29 2004 From: chris at cs.utexas.edu (Chris Edmondson-Yurkanan) Date: Tue, 24 Feb 2004 14:23:29 -0600 Subject: [ih] ARPA initial IMP-IMP line speed Message-ID: <200402242023.i1OKNTnd030835@neverland.cs.utexas.edu> Toru, at this point I want to make sure which documents you have read on the IMP hardware impl. and design... since I suggested [ih] to you. Have you read the 1970 paper on The interface message processor for the ARPA computer network (by Heart, Kahn, Ornstein, Crowther, and Walden) And the Arpanet Completion Report section on phone carriers, etc (on my digital archive) and the RFQ proposal? (on my digital archive) And I expect that you would enjoy the materials that Katie Hafner donated to the www.archive.org: (such as pieces of the BBN response to the RFQ, and BBN's first quarterly report?) (both of which I only found today, while searching for IMP materials for you) Thanks, Chris -- Chris Edmondson-Yurkanan TAY 4.136; +1 512 471 9546 Fax: 471 8885 The University of Texas at Austin Email addresses are: chris at cs.utexas.edu Computer Sciences Department or for fun: dragon at cs.utexas.edu 1 University Station C0500 URL: www.cs.utexas.edu/users/chris/ Austin, TX 78712-0233 Fedex: please send to Taylor Hall 2.124 From m.brescia at comcast.net Tue Feb 24 13:31:43 2004 From: m.brescia at comcast.net (Mike Brescia) Date: Tue, 24 Feb 2004 16:31:43 -0500 Subject: [ih] ARPA initial IMP-IMP line speed In-Reply-To: References: Message-ID: Hi, The decision to use 50kb instead of 9.6kb as mentioned in "The Dream Machine"* was made by Larry Roberts as the ARPA program manager for the network project. "what the phone companies called a 50-kilobit line." [The Dream Machine, J.C.R.Licklider and the Revolution That Made Computing Personal, M.Mitchell Waldrop, Viking, New York, 2001, ISBN 0-670-89976-3 -- page 276] A reference in the 1987 report describing interfaces on a compatable IMP http://www.totallynerd.com/library/files/c30e_int.txt mentions the modem equipment -- Bell System Technical Reference; "Wideband Data Stations 303-Type," August 1966. >Do you know somebody who handled the communication portion of the IMP ? All who I have heard from agree that 50kb was the speed. There were experiments with a 230.4kb service but the concerns were with the economics of renting the lines and the overall computing capacity of the IMPs. Most of us who have responded to your notes are computer people, not modem hardware or telecomms people. We have some recollections of descriptions of what went on in the telco equipment, bandwidth and the hierarchies of voice circuit aggregation. You need to find people who were closer to AT&T, Bell Labs or Western Electric, perhaps Lucent now, to get more detail about wiring, encoding or equipment. Regards, -- Mike (BBN &seq.1978-2002) Disclaimer: I have no connection with Mr. Waldrop or the "Dream Machine" book other than that I have a copy. - m From mills at udel.edu Tue Feb 24 13:39:40 2004 From: mills at udel.edu (David L. Mills) Date: Tue, 24 Feb 2004 21:39:40 +0000 Subject: [ih] ARPA initial IMP-IMP line speed References: <200402242023.i1OKNTnd030835@neverland.cs.utexas.edu> Message-ID: <403BC49C.7DC83FB6@udel.edu> Chris, The IMP interface is described in BBN Report 1822, which I have. There were companion documents for the Digital Equipment IMP-11A interface for the PDP-11 Unibus which I don't have. My first ARPAnet connection at COMSAT Labs was a 4800-bps analog link to BBN; later I connected with 56-kbps DDS and IMP 29 at Mitre. Toward the end of ARPAlife I connected from UDel via 56-kbps DDS to IMP 112 somewhere in WashDC. Scary to think now I have 100-Mbps desktop-desktop just about anywhere in Internet2. How the heck did we get along with only a 50-kbps network? Dave Chris Edmondson-Yurkanan wrote: > > Toru, > at this point I want to make sure which documents you have read on the > IMP hardware impl. and design... since I suggested [ih] to you. > > Have you read the 1970 paper on > The interface message processor for the ARPA computer network > (by Heart, Kahn, Ornstein, Crowther, and Walden) > > And the Arpanet Completion Report section on phone carriers, etc > (on my digital archive) > and the RFQ proposal? (on my digital archive) > > And I expect that you would enjoy the materials that Katie Hafner donated > to the www.archive.org: (such as pieces