[ NNSquad ] Re: Neutrality in Perspective

["Nick Weaver" <nweaver@gmail.com>]

I must disagree.

On Nov 13, 2007 9:07 AM, Lauren Weinstein <lauren@vortex.com> wrote:
>
> It seems to me that when we get down to brass tacks, there isn't as
> much disagreement about what network neutrality really is as some
> players would have us believe.

No.  I'll give you a list in a subsequent email, and the difference
between "what is kosher" and "what is not kosher" for large classes of
traffic descrimination will be apparently be very different to
different people.

> In essence, most of the arguments are indeed about how much
> non-neutrality is "necessary" or otherwise should be permitted by
> ISPs, and how much collateral damage to innocent users should be
> permitted under any given scenario.

> For example, there is absolutely no good reason why a well-behaved
> residential dynamic service user shouldn't be able to operate their
> own mail servers over port 25.  There are utterly valid privacy and
> security reasons for wanting to do this, not to mention much better
> control over mail handling overall.  The problem comes up when ISPs
> simply declare that an entire class of users can't use this port
> or that protocol as designed, without taking into account the
> variation between users.

This actually seems to be an unrealistic example.  There is no good
reason why a RESIDENTIAL CUSTOMER should have the ability to operate
their own mail server, from the point of view of a rational ISP.

The number of users who have a legitimate reason to operate a service
on port 25, or to receive traffic on the Windows ports, is ~.1% or
less.

What interest does an ISP have in even creating exceptions for
residential customers in this class?

Should .1% of the customers require 10% of the customer service budget?

The changes to customer service training?

The changes to router configurations?

The effect when a customer becomes misconfigured and the ISP's entire
address block (including its own SMTP servers) ends up on a spam
blacklist?!

Or should outbound port 25 and inbound windows ports just simply be
ACLed to /dev/null, and the problem of spam-bots, windows worms, and
other general "health of the internet" issues just goes away?


Rather, such customers who need a static IP and such abilities can pay
an extra $10-50/month for a "business grade" service, and should do
so.  Most areas have plenty of DSL ISPs which offer a business-grade
package, for a higher price, that will provide such benefits as having
customer service representatives know what a "trouble ticket" means.

Even the evil phone company offers business grade services where you
can run your own services, not have ports filtered, and get you people
with a remote clue on the other end of the line.

> Satellite Internet is
> relatively expensive and suffers severe bandwidth limitations at the
> consumer level.

Sattelite internet suffers from completely unusable LATENCY for a
large class of applications.  Even with 100 GB sattelite pipes, it
would be unusable for VoIP, as RTT through a geosynchronous sattelite
link is 560 ms.

> So when we talk about what sorts of restrictions on users are
> reasonable for ISPs to impose, and whether or not any given
> restriction or similar activity by an ISP should be viewed as
> unacceptable, I believe that it's important to keep in mind that the
> ISPs are by and large not innocent bystanders being victimized, but
> to a major extent have themselves created the present environment by
> virtue of their various business-related decisions and motives over
> time.

There are very different business motives for different actions.  See
the subsequent email I will send on the subject.