> that should be left to each customer to decide for one's own allocated bandwidth
Bandwidth cannot be conceptualized without a notion of time, which is the essence of QoS. Traffic shaping over time.
I'd argue that consumers choose their QoS when choosing a residential circuit instead of a business class circuit. Business class circuits offer dedicated bandwidth, and residential circuits offer shared bandwidth and are priced accordingly.
> ISPs choose to oversell their residential bandwidth is the ISPs' problem alone
This is not a problem it is the definition of residential bandwidth as compared with dedicated circuit bandwidth which is typically sold to business customers at a higher price. In a business class circuit the traffic typically uses the full capability of the circuit up to the hardware/connection limits.
> customer alone - is the one who ought to decide
I agree circuits like this should be sold, but they would not sell for the same price as residential bandwidth. The only way residential circuits can be sold at a profit is by ISPs doing the arbitrage differently because they are not agreeing to provide a specific amount of peak/average bandwidth 24/7 as they do in a business class contract.
> customer has very few (if any) choices
What better way to identify areas where ISPs are abusing their market power than by observing the QoS settings they use? I'd prefer to let ISPs operate unhindered and then have regulators attack the ones that are specifically abusing their last mile monopolies.
> Too bad the vast majority of the technical people are in favor of net neutrality
This is indeed too bad. It stems from the last mile monopolies, which are a legacy of a heavily regulated telecom space. Note that POTS circuits had a minimum latency requirement and are still lower latency for long distance real time audio than many internet circuits using voip codecs.
I think the name "net neutrality" was wisely chosen and has made many people believe that ISPs do not have an incentive to deliver generally neutral services.
Residential bandwidth needs to be the mass market option without regulatory impediments. It's hard to predict how net neutrality will harm the internet, but over time the harm will be substantial.
"This is not a problem it is the definition of residential bandwidth"
Which is a problem. If an ISP chooses to oversell its bandwidth for a given segment of its residential customer-base, then it's already in the business of deceiving its customers. If the ISPs didn't do this (or at the very least provided some minimum guaranteed bandwidth, with or without a maximum "burst" speed similar to what's happening now for residential customers), then they wouldn't be in a situation where they feel it necessary to be the ones imposing QoS settings upon their customers.
> I agree circuits like this should be sold
It doesn't even have to be a different kind of circuit. Just because the ISP chooses to oversell its bandwidth allocations doesn't mean it has to dictate each residential customer's QoS settings.
If we have a neighborhood of 10 households, each with a 100Mbps plan, all sharing a single 100Mbps pipe, then the question of how the ISP needs to divide this bandwidth at peak hours is simple: divide evenly between the households. Then, each household can decide what to do with the resulting 10Mbps pipe. At non-peak, when maybe two houses are using the Internet heavily, they can then each decide what to do with their respective 50Mbps (or probably a bit less, since there will surely be some mild Internet use in the less-active households).
That means ISPs don't get to force their residential customers to make room for the ISP's own over-the-Internet products, though. What customer would possibly prefer Hulu and Vonage over the ISP's own half-baked shoddy TV and VoIP solutions? The customers obviously don't know what they want!
"I'd prefer to let ISPs operate unhindered and then have regulators attack the ones that are specifically abusing their last mile monopolies."
The fact that ISPs have last mile monopolies in the first place is exactly why they should be regulated like any other public utility. I'd much prefer that actually happen than pretend that the not-actually-free market will magically fix anything (if the last 30 years are anything to go by: it won't).
"has made many people believe that ISPs do not have an incentive to deliver generally neutral services."
Because they don't. Not unless net neutrality is enforced (and even then).
Without net neutrality, ISPs have every incentive to squeeze every last penny out of their customers while not actually improving anything. We as a society have witnessed this for decades. In rural parts (and even quite a few urban parts) of America, this is still an ongoing struggle, and one which I've personally and repeatedly witnessed and experienced firsthand.