of the BBN response to the RFQ, > and BBN's first quarterly report?) (both of which I only found today, > while searching for IMP materials for you) > > Thanks, Chris > > -- > Chris Edmondson-Yurkanan TAY 4.136; +1 512 471 9546 Fax: 471 8885 > The University of Texas at Austin Email addresses are: chris at cs.utexas.edu > Computer Sciences Department or for fun: dragon at cs.utexas.edu > 1 University Station C0500 URL: www.cs.utexas.edu/users/chris/ > Austin, TX 78712-0233 Fedex: please send to Taylor Hall 2.124 From casner at acm.org Tue Feb 24 14:59:24 2004 From: casner at acm.org (Stephen Casner) Date: Tue, 24 Feb 2004 14:59:24 -0800 (PST) Subject: [ih] ARPA initial IMP-IMP line speed In-Reply-To: References: Message-ID: <20040224145847.E75277@ash.packetdesign.com> On Tue, 24 Feb 2004, Toru Nohzawa , wrote: > My point is IMP-IMP fact. > I thought there must be some people who handled the IMPs > and they must know the fact. > > If early ARPAnet used the multiplexed or wideband modem for > IMP-IMP communication, it might be new news to many people > and it should be described on the internet history. I do not know details of the modems, but I remember the large cabinet that contained them. This was at ISI from 1973 onward. BBN Report 1822, "Interface Message Processor: Specifications for the Interconnection of a Host and an IMP", shows a picture of that cabinet on page 1-3 and states the dimensions as 68-1/8" x 28" x 28". That document does not describe the modems that connect the IMP to the "wideband communication lines" that interconnect the IMPs. -- Steve From mills at udel.edu Tue Feb 24 22:52:41 2004 From: mills at udel.edu (David L. Mills) Date: Wed, 25 Feb 2004 06:52:41 +0000 Subject: [ih] ARPA initial IMP-IMP line speed References: <200402242023.i1OKNTnd030835@neverland.cs.utexas.edu> <403BC49C.7DC83FB6@udel.edu> <20040225042106.GJ1668@owl.songbird.com> Message-ID: <403C4639.2ED28536@udel.edu> Kent, Depends who you talk to. AT&T had several special-order circuits, including 48, 50 and 56-kbps circuits, each individually engineered. The 48 and 50 kbps circuits were defintely ananlog, but the 56-kbps circuits were digital derived from 64-kbps voice circuits with a supervisory bit every 192 bits. It might not make much difference from the viewpoint of today and especially because of the ubiquitous nature of AT&T penetration of the 1970s. Dave kent at icann.org wrote: > > On Tue, Feb 24, 2004 at 09:39:40PM +0000, David L. Mills wrote: > > Chris, > > > > The IMP interface is described in BBN Report 1822, which I have. There > > were companion documents for the Digital Equipment IMP-11A interface for > > the PDP-11 Unibus which I don't have. My first ARPAnet connection at > > COMSAT Labs was a 4800-bps analog link to BBN; later I connected with > > 56-kbps DDS and IMP 29 at Mitre. Toward the end of ARPAlife I connected > > from UDel via 56-kbps DDS to IMP 112 somewhere in WashDC. Scary to think > > now I have 100-Mbps desktop-desktop just about anywhere in Internet2. > > How the heck did we get along with only a 50-kbps network? > > I wasn't closely connected at the time, and my memory is dim, but I > always thought that the early long line arpanet connections were all 56 > kbs. Where does 50 kbs come from? > > Kent > > -- > Kent Crispin > kent at icann.org p: +1 310 823 9358 f: +1 310 823 8649 > kent at songbird.com SIP: 81202 at fwd.pulver.com From nohzawa at w-tri.com Wed Feb 25 03:19:37 2004 From: nohzawa at w-tri.com (Toru Nohzawa ,) Date: Wed, 25 Feb 2004 20:19:37 +0900 Subject: [ih] ARPA initial IMP-IMP line speed In-Reply-To: Message-ID: Thanks Mike, According to the 1987 BBN report you mentioned, "Bell Data Station 303-type" supports 230.4K, 50K, 19.2K with internal clock selection. Therefore If IMP did use Bell Data Station 303-type, 50K was reasonable. (By the way, what is "Data Station 303-type" ? We know Bell 103, 212 as a modem. So "303" could be the analog modem we are talking about, I hope. ) Yes, I know Waldrop's "The Dream Machine" mentioning 50Kbps. But his description of "The idea was that you bought a very expensive modem that tied twelve lines together" causes me some confusion. The straight forward interpretation of the sentence is "AT&T provides 9.6Kx12lines". But the word "idea" sound something like "analogy". Then, the sentence could be interpretable