Hacker News Reader

"Comcast has to deliver all this data to tens of millions of customers spread out around the nation."

Yes, this is a definition of an ISP. The ISP transmits requests from users to the internet, and the requested data back to the users. This is the product that Comcast already provides and charges their users for.

The users ask for internet content, the ISP (Comcast) is responsible for delivering it. If the users are clamoring for more content, taking more bandwidth, so be it. Comcast charges end users more money for more bandwidth already, through different pricing Tiers. If their infrastructure isn't big enough to keep up with customers demands then either they aren't meeting their existing contractual obligations with their customers, or they are falling behind technology wise. They have the money to keep up. They tend not to invest enough in their infrastructure, but that is a business issue. If Comcast is unhappy that users are using the service they paid for, they may need to restructure their offering to users, for example paying for bandwidth as a 'per amount transferred'. The critical point here is that it is a relationship only between the ISP (Comcast) and the end user, and the content providers aren't involved.

It is incredibly dangerous to confuse their relationship with their users with outside entities.

Net neutrality doesn't mean Comcast has to handle double the traffic from Netflix. Net neutrality means that Comcast has to handle the traffic from Netflix in the same way it handles it from other places. It could be slow, choppy, and a bad experience. They could choose to invest in a better network or not. They can charge users for more bandwidth, or other related infrastructure services.

What they cannot be allowed to do is discriminate against users and outside companies.

> The critical point here is that it is a relationship only between the ISP (Comcast) and the end user, and the content providers aren't involved.

Isn't comcast also a content provider? They provide cable TV services and own a TV network and a film studio.

Your comment is indicative of why this discussion is so difficult to have and why it's even tougher for the general public to understand or even care.

The way this has always works, and is not changing based on anything 'network neutrality' related, is that generally the sending network (Netflix) pays the receiving network (Comcast) based on something like the 95th percentile of their sending rate. It maybe not be linear, but if Netflix starts sending 2x traffic, they will pay something like 1.8x-2x. That's not changing, and nobody is disputing this.

What we're talking about here is that Comcast could take a user's 20Mbps connection and decide that even though it's fully provisioned in the last mile for 20Mbps, and they have plenty of pipe to get the Netflix data to the Comcast side of the last mile, that Comcast can just decide that Netflix will only get 5Mbps of that 20Mbps, that is unless Netflix pays the ransom money to Comcast to change that arbitrary limit.

> What we're talking about here is that Comcast could take a user's 20Mbps connection and decide that even though it's fully provisioned in the last mile for 20Mbps, and they have plenty of pipe to get the Netflix data to the Comcast side of the last mile, that Comcast can just decide that Netflix will only get 5Mbps of that 20Mbps, that is unless Netflix pays the ransom money to Comcast to change that arbitrary limit.

On the other hand, that's not what happened in the recent Comcast/Netflix dispute, so spinning it that way does not make a good case for net neutrality.

It maybe not be linear, but if Netflix starts sending 2x traffic, they will pay something like 1.8x-2x. That's not changing, and nobody is disputing this.

Are you arguing that Netflix's costs to scale traffic into Comcast are commensurate with Comcast's costs to scale the delivery of that traffic to consumers? I don't think that's true. Didn't Level 3 just blog[1] that they largely operate on a settlement-free basis? Prior to their deal with Comcast, Netflix may have been paying Level 3, and Level 3 may have been paying Comcast, but neither could have been paying the order-of-billions Comcast would need to spend to support, say, doubled Netflix primetime traffic.

[1] http://blog.level3.com/global-connectivity/observations-inte...

Well, there's no gold standard for how any of the contracts are setup, but if I had to guess, it's because you're talking about Level 3 <-> Comcast. The profile of Netflix is far different from Level 3. Generally the fee is based on the net (as in net total, not net internet) traffic.

