Hacker News Reader

Comment by pcurve | Hacker News Reader

Those goofy diagrams are wonderfully effective at explaining the situation.

Worth noting is that Cogent is small potato and basically has no leverage. It barely does $300 mil in revenue. Compared that to most Tier 1 providers which do $10 bil+.

View on HN

That's the part that makes me suspect the talk about this harming new entrants, startups, etc. is overblown. Afaict, the situation for small players is the same as it's always been: you pay for transit from an ISP with good connectivity, and you get good connectivity. If you have some colo racks somewhere with high-quality connectivity, your customers on just about any network have good access to your content. If not, you find a better colo facility.

The controversy here seems to be over what happens when you're a massive player and looking to get your bandwidth for less than typical retail transit prices. Then it enters the whole game of peering politics, which has changed considerably over the past 10 years. But if you're a relatively small startup with some colo racks, I don't see how that game is any more relevant today than 10 years ago: you still just buy transit. I mean, I personally have no trouble delivering my modest amount of content to Comcast users, and I've never paid Comcast.

I do think there is a general barrier to entry on the internet, because large players (YouTube, etc.) get free transit from peering agreements while new entrants don't, so any YouTube competitor is at a huge transit-cost disadvantage to Google. But that's a fundamental problem with the way the internet backbone has been built out of a mesh of private peering agreements, ever since it abandoned having a single neutral backbone run as a utility (originally by the NSF). I think going private was a mistake and benefits both big players and politically savvy ones, at the expense of new entrants paying commodity, but that's a (very) big issue to fix at this point. It's not an issue of "net neutrality", though: the traditional peering system is inherently non-neutral, not utility-esque.

You fail to see the big picture, most people do.

This is not unlike government regulation that "exempts" small business....

All this does is put an effective ceiling on a business. You will never be allowed to grow beyond X with out having to pay your protection money.

This hurts startups because VC and other investors want the business to grow beyond X and knowing their is an upper limit to the growth will make it very hard if not impossible to get investment.

No it is not that. The internet is largely a collection of privately owned network that has interconnects where networks meet. At that meeting point you can either pay for transit on another network, or agree to settlement free peering if it is mutually agreeable.

Google "Cogent peering dispute" you will see they have had problems with nearly every tier 1 out there. They sell bandwidth very cheap and then try to work out settlement free agreements. The problem is their network often times ends up pushing much more traffic onto the peer than they deliver themselves.

Netflix is a customer of Cogent's and as many customers of them know they often have disputes. Netflix is large enough now where they can pay at interconnects for transit and not have to deal with Cogent's oversubscribed network.

This is not the death of net neutrality. This is how the internet has always worked. To peer settlement free the networks need to be on near equal terms. If you are a small business you simply pay someone who has this peering worked out.

In this instance Comcast is not discriminating against traffic from Netflix in particular, it is discriminating against a peer network that is not paying settlement and providing a non mutually beneficial connection. That is business and is how the internet has always worked since it left the NSF.

What exactly does "mutually beneficial" mean, though? It doesn't necessarily mean traffic is flowing in both directions. Theoretically, when Netflix delivers traffic to residential customers, Netflix benefits through subscription fees, and the residential ISP benefits in that it is able to offer a fast internet experience to its paying customers. If it were easy to switch ISPs, then customers might switch to those with better Netflix streaming, giving Netflix bargaining power. Unfortunately, not only is it not easy to switch, but many areas have effective ISP monopolies... and on the other side, switching from Netflix to competing video services is easy, so customers are much more likely to punish Netflix than their ISP for bad interconnection. But that arguably has elements of market distortion to it.

Exactly. Cogent isn't shoving a burden onto Comcast because that "burden" is part of Comcast's value proposition to its customers. Calling it a "burden" is like saying that Zappos is placing a burden on UPS and should pay above the standard rate to have their product delivered.

Of course, UPS doesn't have a monopoly, so they can't get away with that crap. But they reveal the truth: this is all about leverage.

> Unfortunately, not only is it not easy to switch

More to the point, in the US there's often only two players, a DSL and a cable provider, so little meaningful choice.

Mutually agreeable would be something like this: Comcast-we provide access to 30 million residential customers for you to deliver content to. Level 3: We provide connectivity to Europe over our Apollo cable, connectivity to ..... on and on

Forget about Netflix here for a moment this is a network agreement. If you are Cogent Comcast can say: You are providing bulk bandwidth cheaply and delivering a huge amount of traffic to our network. You are not providing our customers or us with significantly greater connectivity, and far less of our traffic is going over your pipes. Pay us or we are not going to take more of your traffic.

As an end user, I say to Comcast: I'm paying you for access to a bunch of stuff on Cogent's network. Make sure you have adequate peering with them.

But you have little leverage over Comcast, even collectively, if they're the only cable provider in your neighborhood.

Which is why government intervention is now an imperative.

I wonder if part of the value equation to Google for Google Fiber is to help balance their traffic.

I am perfectly Aware of how "opeering" works, am I also perfectly aware how that is a bull shit excuse made up by people like Verizon to justify double billing

There is absolutely zero chance Cogent will ever recieve more data from Verizon then Verizon receives from cogent.

So the Teir 1 Providers that also just so happen to be last Mile Providers have created this nice little scam whereby they can double bill and call it "peering"

Netflix accounts for 30-50% of traffic these days yet only now have they needed to cut out a low quality middle-man to provide their service. I doubt startups have anything to worry about. If you manage to become 50% of Internet traffic you can toss the biggest ISP in the US some money to stay in business surely.

"Toss them some money" will turn into "lie down and let them bleed you dry" sooner or later once Comcast decides to launch a competing service.

Yes, and I think Netflix already knows that's inevitable, hence their heavy investment in developing original content.

Not when the media monopoly is taking almost all of your revenue in "license" fees

Netflix does not really make that much in profit. I see a rate hike coming soon

I have no idea how Tier 1 providers work but is it possible that in the future (5 - 10 years) more will be sprung up?

Or is it so costly that it will eventually be a single tier 1 in the future?

There can be only one. The "competitive marketplace" is a fancy name for modern gladiatorial combat, played out with pens and contract forms instead of swords and shields.

How many Tier 1 providers that aren't ILEC's, Cable or Wireless providers are doing $10B in rev? Akamai did $1.6B in rev last year. Edit: Level 3 does around $6B - that's the highest I've found of the independent players.

Reply on news.ycombinator.com