New Report Says Tiered Broadband Bad, Unlikely


The Free Press issued a report this afternoon casting doubt on the theory of network congestion that has been cited by ISPs as the reason behind P2P blocking or broadband caps, and offering more rational solutions for dealing with sporadic congestion. It also claims that tiered broadband and limitation pricing — in which a carrier charges per gigabyte fee after users exceed a certain cap — is unlikely to become reality. Prior to the report coming out, I had spent the afternoon asking people about this issue, trying to figure out if our series of tubes is really clogged or if the carriers are merely seeking financial and/or competitive gain.

Most people believe and some data shows that not only is the Internet not as congested as the carriers want you to believe, but usage isn’t growing as fast as we’re being told. So when I view efforts such as Frontier’s 5 GB data cap or Bell Canada’s usage pricing for smaller carriers using the Bell Canada network, and offering data over 2GB, I don’t see an honest attempt to deal with network congestion — I see anti-competitive behavior. And Frontier’s cap seems particularly stupid given that Time Warner hasn’t yet begun implementing a tiered system in Frontier’s region and will offer a cap that exceeds 5 GB if it does. As the Free Press report states:

The arguments for the “need” to switch to limitation pricing essentially rest on the premise that we’ve somehow reached a magical bandwidth threshold that throws the entire industry pricing model out the window. We are being told that despite predictable growth, supply can no longer keep up with demand. The old “oversubscription” model has failed, and the only way to recoup costs and manage user behavior is through metered pricing. This seems highly implausible.

I agree, especially when I consider AT&T’s plans to upgrade its network and the cable providers talking about their plans to implement DOCSIS 3.0. If the networks are building out capacity, one would think they’d want customers to use it. Prohibitive caps make that use more expensive, and less accessible to the average users who have driven broadband growth. That’s one reason the Free Press indicates that caps will not gain in favor with ISPs. And if all they’re really trying to do to stifle video competition from sources such as Netflix, Hulu or Amazon, then it’s time for the FCC or Congress to get involved.



well we all know cablevison is a greedy compony . just tack a good loock at the billing prosess , and the people they hire .i am shore this is to rase prophits. i dont trust them . big companys just loock to justify greed . i think its about time they upgrade . if im gonna pay for a gig then its only fair that they giv it to me now !

Allan Leinwand

@Victor – I think I am part of the GigaOm group that you claim is taking a “socialist leaning” on this issue. I am not a socialist, but a venture capitalist.

To be painstakingly clear – I do not advocate a socialist view of broadband. We do not need centrally controlled broadband designed to be equal benefit to all (I assume that is the version of socialism you are referencing). What we need is metered bandwidth offerings that are run by organizations focused on capitalism and efficient operations.

This is basically the metered bandwidth argument from the service providers:
1) p2p and video are crushing our networks
2) we need to buildout our networks to handle the crush
3) to fund this buildout we are going to metered bandwidth to extract more money from p2p and video users

The problem is that they have set the meter far too low to extract a disproportional amount of money from consumers (we’ll all end up paying about 20X per byte delivered based on today’s prices). 5GB is a ridiculously low number and it is around for one specific purpose – to generate revenues to support their over-inflated organizations.

From my point of view, this is not a TCP vs UDP vs RTSP protocol problem, not a conservation vs usage problem, but an problem in the operational efficiencies of the large service providers. They absolutely need to normalize their operations staffs (spend one day inside a large operator and ask operations folks what they do) and work on faster processes to deploy new technologies.

I predict that if the large service providers do implement metered bandwidth you will see CLECs will make a dramatic comeback offering better services to consumers based on newer technologies and more streamlined operations.

That should cause a market dynamic that is very capitalistic in nature – those with low margins and high operating costs will be pushed out of the market. Nothing socialistic about that at all.


Capped broadband is a way to increase profits, not benefit users. I would bet real money that it’s not going to increase speeds for other users, or increase the telco/cableco capital expenditures.

And what’s up with this bogus 0.03% number? Don’t you know that 42.42% of all statistics are made up?

And as far as airbags go, the people who have been killed by them would disagree with Victor.

Victor Blake

99.97% (or more) of the people would like a better service for the same price (we can bet on that) They do want change. They obviously do not have a detailed understand of the problem or how to fix it, nor should we expect them to.
Those some people would gladly see their neghbors unreasonable use curtailed if they could get that for free.
This is the essence of what service providers want to do.
Yes, the 0.03% will not be happy.

Not surprisngly most customers do not ask for the changes they need to benefit them. I don’t recall a big lobby of customers asking for anti-lock brakes or air-bags. Nonetheless they were added to vehicles, customers paid more to get them, and I think we can say — that is a good thing. I don’t think it’s necessary to ask customers if they want better performance of the broadband service. I’m sure that if polled, they will all say “yes.”


“That’s why there are so few all you can eat buffets.”

Obviously you’ve never visited Missouri ;-)

The point the report makes is that caps may make people question the need for broadband, or the higher speed tiers. For instance, I have the highest broadband tier for ATT DSL. I subscribed to the higher tier specifically to watch video. In fact, ATT sells this tier and the one lower specifically for watching video.

If ATT caps, which I agree with the report, really is unlikely, I would immediately a) look for a competitor; failing finding such, b) change my subscription to the lowest broadband speed.

Caps are counter-intuitive when the company already has tiered service, specifically in order to make it easier to use the service.

Dr. Rock

Hey Kevin,

Like all MBA student, you must not have paid close attention in microecon class.

In industries with extremely high fixed cost and extremely low marginal costs (such as telecom), providers have large incentives towards encouraging higher levels of consumption.