If Level 3 sends 100Mbps to Comcast, and Comcast sends 100Mbps back to Level 3, or at least approximately even most of the time, they very well may have a settlement free agreement. This is believable because Level 3 is a transit provider, as in if you send them traffic bound to anywhere on the internet, they'll get it there, so the amount of traffic that Level 3 receives from everybody to get to Comcast could be roughly equal to the amount of traffic Comcast wants/needs to route through Level 3 to the rest of the internet.

In the Netflix<->Comcast case, it's not a transit peering. Netflix can only send traffic over that link that is destined for Comcast customers. These peering arrangements are generally drastically cheaper per Mbps than the transit peering. The difference here though is that the only think Comcast sends to Netflix is HTTP request for the most part, and then Netflix sends a video stream back, so there's a HUGE imbalance in the traffic going each direction.

All that said, what Netflix pays to Comcast may scale with Comcast's costs, but there's no direct relationship. To simplify it a bit, Netflix is paying to get traffic to the Comcast central brain. Customers are paying Comcast to get a pipe laid connecting them to the central brain. The two costs are not very directly related.

Basically the answer to this is yes. If Comcast spends all of this money to upgrade their network and the cost to the end user goes up for the better service, then there are really no problems here. That's basically how a market should work. That plan would actually encourage them to upgrade their networks over time and install things like fiber to meet the new demands of customers, instead of providing poor service like they do now.

The problem is that a network that can stream 20Mbit continuously to every customer (so they always can get the "advertised" bandwidth) would be absurdly expensive, and rates would have to go up accordingly. Why should someone who doesn't consume Netflix pay for this? Why should Comcast be afraid to sell you a 100MBit "bursty" connection (essentially the same kind of service they actually advertise now) because they'd be 'forced' to potentially have the capacity to serve 100MBit to every subscriber when Netflix goes to 8K-h.269 or what-have-you?

The problem is that a network that can stream 20Mbit continuously to every customer (so they always can get the "advertised" bandwidth) would be absurdly expensive, and rates would have to go up accordingly

Then maybe they should advertise a different bandwidth?

No doubt. A local company in town recommends customers sign up with some absurd number like 5mpbs per every device connected on a local network. In reality the majority of customers can use the lowest possible plan (in this case 5mpbs) and have trouble free browsing and streaming.

The reality is these ISPs make money by overselling services to customers and not having to deliver on it. And now they're tacking on fees on the backend as well? Absurd.

You're not going to have a good time trying to run multiple concurrent Netflix streams on a 5Mbps connection. That's only enough for a single HD stream, assuming you even get all the bandwidth you're promised.

I'm not sure that the real problem is advertisement consistency. If your only choice is Comcast, and they deliver poor speeds (due to a combination of favoritism and not upgrading) then it's little consolation that you get the advertised low speed.

What difference does it make what they advertise? The existing advertising doesn't guarantee streaming 20Mbit from any host on the internet 24/7. If their advertising were scrupulously accurate, but they delivered the same service, why should you be happier?

twoodin: >The problem is that a network that can stream 20Mbit continuously to every customer (so they always can get the "advertised" bandwidth) would be absurdly expensive, and rates would have to go up accordingly.

dec0dedab0de: >Then maybe they should advertise a different bandwidth?

twoodin: >What difference does it make what they advertise?

If Comcast is having a hard time delivering on what they advertise, as dec0dedab0de said, they should perhaps advertise differently. Offering different plans other then trying to sell a bandwidth that they can not deliver on with it being, as you say, "absurdly expensive."

There was a reason I put "advertised" in scare quotes: Comcast has always advertised that rate as a "peak rate", not something you can get continuously to every host on the internet. Peak rate matters to a lot of people who are using it for things (for example, browsing a multi-MB rich web page) other than streaming, where a consistent rate is more important.

This is sort of like car advertisements, that always touts "350hp V8" or whatever, even though in the vast majority of real-life usage you can use that power for at most a second or two at some freeway onramp or in a passing lane.

Saying that the connection is technically capable of 20Mbit/s is different from guaranteeing 20Mbit/s in any possible usage situation. The problem with my analogy is that people generally have an idea that you can't go faster than maybe 80mph regardless of what car you have, but there is no way to know if the "up to 20Mbit/s" connection means actual usage will show 15, 5, or 1... So yeah, they should somehow be required to tell you, at least at signup, what the typical throughput on your connection will be.

