[ih] Fwd: History of "accounts"

[Amelia A Lewis]

There are a lot of possible responses, here, but Guy's is tempered. 
I'll try to temper mine, as well.

Certainly roads and grazing areas are similar examples to the 
TCP/IP-based internet. It's true that there's no per-user tariff from 
end-to-end, but there are indirect costs, and 'accounts' for entry to 
the commons, at least conceptually.

The end-to-end, or host-to-host definition of TCP/IP made accounting 
irrelevant, rightly. Accounting could happen at the access level (below 
IP, effectively ... like SS7 over Frame Relay (or over anything else, 
for that matter), perhaps best characterized as the charges (metered or 
unmetered) that access providers (ISPs) charged). Accounting could 
happen at the application level (above TCP), but that hardly suggests a 
need for account information to be supplied over IP or TCP. Major 
backbone providers negotiated agreements based on traffic, and modified 
them based on measurement, and charged "lesser entities" (ISPs) based 
on measurement of *their* traffic.

I don't think the absence of a bean-counting-bit impeded the growth of 
the internet, or the remuneration of those who built the infrastructure 
to enable it. In fact, I'd say that *at the level of IP and TCP*, the 
ideology of the shared-information sorts actually enabled a network 
that could be billed at connection- and application-levels.

Amy!
(with apologies for typos; I've got an ouch on a finger and seem to 
*mostly* catch the errors, but if I missed one please account it to 
damage from Dangerous Kitchen Implements)
On Tue, 11 Feb 2014 20:54:52 -0600, Guy Almes wrote:
> Jack,
>   Interesting observations/questions.
> 
>   I recall a meeting at Harvard's Kennedy School sometime about 1990 
> or so with people from several different countries discussing ways to 
> ensure that the net was not overcongested (if I can be forgiven the 
> arguable redundancy).
>   It fascinated me.
> 
>   Some of the speakers were from parts of the world where the net was 
> built around 64kb/s X.25 networks with a *relatively* strong notion 
> of accounting.  At least at the X.25 layer, there were 'calls' and 
> per-packet charges.
>   They argued for the value of accounting.  They were coming from a 
> mindset and reality in which resources were limited and careful 
> accounting and charging was seen as a means of avoiding congestion.
>   I say reality here to include the charging of the PTTs that 
> dominated their world.  They could be forgiven seeing that PTT world 
> as "the way things were".
>   By the way, the very long distance intercontinental circuits were a 
> particular focus of concern at that time.
> 
>   Most of the Americans in the room were fresh off the NSFnet effort 
> in which stoking the infrastructure with bandwidth (between 1987 and 
> 1989 we'd seen the backbone bandwidth increase by a factor of 24 
> (56kb/s to T1) and three years later would see it increase again by a 
> factor of 28 (with T3).
>   They argued for highly scaled best-effort service aimed at avoiding 
> congestion by having lots of bandwidth.
>   Accounting mechanisms, in their view, would add more congestion 
> than it would remove.
> 
>   The Internet of today, largely, reflect that American 
> experience/perspective.  Sometimes it's cheaper to stoke bandwidth in 
> a simple network than to carefully account for usage in a more 
> complex network.
> 
>   Decades later that Internet approach (by no means purely or even 
> primarily American now) continues to work quite well.
>   What accounting there is is done at the "entrance ramps" -- with 
> local ISPs in various different contexts -- and not so much with 
> respect to the wide area infrastructure.
> 
>   Let me not predict the future (especially in the 
> post-net-neutrality era), but I'm glad that the government agencies 
> of 1990 didn't agree to impose accounting at that time.
>   	-- Guy
> 
> On 2/11/14, 6:40 PM, Jack Haverty wrote:
>> 
>> On Tue, Feb 11, 2014 at 1:29 PM, John R. Levine <johnl at iecc.com
>> <mailto:johnl at iecc.com>> wrote:
>> 
>>     There really is nothing new about job accounting for expensive
>>     equipment used for different jobs.
>> 
>> 
>> Very true.   Management is about using resources, and you can't manage
>> what you can't measure.  Fascinating discussion...   it got me thinking
>> about the Internet (this *is* the internet-history list).
>> 
>> In particular, is The Internet the first and only "infrastructure"
>> (widespread resource used by everybody) that has been developed with no
>> associated mechanism for accounting?   I can't recall a single protocol,
>> packet header, or such mechanism, at least from the early days of
>> 70s/80s, that had anything even resembling an "account" field to enable
>> usage to be associated back to some specific "account".
>> 
>> There were some attempts (I pushed on "usage accounting" back in the 80s
>> but it got pretty much ignored), but I think nothing much ever developed
>> ingrained in the Internet architecture.  The culture of the 60s/70s/80s
>> was simply against it, and the ARPANET started, and ended, with no
>> accounting.   Computers had accounting.  ARPANET, and the Internet, do
>> not.  How come?
>> 
>> I also can't offhand think of any other infrastructure without some kind
>> of accounting or at least "feedback" mechanism to make usage visible to
>> the user.   Transportation, energy, etc., all have had such mechanisms
>> from their early days.   Anything that involves sharing a resource is
>> likely to have some kind of accounting.
>> 
>> I think this is changing now in The Internet, as usage skyrockets with
>> video, and cellular carriers notice the costs of provisioning for use of
>> such resources by masses of people as an everyday activity.    Millions
>> of devices that all seem to need their software upgraded daily is
>> probably a factor too.
>> 
>> Is this finally the beginning of Internet accounting?   Is it A Bad
>> Idea?   And is the Internet the first or only infrastructure to make it
>> this far (approaching 1/2 of the world population!) without any such
>> mechanism?
>> 
>> /Jack Haverty
>>