> Which is a problem. If an ISP chooses to oversell its bandwidth for a given segment of its residential customer-base, then it's already in the business of deceiving its customers. If the ISPs didn't do this (or at the very least provided some minimum guaranteed bandwidth, with or without a maximum "burst" speed similar to what's happening now for residential customers), then they wouldn't be in a situation where they feel it necessary to be the ones imposing QoS settings upon their customers.
It seems that you oppose the existence of residential bandwidth plans and would prefer that ISPs only offer dedicated bandwidth plans. There is nothing wrong with believing this, but doing so would drastically reduce the value of the last mile circuitry for consumers.
My home internet (cable modem) costs $90/month and I get upwards of 120 Mbps (close to OC3 bandwidth) downstream bandwidth. Upstream is about 50 Mbps. A similarly broadband dedicated circuit would cost thousands per month. I'm certainly not under the impression that I am getting (or should be getting) a dedicated bandwidth level of quality with this cable modem, and I think it's fairly ridiculous to expect that ISPs could offer dedicated bandwidth for the sorts of prices that residential consumers are used to paying.
"It seems that you oppose the existence of residential bandwidth plans and would prefer that ISPs only offer dedicated bandwidth plans"
Not necessarily. Like I said above, even having a certain amount of dedicated bandwidth (even as low as, say, 5Mbps) plus a burst to, say, 100+Mbps during non-peak times would be a lot better than the current situation of no guarantees at all. The exact numbers should be easy to calculate: Min = Total / NumHouses, Burst = Total.
All this is still tangential to my main point, though: of that slice of bandwidth the customer can access during peak hours, the customer - and only the customer - is the only person who should be able to decide how that slice is allocated. If my neighborhood's Internet is really busy and I'm left with 1Mbps, then I should have full control over that 1Mbps. Cable company wants to reserve some of my 1Mbps for its half-baked VoIP offering? Boo hoo; my Counterstrike takes priority.
"for the sorts of prices that residential consumers are used to paying."
Cable companies are already overcharging for Internet plans. They can do this because they have all the benefits of a public utility (namely: the local monopolies) without any of the regulation.
So no, I don't believe for one second that the cable companies can't afford to provide at least some guaranteed minimum bandwidth. They just choose not to because they can get away with it.
Bandwidth cannot be conceptualized without a notion of time, which is the essence of QoS. Traffic shaping over time.
I'd argue that consumers choose their QoS when choosing a residential circuit instead of a business class circuit. Business class circuits offer dedicated bandwidth, and residential circuits offer shared bandwidth and are priced accordingly.
> ISPs choose to oversell their residential bandwidth is the ISPs' problem alone
This is not a problem it is the definition of residential bandwidth as compared with dedicated circuit bandwidth which is typically sold to business customers at a higher price. In a business class circuit the traffic typically uses the full capability of the circuit up to the hardware/connection limits.
> customer alone - is the one who ought to decide
I agree circuits like this should be sold, but they would not sell for the same price as residential bandwidth. The only way residential circuits can be sold at a profit is by ISPs doing the arbitrage differently because they are not agreeing to provide a specific amount of peak/average bandwidth 24/7 as they do in a business class contract.
> customer has very few (if any) choices
What better way to identify areas where ISPs are abusing their market power than by observing the QoS settings they use? I'd prefer to let ISPs operate unhindered and then have regulators attack the ones that are specifically abusing their last mile monopolies.
> Too bad the vast majority of the technical people are in favor of net neutrality
This is indeed too bad. It stems from the last mile monopolies, which are a legacy of a heavily regulated telecom space. Note that POTS circuits had a minimum latency requirement and are still lower latency for long distance real time audio than many internet circuits using voip codecs.
I think the name "net neutrality" was wisely chosen and has made many people believe that ISPs do not have an incentive to deliver generally neutral services.
Residential bandwidth needs to be the mass market option without regulatory impediments. It's hard to predict how net neutrality will harm the internet, but over time the harm will be substantial.