But this is just a problem with awareness. With the rise of Tesla cars, there has been an increase of public awareness of low speed torque. As more people have become aware of solar, more people have become aware of base-load generation. If ISPs started advertising "guaranteed bandwidth" customers would definitely pay attention, especially with the crappy product they've been getting thus far.

So you're ok with buying a 350hp v8 that can only really ever do 50hp? Except on toll roads you have to pay for again that you may use 'up to' 350hp, but you're still not guaranteed that.

The FCC isn't talking about speed that's a red herring. It's about access. What if the FCC say "Fine consumers must have a minimum of 1gbps" you're like "Huzar" and the FCC say "And anyone who pays a huge amount of money get a minimum of 100gps." ..hold the phone

So, charge consumers for what they use. Why does this seem to rarely come up?

The problem with the networks comcast and others currently maintain is they are saturated. And I'm not getting what I pay for as a result. They need a new pricing structure that properly values their product (bandwidth) and charges me for it.

Can anyone explain why the ISPs don't go down this road?

Because they can make more money selling pretty lies than ugly truths.

Of course, that's true for a lot of products, but the US ISP market in particular is problematic because there's little or no competition. And the big ISPs like Comcast spend a lot of money buying political support, so there's little effective regulation.

You're asking why Comcast should be responsible for meeting the expectations they're creating? When your advertising solely conveys "always fast Internet" and "up to 100Mb/s speeds", what exactly do you think people will take away from it.

If Comcast doesn't want to or can't provide these speeds for general use, they shouldn't advertise these speeds for general use. Can you explain the problem with that logic?

> The problem is that a network that can stream 20Mbit continuously to every customer would be absurdly expensive

Will it really? Because every time someone talks about ISP costs, they remark that what's expensive is the last mile. That's the entire reason ISPs claim to be so much more expensive than enterprize providers, so they need higher prices.

Now they claim that the backbone is the expensive part?

> Why should Comcast be afraid to sell you a 100MBit "bursty" connection (essentially the same kind of service they actually advertise now) because they'd be 'forced' to potentially have the capacity to serve 100MBit to every subscriber when Netflix goes to 8K-h.269 or what-have-you?

Who would be "forcing" them to do this? In a sane world they'd already be selling these two types of services as two types of service--you either choose the "bursty" option or you choose the (more expensive) "streaming" option. Then they would know how to provision for each type of user, and they could price each option based on the capabilities of their infrastructure.

Of course, what would happen then, if you are correct, is that the absurdly high price of the "streaming" option would drive their customers to alternative providers wherever possible--or they would be forced to actually spend that extra money on infrastructure. So basically they are trying to obfuscate the issue because actually serving their customers' needs is way too much like work.

It's actually not so expensive relative to the profit that these companies have been making. But they have a monopoly - so there's almost no incentive for them to upgrade their networks.

Exactly, customers in the market can then make cost/value choices for products that directly relate to how that individual ISP is managing costs.

With fast lane proposals, additional cost gets pushed back to successful websites. So for example, when Netflix pays a toll to Comcast, that additional cost is distributed among all the users of Netflix even if they're not doing business with Comcast. How does this work to support a market where consumer choice rewards efficient ISPs? It basically makes the market for network connections look more like the healthcare market where a fundamental problem in it's efficiency is how endpoint costs are highly disconnected from how efficient individual players in the market are operating.

The part that makes this different from healthcare options, is that there are usually no or few options. I know where I live I have the choice of timewarner or fios. That's it. If they decide they are slowing down the sites I use, I'm SOL.

I see a future where we are given a list of "premium" supported sites before making ISP decisions. I wonder if well end up with another dimension of price tier where one dimension is speed of service and the other is what major sites support it.

Comcast's network is designed to deliver a certain amount of data per customer.

It can hit this number when the data is coming from distributed sources.

It should be able to hit this same number even if all the data was coming from 60 interlinks to netflix.

