NNSquad - Network Neutrality Squad
[ NNSquad ] Re: Is network unneutrality necessarily bad?
On Thu, Nov 08, 2007 at 02:40:15PM -0800, Nick Weaver wrote:
> On Nov 8, 2007 1:36 PM, The Anarcat <anarcat@anarcat.ath.cx> wrote:
> > Why? Because they are in a conflict of interest. Their interest is not
> > to provide good QoS to everyone, it's to maximize their return on
> > investment. So what ISPs usually do is under-provisionning: say they
> > sell 1000 3mbps ADSL connections, they're *not* going to buy an
> > associated 3gbps pipe on the other site. They'll buy less bandwidth and
> > hope that the users don't go all together download stuff at the same
> > time.
> ...
> > So to answer your questions more clearly:
> >
> > Yes, network unneutrality is necessarly bad, unless it's sold as such.
>=20
> So why is underprovisioning necessarily bad?
Did I say this? All I'm saying is that there is currently abuse in that
context. Of course I understand that underprovisionning isn't necessarly
bad, I even agree that at some point it's *necessary* to avoid wasting
resources. But when you get to a point where you can't provide proper
service to your customer base, it's simply abuse.
> If you want a fully provisioned link, you can get one today. But
> there is a reason why a dedicated T1 costs an order of magnitude more
> than your home DSL line. Do you pay $500/month for a T1?
Maybe, but that's not the point. There's a reason why "broadband" is so
popular (and why it's called broadband in the first place), it's because
it's supposed to provide you with a capable link.
> One can argue that the consumer needs visibility into the
> underprovisioning and a way to compare the effective bandwidth from
> their (two) competing providers, but explaining it in a useful way is
> hard (measuring is hard too).
I beg to differ. It's perfectly possible to explain it, even if it would
mean putting it in fine print somewhere. Right now, it's deliberatly
hidden and misconceived in the eyes of the consumer.
> But why should one expect an ISP to not be underprovisioned?
As I mentionned before, because that's what they advertise.
> Any ISP that tried to provision for peak load of all users in the
> worst case would quickly find itself out of business, either from
> losing money on the upstream link or from losing customers to vastly
> cheaper alternatives.
One could also argue that some people are ready to pay the price for
provisionned links, as you mentionned above, and that people interested
in "network neutrality" could just pay extra to get it (which would be a
bit cynical but let's try that path).
The problem with that approach is that it assumes that the broadband
market is a free market with "vastly cheaper alternatives" always
readily available. It's far from being the case. DSL (where I live, at
least) is vastly monopolised by one big provider (Bell Sympatico) that
is forced to resell the service to competitors. Bell always has a profit
margin, even when a competitor provides the service. It's also then less
encourage to provide efficient tech support for line problems, leading
to implicitely less reliable service for competitors.
The cable providers are even worse, having just recently starting
reselling their services for even higher prices.
> And today's "correctly provisioned" can also quickly translate to
> tomorrow's underprovisioned, as usages grow to consume the available
> bandwidth, getting us back to the traffic prioritization problem.
.. which would be simpler if there was a free market and alternatives
to choose from with regards to priorities. Of course, it's not possible
with existing regulations. One could even argue that telecommunication
equipement in urban centers is implicitely monopolistic in nature (ie.
you can't possibly have 3 copper wires getting into your house from 3
different providers) and that priorization, in that context, will
systematically result in conflicts of interests and abuse.
Therefore, the infrastructure must be public or regularized.
--
It is a miracle that curiosity survives formal education
- Albert Einstein