like "actually 50K but analogous to 9.6x12". In addition, 9.6x12=115.2, 9.6x6=57.6. It doesn't agree with 50K. Therefore, I think this portion of the book gives some confusion to the careful non-modem-background readers. I recommend we have better to describe the Iine speed of IMP-IMP at 1969 more clearly such as " 50Kbps through the specially ordered analog wideband.least line" , to avoid confusion with current 56K voice band modem. Toru ----Original Message----- From: Mike Brescia [mailto:m.brescia at comcast.net] Sent: Wednesday, February 25, 2004 6:32 AM To: Toru Nohzawa , Cc: internet-history at postel.org Subject: RE: [ih] ARPA initial IMP-IMP line speed Hi, The decision to use 50kb instead of 9.6kb as mentioned in "The Dream Machine"* was made by Larry Roberts as the ARPA program manager for the network project. "what the phone companies called a 50-kilobit line." [The Dream Machine, J.C.R.Licklider and the Revolution That Made Computing Personal, M.Mitchell Waldrop, Viking, New York, 2001, ISBN 0-670-89976-3 -- page 276] A reference in the 1987 report describing interfaces on a compatable IMP http://www.totallynerd.com/library/files/c30e_int.txt mentions the modem equipment -- Bell System Technical Reference; "Wideband Data Stations 303-Type," August 1966. >Do you know somebody who handled the communication portion of the IMP ? All who I have heard from agree that 50kb was the speed. There were experiments with a 230.4kb service but the concerns were with the economics of renting the lines and the overall computing capacity of the IMPs. Most of us who have responded to your notes are computer people, not modem hardware or telecomms people. We have some recollections of descriptions of what went on in the telco equipment, bandwidth and the hierarchies of voice circuit aggregation. You need to find people who were closer to AT&T, Bell Labs or Western Electric, perhaps Lucent now, to get more detail about wiring, encoding or equipment. Regards, -- Mike (BBN &seq.1978-2002) Disclaimer: I have no connection with Mr. Waldrop or the "Dream Machine" book other than that I have a copy. - m From mills at udel.edu Wed Feb 25 07:54:13 2004 From: mills at udel.edu (David L. Mills) Date: Wed, 25 Feb 2004 15:54:13 +0000 Subject: [ih] ARPA initial IMP-IMP line speed References: <200402242023.i1OKNTnd030835@neverland.cs.utexas.edu> <403BC49C.7DC83FB6@udel.edu> <20040225042106.GJ1668@owl.songbird.com> <403C4639.2ED28536@udel.edu> <168990000.1077714705@localhost> Message-ID: <403CC525.686D7198@udel.edu> Guy, You are correct; the eigth bit was for scrambling and the supervisory bits were every 193 bits. This was the original DS-1 specification; later the Extended Superframe spec got even more ingenious. The real lesson of the analog era was that whatever you did had to fid in the multiplexor hierarchy at the group, super and jumbo bandwidths. That's why the silly breakout at 50 kbps and 230 kbps. The really interesting issues were how the transition between analog and digital domains were managed and, for example, how TV transitioned from 4.5 MHz analog to 2xT3 (92 Mbps) with basically old technology. That was clever. One of the digital breakout rates is 6.3 Mbps, originally I am told, to support Picturephone. Remember that? Dave Guy T Almes wrote: > > Dave, > As I recall, you got only 56 Kb/s from a 64-Kb/s DS0 circuit by wasting > every 8th bit (to avoid long strings of zeroes that messed up the > encompassing T1 circuit). > The supervisory bit was one bit in every 193-bit T1 frame (24 8-bit > chunks plus that pesky supervisory bit). > But this was the era of digital service, at least. > I'm really enjoying learning more about the pre-digital service era. I > recall the monster IMP equipment in the CMU CS machine room and always > imagined that squeezing 50 Kb/s out of an analog circuit was pretty tense. > Regards, > -- Guy > > --On Wednesday, February 25, 2004 06:52:41 +0000 "David L. Mills" > wrote: > > > Kent, > > > > Depends who you talk to. AT&T had several special-order circuits, > > including 48, 50 and 56-kbps circuits, each individually engineered. The > > 48 and 50 kbps circuits were defintely ananlog, but the 56-kbps circuits > > were digital derived from 64-kbps voice circuits with a supervisory bit > > every 192 bits. It might not make much difference from the viewpoint of > > today and especially because of the ubiquitous nature of AT&T > > penetration of the 1970s. > > > > Dave > > > > kent at icann.org wrote: > >> > >> On Tue, Feb 24, 2004 at 09:39:40PM +0000, David L. Mills wrote: > >> > Chris, > >> > > >> > The IMP interface is described in BBN Report 1822, which I have. There > >> > were companion documents for the Digital Equipment IMP-11A interface > >> > for the PDP-11 Unibus which I don't have. My first ARPAnet connection > >> > at COMSAT Labs was a 4800-bps analog link to BBN; later I connected > >> > with 56-kbps DDS and IMP 29 at Mitre. Toward the end of ARPAlife I > >> > connected from UDel via 56-kbps DDS to IMP 112 somewhere in WashDC. > >> > Scary to think now I have 100-Mbps desktop-desktop just about anywhere > >> > in Internet2. How the heck did we get along with only a 50-kbps > >> > network? > >> > >> I wasn't closely connected at the time, and my memory is dim, but I > >> always thought that the early long line arpanet connections were all 56 > >> kbs. Where does 50 kbs come from? > >> > >> Kent > >> > >> -- > >> Kent Crispin > >> kent at icann.org p: +1 310 823 9358 f: +1 310 823 8649 > >> kent at songbird.com SIP: 81202 at fwd.pulver.com > > > > From casner at packetdesign.com Tue Feb 24 14:50:54 2004 From: casner at packetdesign.com (Stephen Casner) Date: Tue, 24 Feb 2004 14:50:54 -0800 (PST) Subject: [ih] ARPA initial IMP-IMP line speed In-Reply-To: References: Message-ID: <20040224144412.J75277@ash.packetdesign.com> On Tue, 24 Feb 2004, Toru Nohzawa , wrote: > My point is IMP-IMP fact. > I thought there must be some people who handled the IMPs > and they must know the fact. > > If early ARPAnet used the multiplexed or wideband modem for > IMP-IMP communication, it might be new news to many people > and it should be described on the internet history. I do not know details of the modems, but I remember the large cabinet that contained them. This was at ISI from 1973 onward. BBN Report 1822, "Interface Message Processor: Specifications for the Interconnection of a Host and an IMP", shows a picture of that cabinet on page 1-3 and states the dimensions as 68-1/8" x 28" x 28". That document does not describe the modems that connect the IMP to the "wideband communication lines" that interconnect the IMPs. -- Steve From kent at icann.org Tue Feb 24 20:21:07 2004 From: kent at icann.org (kent at icann.org) Date: Tue, 24 Feb 2004 20:21:07 -0800 Subject: [ih] ARPA initial IMP-IMP line speed In-Reply-To: <403BC49C.7DC83FB6@udel.edu> References: <200402242023.i1OKNTnd030835@neverland.cs.utexas.edu> <403BC49C.7DC83FB6@udel.edu> Message-ID: <20040225042106.GJ1668@owl.songbird.com> On Tue, Feb 24, 2004 at 09:39:40PM +0000, David L. Mills wrote: > Chris, > > The IMP interface is described in BBN Report 1822, which I have. There > were companion documents for the Digital Equipment IMP-11A interface for > the PDP-11 Unibus which I don't have. My first ARPAnet connection at > COMSAT Labs was a 4800-bps analog link to BBN; later I connected with > 56-kbps DDS and IMP 29 at Mitre. Toward the end of ARPAlife I connected > from UDel via 56-kbps DDS to IMP 112 somewhere in WashDC. Scary to think > now I have 100-Mbps desktop-desktop just about anywhere in Internet2. > How the heck did we get along with only a 50-kbps network? I wasn't closely connected at the time, and my memory is dim, but I always thought that the early long line arpanet connections were all 56 kbs. Where does 50 kbs come from? Kent -- Kent Crispin kent at icann.org p: +1 310 823 9358 f: +1 310 823 8649 kent at songbird.com SIP: 81202 at fwd.pulver.com From almes at internet2.edu Wed Feb 25 05:11:45 2004 From: almes at internet2.edu (Guy T Almes) Date: Wed, 25 Feb 2004 08:11:45 -0500 Subject: [ih] ARPA initial IMP-IMP line speed In-Reply-To: <403C4639.2ED28536@udel.edu> References: <200402242023.i1OKNTnd030835@neverland.cs.utexas.edu> <403BC49C.7DC83FB6@udel.edu> <20040225042106.GJ1668@owl.songbird.com> <403C4639.2ED28536@udel.edu> Message-ID: <168990000.1077714705@localhost> Dave, As I recall, you got only 56 Kb/s from a 64-Kb/s DS0 circuit by wasting every 8th bit (to avoid long strings of zeroes that messed up the encompassing T1 circuit). The supervisory bit was one bit in every 193-bit T1 frame (24 8-bit chunks plus that pesky supervisory bit). But this was the era of digital