Hacker News Reader

Comment by cl42 | Hacker News Reader

The Economist's recent article on net neutrality and common carriage is fascinating and related: http://www.economist.com/news/leaders/21641201-why-network-n...

Part of the interesting piece is the history of "common carriage": "The idea that certain businesses are so essential that they must not discriminate between customers is as old as ferries. With only one vessel in town, a boatman was generally not allowed to charge a butcher more than a carpenter to move goods. This concept, called 'common carriage', has served the world well, most recently on the internet."

I've never doubted my support for net neutrality, and the legal history of "common carriage" makes this even more obvious.

View on HN

The Economist article is written with the presumption that there are are problems that need non-neutrality in order to be solved and that there is thus a tradeoff of some sort to be made. They cite better network management, and give as examples latency-sensitive applications that suffer from the lack of guaranteed latency.

Those are fallacies. The lack of guaranteed latency is fundamentally no bigger of a problem than the lack of guaranteed bandwidth. Non-neutral traffic shaping is not necessary to maintain good performance, and in practice a best-effort network can achieve acceptable latency for things like VoIP without having to explicitly identify, classify, and prioritize VoIP protocols. ISPs may be accustomed to using such explicit techniques, but neutral alternatives exist and work better for general purpose connections, so we shouldn't carve out any exceptions here.

The Economist also points out the difficulty of determining when traffic is or should be illegal and blockable. It is not at all apparent to me that ISPs have a need to be involved in subjective decisions like that in order for us to have a functional Internet. Is erring on the side of allowing traffic and prosecuting later really going to cause serious problems, or can we come up with an objective set of guidelines for what constitutes a DoS worth blocking?

Re: requirements of ISPs blocking illegal traffic: http://en.m.wikipedia.org/wiki/Online_Copyright_Infringement... (TL;DR they're exempt, just like e.g. Web hosting companies. It's the "don't shoot the messenger" concept.)

That covers copyright infringement but not criminal activity. They haven't explicitly said what they intend to do with section 230, which provides safe harbor for ISPs that try to block obscenities and offensive content.

People keep using DoS as the canonical example of the need for this, but in that case one of the endpoints is requesting that the traffic be dropped. The ISP isn't making the decision for anybody.

Except that there isn't actually a real and usable protocol for pushing drop rules up to your ISP, so in practice you can't authorize your ISP to block a DoS on your behalf without getting humans involved.

Which is why you pretty much always get humans involved.

How is the ISP even supposed to know it's a DoS otherwise? Maybe you just made the front page of HN and you want the traffic.

>The Economist article is written with the presumption that there are are problems that need non-neutrality in order to be solved and that there is thus a tradeoff of some sort to be made.

This is the absolute classic form of media psychic manipulation: present a false choice and allow endless debate within this window but never question what might lie outside it. Additionally, one side of the debate is accepted as the default and alternatives are only evaluated relative to this one viewpoint.

Another current example is the "work life balance" "debate" as it relates to social media. We are constantly told that we should "unplug" and "find time for ourselves," assuming that the majority of the time we are "plugged" and devoting our energies, via elaborate and exciting abstractions, to some distant venture capital fund.

One relevant term of art here is the https://en.wikipedia.org/wiki/Overton_window

Depending on your taste in writing styles, you might enjoy this: http://thelastpsychiatrist.com/2014/01/randi_zuckerberg.html

> It is not at all apparent to me that ISPs have a need to be involved in subjective decisions like [determining when traffic is or should be illegal and blockable]

What makes you think they are or would be under this proposal?

The announcement said, "These enforceable, bright-line rules will ban paid prioritization, and the blocking and throttling of lawful content and services."

It says nothing about unlawful content. If blocking lawful content is illegal, but not-blocking unlawful content is not illegal, obviously ISPs will err on the side of not blocking grey-area content. It will only be obviously unlawful content that is blocked, which is really no different than today's situation.

The other key fallacy that the telecoms seed the media with is that with common carriage, there is no way to provide quality service.

Net neutrality doesn't mean that you cannot optimize the path to get somewhere, just that you don't discriminate. If you need a lower latency or need to lower utilization on certain links, the ISP can do things like peer with the chatty networks.

But strangely the author of the Economist post concludes that the market should be left to regulate itself. Even stranger, they immediately lament the fact that it's failed to do so in the 75% of America where there is a single broadband provider.

Which makes me wonder - will the FCC action do anything to expand the provider market, or merely ensure the quality of currently available access?

> Which makes me wonder - will the FCC action do anything to expand the provider market,

No, the provider market is unlikely to significantly expand without local loop unbundling; fair access to poles and conduit may help in some cases, but overbuilding is still going to be expensive and limited.

> or merely ensure the quality of currently available access?

We'll have to wait and see what falls under acceptable and unacceptable network management, to see what happens to quality.

I think there's an interesting interplay between this FCC announcement and the earlier one on classifying "broadband" connection speeds of 25 Mbps and above.

If an ISP is required to abide by net neutrality and can only sell broadband when it meets the above requirement, then the FCC is essentially mandating quality and consistency of service.

Does a service meet the requirement when the highest possible rate is 25 Mbps or when the average rate is 25 Mbps? Underprovisioning is standard practice, and it's possible to get only a tiny fraction of what a link is physically capable of.

Also, I don't see anything there about QoS or uptime guarantees. (Uptime guarantee: "One nine." "90%?" "9%.")

Since I can't reply to the reply to this message.

Not sure whether guarantees or provisioning levels are defined, but today's definition of "broadband" is 4 Mbps, so all else being equal, the FCC is trying to modernize the language.

ISPs can still sell sub-broadband; it just doesn't count for the broadband map ( http://www.broadbandmap.gov/ ). This map is also used to determine how much competition there is in the market. So basically DSL no longer counts and many locations now only have one "broadband" provider.

Does this mean that there's a chance that previous decisions that "DSL is broadband and is thus competition to cable" might be able to be turned over, and therefore it might be possible to sue cable companies under anti-trust laws if they act in a way that prevents competition or otherwise abuse their monopoly?

DSL is not broadband; that's a done deal. Now that DSL doesn't count, Comcast has 56% "broadband" market share. http://arstechnica.com/business/2015/01/comcast-now-has-more... (I question these TWC numbers, though, since almost all TWC customers have access to 25 Mbps.)

Reply on news.ycombinator.com