[ih] Nit-picking an origin story

[John Linn]

Another potential opportunity for speed confusion is the fact that 
modems for use across single analog phone lines subsequently got beyond 
19.2 Kb to reach the 56 Kb range, though this took place in the 1990s, 
long after the original Arpanet lines were deployed.

--jl

On 8/24/25 9:32 AM, Steve Crocker via Internet-history wrote:
> It's very common for people to assume the Arpanet lines were 56Kbps, so it
> might be helpful to explain why they were 50 kbps instead of 56 Kbps.  I'd
> guess a large fraction of the people on this list are familiar with the
> following, but for youngsters who didn't grow up before the telephone
> system became fully digital, here's the brief story.
>
> *Digital* lines are 64 kbps, with 7 bits out of 8 used for data.  Hence 56
> kbps.
>
> Previously, the lines were analog, and data was sent by encoding bits into
> sounds.  The encoder/decoder was called a "modem," standing for
> modulate/demodulate.  The encoding improved over time from 300 bps to 1200,
> and then to 2400 bps.  With careful "conditioning" of the lines and further
> advances in code, the data rates were pushed up to 9600 and then to 19200
> bits per second.
>
> When the Arpanet was designed, one of the offerings from AT&T was 50,000
> bits per second.  They achieved this data rate by using twelve(!) voice
> grade lines and spreading the data across all twelve.  The modem, Western
> Electric series 303A, I believe, was a complicated piece of equipment.  The
> box that held could accommodate four of these modems and had a footprint
> about the same as a refrigerator and stood about half as tall as the IMP.
> There was no relationship between the techniques used to create a 50,000
> bit channel using twelve voice grade lines and the later technique of using
> digital lines to transmit 56,000 bits per second.  Nonetheless, because the
> numbers are close and 56,000 bits per second became the universal building
> block for digital transmission, it's very common for people to assume the
> Arpanet had 56 kbps lines.
>
> (I expect others on this list know more about the details than I do and may
> be able to add some color or correct any errors I've made.)
>
> Steve
>
>
> On Sun, Aug 24, 2025 at 8:44 AM Vint Cerf via Internet-history <
> internet-history at elists.isoc.org> wrote:
>
>> lars is correct
>> v
>>
>>
>> On Sun, Aug 24, 2025 at 8:25 AM Lars Brinkhoff via Internet-history <
>> internet-history at elists.isoc.org> wrote:
>>
>>> Jack Haverty wrote:
>>>> Performance was also an issue as the ARPANET grew and traffic
>>>> increased.  One of the limiting factors to performance was the routing
>>>> algorithm.   Packets were always sent on the "shortest" path.   But
>>>> that meant that the aggregate performance was also limited to
>>>> 56kb/sec, which was the maximum line speed of any path.
>>> I beleve the correct number is 50,000 bits/second.
>>>
>>> ARPANET Information Brochure, from 1980.
>>> "The complete network is formed by interconnecting the nodes through
>>> wideband communication lines, normally 50,000 bits per second (50KBPS),
>>> supplied by common carriers."
>>> https://apps.dtic.mil/sti/pdfs/ADA096798.pdf#page=19
>>>
>>> BBN Report 1928, from 1970.
>>> "In the last quarter, we designed and implemented a test program to
>>> obtain data on the performance of the fifty kilobit communication
>>> circuits"
>>>
>>>
>> http://bärwolff.de/bbn-arpanet-reports-collection/BBN%20(1970)%20Interface%20Message%20Processors%20for%20the%20ARPA%20Computer%20Network%20(Report%201928,%20Quarterly%20Technical%20Report%204).pdf#page=10
>> <http://xn--brwolff-5wa.de/bbn-arpanet-reports-collection/BBN%20(1970)%20Interface%20Message%20Processors%20for%20the%20ARPA%20Computer%20Network%20(Report%201928,%20Quarterly%20Technical%20Report%204).pdf#page=10>
>>> <
>> http://xn--brwolff-5wa.de/bbn-arpanet-reports-collection/BBN%20(1970)%20Interface%20Message%20Processors%20for%20the%20ARPA%20Computer%20Network%20(Report%201928,%20Quarterly%20Technical%20Report%204).pdf#page=10
>>>
>>> This was due to the Bell/Western Electric 303C wideband modem using a
>>> group service of 12 voice circuits.
>>>
>>>
>> https://bitsavers.org/communications/westernElectric/modems/303_Wideband_Data_Stations_Technical_Reference_Aug66.pdf#page=6
>>> --
>>> Internet-history mailing list
>>> Internet-history at elists.isoc.org
>>> https://elists.isoc.org/mailman/listinfo/internet-history
>>> -
>>> Unsubscribe:
>>>
>> https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history
>>
>> --
>> Please send any postal/overnight deliveries to:
>> Vint Cerf
>> Google, LLC
>> 1900 Reston Metro Plaza, 16th Floor
>> Reston, VA 20190
>> +1 (571) 213 1346
>>
>>
>> until further notice
>> --
>> Internet-history mailing list
>> Internet-history at elists.isoc.org
>> https://elists.isoc.org/mailman/listinfo/internet-history
>> -
>> Unsubscribe:
>> https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history
>>
>