Live data from Hacker News

Netflix Agrees to Pay Comcast to End Web Traffic Jam

online.wsj.com

271–280 of 351 posts

Re: Netflix Agrees to Pay Comcast to End Web Traffic Jam

#272

Earlier quoted context omitted.

For a business...it was $300 cheaper to install Comcast with TV than without as of Oct 2013. I pay $2/month for TV but got free install and I can cancel TV in a year. It seems to me they need another mouth to feed.

Maybe I'm missing something but why would it be cheaper for Comcast to include TV?

It probably means that some mid-level manager at Comcast is catching heat from the home office for letting their TV numbers slip.

At least once a month for the past couple of years, Comcast sends me some junk mail begging for an "Account Review," in which they would add triple-play service to my business Internet account in exchange for a discount on all three. Those letters go in the trash on general principles.

Re: Netflix Agrees to Pay Comcast to End Web Traffic Jam

#273

Earlier quoted context omitted.

The difference is that it's the actual web sites making that decision. ISPs aren't blocking ESPN, but rather ESPN is making the deliberate choice to only allow access to their web site from certain ISPs. This is dumb, I think, and annoying, but there's nothing fundamentally wrong with it. Where you start having problems is when your ISP , who you are supposedly paying for access to the entire internet, starts blockin…

Comcast wasn't blocking or rate limiting Netflix (in this scenario) - they had over utilized circuits with Cogent, who is a transit-provider that sells cut-rate transit of lower quality (and much lower cost) than other transit providers.

I got better and more reliable service from Cogent than I ever got from the likes of Level3, and having both been a Comcast customer AND seen the inside of Comcast, I don't know in what universe it could be called "quality".

Re: Netflix Agrees to Pay Comcast to End Web Traffic Jam

#274
post #263
post #251

Earlier quoted context omitted.

> Just because Comcast and Verizon are sitting there refusing to turn the ports on until you cough up $ (basically a hostage negotiation) doesn't mean you should. Would you please enlighten me as to why an ISP should be required to provide access to a resource they aren't being compensated for?

Because their customers are paying for it.

Nope. ISP customers pay for last-mile infrastructure and access to the connection that the ISP has brokered with larger traffic carriers, whatever that may be. You should read up on peering agreements.

Re: Netflix Agrees to Pay Comcast to End Web Traffic Jam

#275
post #50

Earlier quoted context omitted.

Not when their competitors will do the same. Heck, what competitors? Many areas only have Comcast service!

>> "Many areas only have Comcast service!" Do a lot of people in the US only have access to one or two ISP's?? If so why? I'm in a not particularly urban part of Ireland and I can choose between at least 5 ISP's, decent speeds, and a good variety of prices.

http://www.broadbandmap.gov/number-of-providers/wireline/ser...

Unfortunately, the min/max slider forces hard-coded values during page load, so I can't link to specific graphs. I suggest checking out "served areas"/min:1/max:1 and "unserved areas"/min:2/max:whatever.

Those maps suggest the "1 ISP" set (or any set, really) is fairly uniformly distributed across the nation. In my opinion, it's not quite as biased towards the urban/rural divide I expected.

http://www.broadbandmap.gov/summarize/nationwide

    #ISP %USA
    0    3.3%
    1    8.9%
    2   32.0%
    3   35.6%
    4   14.3%
    5    4.1%
    6    1.3%
    7    0.4%
    8+   0.2%
So it's a minority, but a substantial one. I would also guess that at least "some" of the 2-ISP group have various barriers that make it de facto 1-ISP.

Regardless, 8.9% of the nation is still a LOT of people, and "2 or 3 providers across the nation" isn't really a thriving economy full of healthy competition.

The interesting data points, I think, are the high end. A 4-ISPs isn't THAT different from 1-ISP (8.9% vs 14.3%), and those numbers drop off fast in the 5+ range.

Oh, and yes, I'm excluding radio, in any form ("G[0-9]", various satellite/asymmetric services, etc). Those are a different product entirely, and any similarity in feature set is only superficial. You can't play games or any other low-latency task over most wireless, and VoIP doesn't like the "several second" latencies of satellite, for example.

Re: Netflix Agrees to Pay Comcast to End Web Traffic Jam

#276
Cogent has had peering disputes not only with consumer ISPs like Comcast, but with nearly every other tier one provider out there. Cogent is known to provide very cheap bandwidth but on the flip side they are also known to not be a mutually beneficial partner for settlement free peering. If you have cogent as a provider you should know this. Netflix knows that and thus has offered to provide settlement to Comcast to get better peering rather than be in the overcrowded Cogent link.

Re: Netflix Agrees to Pay Comcast to End Web Traffic Jam

#277
post #274
post #263

Earlier quoted context omitted.

Because their customers are paying for it.

Nope. ISP customers pay for last-mile infrastructure and access to the connection that the ISP has brokered with larger traffic carriers, whatever that may be. You should read up on peering agreements.

And part of that payment is for the ISP's peering agreements with other providers. Why would I pay just to connect to Verizon? I want my connection beyond Verizon to the peer networks hosting Netflix, Facebook, YouTube, etc. No customer would only pay for the last mile if that's all they got.

Re: Netflix Agrees to Pay Comcast to End Web Traffic Jam

#278

Earlier quoted context omitted.

That's already the case if you want websites like ESPN3 or NBC Sports Live Extra.

The difference is that it's the actual web sites making that decision. ISPs aren't blocking ESPN, but rather ESPN is making the deliberate choice to only allow access to their web site from certain ISPs. This is dumb, I think, and annoying, but there's nothing fundamentally wrong with it. Where you start having problems is when your ISP , who you are supposedly paying for access to the entire internet, starts blockin…

I thought most did this due to licensing and 'blackouts' by the leagues. Although I'm not sure if this is correct for ESPN.

Re: Netflix Agrees to Pay Comcast to End Web Traffic Jam

#280
post #274

Earlier quoted context omitted.

Nope. ISP customers pay for last-mile infrastructure and access to the connection that the ISP has brokered with larger traffic carriers, whatever that may be. You should read up on peering agreements.

And part of that payment is for the ISP's peering agreements with other providers. Why would I pay just to connect to Verizon? I want my connection beyond Verizon to the peer networks hosting Netflix, Facebook, YouTube, etc. No customer would only pay for the last mile if that's all they got.

Yes, part of that payment is for access to those negotiated peers. However, just because a customer is paying indirectly for peer access doesn't establish an obligation on the ISP to support arbitrary uncompensated resource consumption on the behalf of their contracted peers. It's an entirely separate contract.

Cogent should pay Comcast market peering costs for the unreciprocated bandwidth they send to Comcast. That's what this whole dispute is about. Arguing that Comcast's refusal to build out and maintain significant infrastructure to support a peering imbalance imposed on them by a peering partner via abuse of a good-faith concession, at their own expense, is somehow "taking hostages" is just silly. Cogent oversold.

I think that Comcast sucks as much as the next guy, but I really don't think they're the bad guy here (as much as it pains me to say that). It looks that way to consumers because Comcast is the only contact affected users have with the whole industry (how many people who have been upset at crappy Netflix quality know who Cogent is, or what a transit agreement is?), but that doesn't mean that's actually an accurate perspective.

Post reply on HN