service, at least. I'm really enjoying learning more about the pre-digital service era. I recall the monster IMP equipment in the CMU CS machine room and always imagined that squeezing 50 Kb/s out of an analog circuit was pretty tense. Regards, -- Guy --On Wednesday, February 25, 2004 06:52:41 +0000 "David L. Mills" wrote: > Kent, > > Depends who you talk to. AT&T had several special-order circuits, > including 48, 50 and 56-kbps circuits, each individually engineered. The > 48 and 50 kbps circuits were defintely ananlog, but the 56-kbps circuits > were digital derived from 64-kbps voice circuits with a supervisory bit > every 192 bits. It might not make much difference from the viewpoint of > today and especially because of the ubiquitous nature of AT&T > penetration of the 1970s. > > Dave > > kent at icann.org wrote: >> >> On Tue, Feb 24, 2004 at 09:39:40PM +0000, David L. Mills wrote: >> > Chris, >> > >> > The IMP interface is described in BBN Report 1822, which I have. There >> > were companion documents for the Digital Equipment IMP-11A interface >> > for the PDP-11 Unibus which I don't have. My first ARPAnet connection >> > at COMSAT Labs was a 4800-bps analog link to BBN; later I connected >> > with 56-kbps DDS and IMP 29 at Mitre. Toward the end of ARPAlife I >> > connected from UDel via 56-kbps DDS to IMP 112 somewhere in WashDC. >> > Scary to think now I have 100-Mbps desktop-desktop just about anywhere >> > in Internet2. How the heck did we get along with only a 50-kbps >> > network? >> >> I wasn't closely connected at the time, and my memory is dim, but I >> always thought that the early long line arpanet connections were all 56 >> kbs. Where does 50 kbs come from? >> >> Kent >> >> -- >> Kent Crispin >> kent at icann.org p: +1 310 823 9358 f: +1 310 823 8649 >> kent at songbird.com SIP: 81202 at fwd.pulver.com > > From touch at ISI.EDU Thu Feb 26 11:00:47 2004 From: touch at ISI.EDU (Joe Touch) Date: Thu, 26 Feb 2004 11:00:47 -0800 Subject: [ih] info for 'held posts' Message-ID: <403E425F.5070500@isi.edu> Hi, all, Because the list is now subscriber-only in order to reduce spam, there is new info on what to do when a post is 'held' - for ANY reason: WHAT THIS MEANS FOR POSTERS: HELD POSTS ARE DISCARDED All posts go to the same queue - spam, policy-based filters, and non-member posts. This list gets large quickly, and it is not feasible to manually triage. As a result, ALL HELD POSTS ARE DISCARDED. Please see the following on how to repost: --------- IF YOUR POST IS HELD FOR "NON-MEMBER": If you receive a "non member post" reply, please subscribe and repost. Messages held for non-member post review will be accumulated along with spam, which means it will not be possible to actively or regularly review them for content. IF YOUR POST IS HELD FOR ANY OTHER REASON (usually "SUSPICIOUS HEADER"): If you receive a 'held for post' message and are a membe - for any reason (e.g., suspicious header), please forward the message to internet-history-admin at postel.org, and it will be posted indirectly. Such requests should NOT include attachments or be otherwise MIME-encoded, and should be brief. Joe (list admin) From lpress at csudh.edu Thu Feb 26 11:14:03 2004 From: lpress at csudh.edu (Larry Press) Date: Thu, 26 Feb 2004 11:14:03 -0800 Subject: [ih] Documents available on line? References: <20040224144412.J75277@ash.packetdesign.com> Message-ID: <258001c3fc9c$b2ef70c0$c815879b@LPNewPortable> Folks, Several of you have referred to reports in the modem thread. I have a copy of the final "completion" report, but no others. Are the BBN and other reports on the Web? If not, could they be? Larry Press From touch at ISI.EDU Thu Feb 26 14:20:50 2004 From: touch at ISI.EDU (Joe Touch) Date: Thu, 26 Feb 2004 14:20:50 -0800 Subject: [ih] Documents available on line? In-Reply-To: <258001c3fc9c$b2ef70c0$c815879b@LPNewPortable> References: <20040224144412.J75277@ash.packetdesign.com> <258001c3fc9c$b2ef70c0$c815879b@LPNewPortable> Message-ID: <403E7142.9070706@isi.edu> Larry Press wrote: > Folks, > > Several of you have referred to reports in the modem thread. > > I have a copy of the final "completion" report, but no others. Are the BBN > and other reports on the Web? > > If not, could they be? > > Larry Press If anyone has electronic versions, I'd be glad to post them on the Postel Center website.