[ih] capacity v bandwidth

[Brian E Carpenter]

Capacity defined as "the number of bits per second which can be transmitted by means of the channel" goes back at least to 1961 as a term of art, according to Shannon's colleague Pierce. A mathematically rigorous definition is in Shannon's 1949 paper but again its unit is bits per second.

I agree that the capacity of a tank is measured in cubic metres and the capacity of a computer memory is measured in bits. But pipeline capacity is measured in cubic metres per second, or barrels per day. Nobody talks about the Strait of Hormuz having a normal bandwidth of 20 million barrels of crude per day.

Regards/Ngā mihi
    Brian Carpenter

On 01-Jun-26 11:55, John Day via Internet-history wrote:
> Like the capacity of a pipeline?
> 
>> On May 31, 2026, at 17:45, Jack Haverty via Internet-history <internet-history at elists.isoc.org> wrote:
>>
>> Getting back to the original capacity<->bandwidth topic...
>>
>> I don't think "capacity" is the right terminology either.  Capacity usually means how much "stuff" can be held, i.e.:
>>
>>
>>            noun
>>
>> 1. The ability to receive, hold, or absorb something.
>>    the storage capacity of a car's trunk.
>> 2. The maximum amount that can be contained.
>>    a bin filled to capacity.
>>
>> So it's appropriate to talk about the capacity of a 5TB drive, or 64GB of memory in a computer.  But "capacity" says nothing about rate of flow of the "stuff" involved, whether it's bits on a wire or gallons of oil in a tank car.
>>
>> As an industry or science or engineering discipline, it seems to me that "networking" has avoided defining and adopting its own terminology for important constructs.   Terms like "bits" are generic.  Terms like "packets" are borrowed from earlier disciplines.   Pony Express riders carried "packets" via horseback:
>>
>> "In the context of thePony Express, the mail packet was called a*mochila*(Spanish for "backpack" or "pouch"). It was a specialized leather saddlebag designed to hold approximately*20 pounds*of mail, with the horse carrying an additional*20 pounds*of material."
>>
>> Back in the 1980s, I remember discussions using terms such as something like "kilobytes per minute per mile".   That reflected the reality of transmitting data over distances.  The further you had to move data, the harder it was to do and the more it cost to do it.  As far as I remember, no one created a term for such metrics.
>>
>> Other disciplines have created terminology.  Medical of course has been quite prolific.  Same with Law.  Even disciplines such as Electricity has unique terminology.  For example, "ampacity" is a term in electrical codes, used to define how much current a particular wire can carry.  That dictates how much power can be moved through the wire.  A wire has lots of relevant parameters that factor into its ampacity.
>>
>> Why hasn't our networking industry developed its own terminology...?
>>
>> So, I don't know when or why "bandwidth" became common.
>>
>> But I do remember the history of another networking term.  In the early Internet, the boxes interconnecting networks were called "gateways".   Today they're called "routers".  But why the change...?
>>
>> In the early/mid 1980s, BBN was selling X.25 networks to all sorts of corporations and government users.  Our small sales force would find prospects.  Sometimes they asked about this new "Internet" thing, and I got called in to explain what it was.  All users were struggling with how to integrate their suddenly affordable PCs and workstations.   The Internet seemed like a possible solution.
>>
>> Our sales force reported back that prospects liked the Internet story, but hated the thought of having "gateways" in their IT world.  Other companies (e.g., IBM) had products called "gateways", and the users had apparently experienced lots of problems with such systems.  The term "gateway" was a potent killer of sales.
>>
>> So we started callig them "routers".  Other companies (cisco, proteon, ...) probably had similar experiences in their sales activities.  Gateways faded into historical oblivion, to be replaced by routers - same products, but much more palatable to customers.
>>
>> What's in a Name?  A lot.
>>
>> /Jack Haverty
>>
>>
>>
>> On 5/31/26 13:57, Brian E Carpenter via Internet-history wrote:
>>> Unfortunately we failed to correct that terminological error in RFC 6214
>>> (the IPv6 version of RFC 1149), and worse we discussed "bandwidth-delay product".
>>>
>>> Regards/Ngā mihi
>>>     Brian Carpenter
>>>
>>> On 01-Jun-26 07:53, D Waitzman via Internet-history wrote:
>>>> In RFC-1149, I wrote "The bandwidth is limited to the leg length."
>>>>
>>>> I stand corrected that I should have used "capacity", but then the pun would not work.
>>>> What's more important, the pun or technically accuracy? :-)
>>>>
>>>> -- 
>>>> David Waitzman
>>>>
>>
>> -- 
>> 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
>