It's not about how fast the netflix access is in absolute terms, it's about whether it's throttled compared to other things.

Comcast is not obligated to speed up their entire network when their customers request more data from a provider, but they should be obligated to peer.

Any argument that talks about the burden of Comcast having to upgrade their intermediate networks inevitably ignores that this is precisely what customers are paying Comcast for, which makes trapped customers rightfully upset.

The phenomenon you describe is best analyzed as the assumptions behind the oversell ratio needing to change - customers' usage is becoming more correlated in time, and they have found a type of service that will expand to use all available bandwidth until each user has >500Mbit.

The appropriate thing to do is to either lower the advertised/supplied bandwidth rates, or to institute metered billing. The first one is a market non-starter as customers shop by large numbers (hence why cable got so popular over DSL). And people are rightfully worried about the second because there's very little competition in this market, and instances of such have appeared to take a punishment-based screw-the-customer approach rather than a selling-more-product-to-customer one.

I think the only legislative solution is to create competition in the consumer ISP space by making last-mile delivery a regulated wholesale utility with many transit providers competing for service.

> But Comcast has to deliver all this data to tens of millions of customers spread out around the nation. While it would be nice if their networks were sufficiently provisioned to serve, say, half their customers 20Mbit/sec continuously during primetime, that is just not the network they have today. It will always be cheaper for Netflix to turn up the spigot than for Comcast to build out its infrastructure. Should "network neutrality" force Comcast to spend billions every time Netflix doubles its streaming rate?

That has nothing to do with intentionally throttling netflix traffic. Network neutrality has never forced Comcast to upgrade its infrastructure, so I'm not sure why that is even a question in your mind. We've had net neutrality for 20 years. Why would it all of a sudden force them to spend billions?

> Should "network neutrality" force Comcast to spend billions every time Netflix doubles its streaming rate?

It wouldn't force an ISP take any action. Each ISP would only be forced to decide how much bandwidth to offer each consumer, and at what price point. If a particular ISP did not feel it wise to upgrade its infrastructure simply because Netflix increased its streaming rate, that ISP would be free not to make the investment.

So how do you deal with network congestion in that scenario? ISPs can continue to offer different pricing plans with different download rates. They can be transparent about the real download speeds for each plan. Perhaps they can even offer a speed estimator tool where the user picks a geographic endpoint and a time of day.

That solves part of the problem of congestion, in that consumer are (roughly) paying for the bandwidth they use. (Theoretically, the users' fees would be scaled to account for the cost of infrastructure upgrades.) But I'd still like to be able to quickly download, say, an API reference page from rubydoc.org without having my connection starved by torrents and video streaming. To that end, perhaps ISPs could grant each consumer a certain GB/month quota of "guaranteed max speed." I.e. until you exceed your quota of X GB per month, your traffic never falls below Y MB/sec, regardless of current network congestion. If you wanted to get really slick, you could even let customers opt certain traffic types out of that program. So I could say, for example, "use my max speed quota for regular HTTP, SSL, etc. usage, but never use it for video."

Ideally, each consumer would have several ISPs to choose from. In that case, the market could discover the optimal plan selections.

Then ISPs should stop claiming speeds they can't deliver.

This is a false choice, born of deception. Comcast's allocations are not zero sum between upgrading and user fees. They're taking record profits every quarter, and customer fees are higher than they've ever been. They are choosing to shoulder the costs of network infrastructure on everyone except their investors.

> Should "network neutrality" force Comcast to spend billions every time Netflix doubles its streaming rate?

This is based on a misunderstanding of what network neutrality means. The question isn't whether Comcast invests in their network but whether they can charge different rates based on the kind of traffic rather than the total volume.

If Comcast wants to skimp on network infrastructure, no problem.

If Comcast wants to charge their customers more for a high speed or uncapped plan, no problem.

If Comcast wants to charge their customers more to watch Netflix than Google Play, problem.

Comcast and other ISPs can charge their subscribers more.

Then they can build out their infrastructure with the new revenue.