Graph of Netflix speeds shows the importance of net neutrality | Hacker News Reader

Graph of Netflix speeds shows the importance of net neutrality

knowmore.washingtonpost.com

82 pointsLibertatea28 comments

I know some websites (like ESPN3) only allow you access if you're on a subscribing ISP.

Could Netflix take a play out of that book and charge a different rate for Comcast customers vs others, and in particular, make it enough of a PR issue that other ISPs would be more willing to cave?

Netflix shouldn't have caved and should have displayed a simple graphic about the slowness in their player along with the number for Comcast support. Displace the anger and support issues around the slowdown to the proper party, the one that's at fault. Cost them some customers.

Of course, the problem with that is that Comcast has a monopoly in many areas they serve. Just like Time Warner does. And folks simply can't switch. So, even though Comcast's customer satisfaction rates are absolutely abysmal, they don't lose customers over it.

Your solution would make more sense. Netflix should charge Comcast customers more and itemize it on the bills as a 'Comcast Network Slowdown Tax'.

I think that's what we all wanted them to do - except for the frustrating few on here that seem to think what happened was right and natural, at least - but I'm not sure it was the correct thing to do for Netflix. I see this as a 'fix it, then make it better' attempt by Netflix. They stemmed the bleeding with a large (the largest?) subscriber base, and now they are looking for alternative long term solutions through public campaigning, product development and probably political means.

At the end of the day Netflix is a business trying to sell product. Taking their ball and going home when the market bullies them seems like corporate martyrdom to me - not a successful growth strategy.

As a Comcast customer I would have called and complained. If enough people complained, it might have had some effect. However, in my area, there's only two choices: crappy DSL from AT&T and Comcast. There's really no comparing the two services. I had the DSL offering for a few years and it was absolute crap. With Comcast I get amazing speeds, but when service is needed, it's Comcast. In other words, it's horrible. Here's what happens when you call them due to an outage:

You tell them you power cycled the modem. They make you do it anyway. You ask them to do a traceroute and they either say "I don't know what that is" or "I can't do that" or "Yeah, we'll get to that after a few other things".

Last time I called, 20 minutes into the call he does a traceroute and finds the problem is a few hops from my house. Yeah, he wouldn't do it it first thing, so he wasted a bunch of time.

Sometimes, if they won't do the traceroute, they'll send someone out, in a 4 hour window, only to tell you when they get there at the problem is not near your house. Or, the problem is fixed by the time they arrive.

It's one of the most frustrating customer service experiences you can have.

I agree that Comcast is horrible but you can't expect them to skip the most basic troubleshooting steps. That probably solves the problem for most non-technical users who call in.

It would be great though if there was an option in general to say, look, I'm an expert, I've tried this, this, and this already. But I see lots of reasons why this wouldn't be practical.

Yeah I think the time you'd have to spend convincing them to skip steps in their script would probably take longer than just doing them.

A traceroute takes a few seconds. 99% of the time I call it's mainly to tell them to please track down the problem (outside of my home... only once over the years was it my modem) and fix it. To have to wade through 20 minutes of prelims, when a 20 second command could make it obviously unnecessary, that's just dumb.

The fact that you thought I was suggesting skipping the basics is bewildering.

I deal with the same issue on Verizon. I found that using LiveChat, the support there seemed to be more tech. focused. If I start by sending a screen shot of my traceroute, I usually can get what I need within a few minutes.

You must keep in mind that all it takes is a another player attempting to enter the movie streaming market (e.g. Amazon) that is willing to pay the ISPs. This just creates a whole slew of issues that Netflix might not be willing to deal with.

Definitely an interesting idea, retaliating in the same spirit. But would it help Netflix? Switching costs for ISPs are higher than switching costs for streaming services (like Netflix or Amazon), not just in terms of dollars but time and hassle. I don't know that many people would be willing to switch ISPs just because Netflix costs a bit more. And in some areas, Comcast is the only ISP.

Nothing spurs innovation more than a challenge. Maybe we'll see it with large-scale peer-to-peer technology that Netflix may help innovate to respond to higher streaming costs, or it may be an opportunity for an different ISP model to enter the market or expand their existing footprint (like Google).

The dependency of the Internet on a small number of ISPs is an anachronistic, unsustainable requirement. When the Internet was first starting, they were necessary to lay cable and maintain network infrastructure. As a nice consequence of building the infrastructure, the ISPs got to charge for routing traffic as well. But this makes no sense. Why should the ISP's have so much bandwidth routing power?

Routing inherently lends itself to decentralized algorithms, and the research is starting to catch up to the ISP's. Mesh networks are growing in popularity, and once they reach a critical mass they will be unstoppable. Expect to see a rise of reliance on mesh networking in universities and urban centers.

BitCoin is going to revolutionize bandwidth routing. In decentralized routing algorithms, payment for bandwidth is a difficult problem to solve because it depends on centralized components of the system. But BitCoin enables 1) micropayments, and 2) distributed transaction storage, which both benefit bandwidth routing research. Some point soon, routing will be completely decentralized, and the infrastructure providers will receive micropayments of BitCoin in return.

As my senior thesis, I'm researching "TorCoin," a proof-of-bandwidth cryptocurrency mined by transmitting bandwidth over the Tor network. This summer I am working to apply this research in a business environment. If any of this interests you, definitely reach out to me and we'll schedule a chat: miles.richardson@yale.edu

Planet Money recently did a really great show [1] about the ISP routing power in the US vs. other countries, and what led to those decisions. The irony is that they were doing everything they could to not be anachronistic, and to plan for the future. The US chose poorly. Anyway, it's worth a listen.

[1] http://www.npr.org/blogs/money/2014/04/04/299060527/episode-...

Algorithms can't change the fact that backbone providers control the main pipes. Smaller pipes and indirect routes are less efficient.

IMHO I think the best response to that would be for Netflix to charge the ISPs that want to carry their service. Reverse subscription if you will. The demand is there, ISPs would be reluctant to alienate the customers even more.

Or you know, not allowing more consolidation in the cable industry.

Comcast's CEO seemed to indicate that Netflix did this on purpose. Yeah, likely.

This is what happens when there is an effective monopoly and it is applied. But seeing as how Windows roamed free for decades, the US doesn't seem to have any problem with huge monopolies.

I don't understand what is "hilarious" about this graph.

Is it me or the graph in the post doesn't say anything about actual speeds, just their changes since Jan 2013.. misleading title?

That is the most irritating page layout I have come across yet. What is that red bar even there for?

"The data are from Netflix."

Really, Washington Post?!

Is it actually wrong? Data is plural.

I kind of assumed it was one of those 'correct grammar' things like 'an historical' that looks wrong but isn't.

Correct, data is a plural. OP either doesn't know that or is making a more nuanced complaint about stodgy old style guides that used Latin to decide things like this.

"An historical" actually is 'wrong' (for as wrong as anything can be in English, which generally isn't very). "Historical" doesn't make a vowel sound when read (which is the rule).

It does if your dialect has a silent H there. A lot of people that pronounce "an historical" do so with a silent H.

Maybe not silent, but definitely muted. It's a soft "H" and probably borrows its pronunciation from Spanish.

Data was plural. It's definitely singular now. The old way is not necessarily "wrong", just awkward. Correcting people to use the old way is wrong, however.

Remember, if you're pedantically holding on to archaic conventions for data, and you don't use "agenda" as plural, you're a hypocrite ;)

Reply on news.ycombinator.com