[173173] in North American Network Operators' Group
Re: Verizon Public Policy on Netflix
daemon@ATHENA.MIT.EDU (Owen DeLong)
Fri Jul 18 19:32:45 2014
X-Original-To: nanog@nanog.org
From: Owen DeLong <owen@delong.com>
In-Reply-To: <15828942.6506.1405725122355.JavaMail.root@benjamin.baylink.com>
Date: Fri, 18 Jul 2014 16:25:45 -0700
To: Jay Ashworth <jra@baylink.com>
Cc: NANOG <nanog@nanog.org>
Errors-To: nanog-bounces@nanog.org
On Jul 18, 2014, at 16:12 , Jay Ashworth <jra@baylink.com> wrote:
> ----- Original Message -----
>> From: "Owen DeLong" <owen@delong.com>
>=20
>> On Jul 18, 2014, at 11:32 , Jay Ashworth <jra@baylink.com> wrote:
>>=20
>>> ----- Original Message -----
>>>> From: "Owen DeLong" <owen@delong.com>
>>>=20
>>>> But the part that will really bend your mind is when you realize
>>>> that
>>>> there is no such thing as "THE Internet".
>>>=20
>>> "The Internet as "the largest equivalence class in the reflexive,
>>> transitive, symmetric closure of the relationship 'can be reached by
>>> an IP packet from'"
>>> -- Seth Breidbart.
>>=20
>> Note that the sentence is incomplete
>=20
> It actually isn't, no.
>=20
> The quoted segment is, as noted, a *relationship*; ie: a function =
applied=20
> to a domain of IP addresses to produce a range of other IP addresses; =
it's
> a *function*, and the closure applies it to produce a result.
>=20
>> and as soon as you put something
>> after "from" that is actually meaningful, you end up with different
>> answers for the left hand side of that statement depending on what =
you
>> put at the right hand side.
>>=20
>> Further, even that definition doesn't define a single cohesive entity
>> and the definition of "can be reached by an IP packet" is highly
>> variable and more subjective than you may realize.
>=20
> Not really.
>=20
>> What we commonly refer to as "THE Internet" is really many different
>> equivalence classes similar to what is described above, but each of
>> them is made up of a collection of independently owned and operated
>> networks that happen to cooperate on traffic delivery to varying
>> extents and happen to have agreed to a common protocol and =
participate
>> in some of the same management schemes for things like namespace
>> collision avoidance and address distribution.
>=20
> Hence "transitive". It's not really an accident that "transit" comes
> from the same root.
>=20
> "The Internet" for all the purposes we generally use it here is =
composed
> of all the machines with publicly routable IP addresses between which =
you
> can move packets, regardless of what they're hooked to, or who they =
pay;
> that was the point Seth made in a much more mathematical-sounding way
> in his oft-quoted statement.
And my point is that when you look at it in detail, there's no such =
thing. There are many hosts which have public IP addresses which can =
reach different subsets of "the internet" than other hosts which also =
have public IP addresses and can talk to each other.
It is very easy to choose a selection of hosts and be unable to solve =
that function with a single solution set for the entire set of hosts, =
yet by any vernacular definition of "the internet", all of the hosts in =
question would be "on the internet".
That's my point. The devil is in the details, but in reality, the =
internet is much more precarious, variable, and generally a convenient =
term of art for something that mostly otherwise defies description.
In fact, I've always loved the description of "You can tell how much =
someone understands the detailed workings of the internet by what amazes =
them."
Almost no detailed knowledge: Amazed by everything one can do.
Some detailed knowledge: Amazed by all the different =
places one can reach and how much information is available.
Near complete knowledge: Amazed that it works at all.
Owen