For detailed explanation of this, see the infamous paper “Too Expensive to Meter” by U Minn’s Andrew Odlyzko:

It is not true that all suppliers of flat-rate services benefit when users minimize consumption — it depends on the underlying cost structure. Of course all you can eat buffets don’t want you to eat in large quantities, but the food service industry has substantial marginal costs, the exact opposite of the telecom and most other network markets.

Kevin Walsh

Hang on there, Shelley. A statement that “ISPs have a strong incentive to keep people using their services, a feature common in markets with high fixed but very low marginal costs. The more time spent online, the higher the perceived need and value of the service” hardly qualifies as economic analysis. Further, it’s wrong; suppliers of flat-rate services benefit when users minimize consumption, not maximize consumption. That’s why there are so few all you can eat buffets.


Folks, read the report first, then comment.

Using throttling during peak times is covered, in addition to other alternatives (including some already practiced). The report also goes into great detail about why capping isn’t going to solve the problem. Think on it: this means that the beginning of every month, we’ll all have a wide open pipe all the time. Now if congestion is really a problem, does this make sense? Seriously?

Kevin Walsh, the report does cover the issue of expense, by explicitly stating that it’s in the ISPs best interest to encourage use, and increase number of customers–not engage in practices that will only chase us away.

This is a thoughtful, detailed report that considers all these issues, including the possible need to put in some form of control, but control that actually focus on fixing so-called congestion problems, not just arbitrarily generating income. Or arbitrarily inhibiting perceived competitors.

Thanks for posting a link to the report, Stacey.


Wouldn’t it be a better approach that Cap should be there but after crossing the cap service should be downgraded instead of making addtional charges (i.e. priority of traffic from specific broadband connection/ip is lowered). However leaving it to customer that if he wants same QoS after crossing cap then he should to pay additional amount.

I am not sure about US but in UK Service Providers have this approach. Although It didn’t happen at my home but it happened at my office and I completely agree with Service Provider’s argument. We had a server that used to synchronize with US server and downloads several GB of data over weekend. This was causing caps overshoot and we were having very slow network speed in our daily work. When we discussed with operator he told us that you already exceeding the cap for this month, so we downgraded your connection and your normal speed would be resumed in next month.

Kevin Walsh

[…rest of truncated final paragraph…]

An inconvenient economic reality is that without the means to generate returns on capital investment, i.e., increased subscriber revenue, those investments will simply not be made. And anyone watching the rush among digital content owners and storefronts to deliver (increasingly long-form, high-def) content to consumers over broadband connections knows that this will not happen without additional investment.

Kevin Walsh

Broadband providers are unlikely to successfully move to usage-based billing models but not for any of the reasons cited in the Free Press report. The main reason is that customers don’t like them. Businesses that ignore customer desires tend to do poorly.

It is, though, incorrect to state that blocking (Comcast-style) or capping and metering are presented by broadband operators as the only options. Comcast, in an admittedly ham-fisted way, was trying to ameliorate real bandwidth allocation problems. Providers looking at cap-and-meter schemes, on the other hand, are trying to find better ways to monetize their network assets. The two subjects are unrelated.

The Free Press report also ignores the economic realities involved in building and operating massively expensive network infrastructure. I suppose this should be expected; a scan of the ten-member board of directors shows none have any experience running capital-intensive businesses, or any businesses for that matter. Most are academics.

An inconvenient economic reality is that without the means to generate returns on capital investment, i.e., increased subscriber revenue, those investments will simply not be made. And anyone watching the rush among digital content owners and storefronts to deliver (increasingly long-form, high-def) content to consumers over broadband

Wes Felter

Free Press corrected the article, so you can ignore my previous comment. (I’m not looking for a prize, BTW, just accuracy.)

Stacey Higginbotham

Victor, I have never had an issue with network management much like I don’t have a problem with alcohol until someone has a few too many and does something stupid. I recognize that it’s impossible to “support 100% bandwidth 100% of the time from all subscribers simultaneously,” but that’s not what we’re talking about here is it? Broadband isn’t a reservoir that goes dry. Conserving it on Tuesday isn’t going to mean there’s more when I want it on Friday.

There are applications that will take up all available bandwidth and there are ways to deal with it without blocking that traffic or setting up caps. Let’s talk about those options. I welcome your discussion on technical means of managing a network and figuring out how those options might affect both user behavior and the network’s traffic.

Filet O Fish

“The only way to make performance better for approximately 99.97% of the subscribers is to clamp down on the selfish behavior of the 0.03%.”

If such was the case, wouldn’t it be the 99.97% of customers experiencing network issues advocating for change? I don’t hear many broadband customers complaining about that .03%. Rather, it’s the telcos who want to charge more…it appears there may be an ulterior motive.

Bill Dollar

Not sure if folks read the report, but it doesn’t treat bandwidth as unlimited. It just suggests that there are better ways to manage congestion other than blocking or metering:

“Another option is exemplified by Cable One… manages congestion with a “limitation throttling” approach. A user subscribing to its 8Mbps (megabits per second) download/500Kbps (kilobits per second) upload service tier is allowed unlimited data transfers between midnight and noon. Between noon and midnight, they are given limits of 3.6GB (gigabytes or 3600 megabytes) download and 219MB (megabytes) upload. If they exceed these limits (which are the equivalent of maxing out the connection in both directions for one hour), customers are throttled back to Cable One’s “standard speeds” — about half of the maximum speed.

The limitation throttling approach is preferable to limitation pricing, because it has a much more tangible
impact on congestion during peak usage times. It only impacts those who exceed the cap in a short time
window, and it narrowly modifies the behavior of those few users that may be causing the congestion.
Most importantly, it does not select winners and losers on the Internet by targeting specific applications.”

Victor Blake

You have managed to write an article with no actual discussion of the facts about network operations. In shared networks there is unfair usage. Your oversimplification might be a quick route to your opinion, but it is not a constructive contribution to the issue of performance. Are you happy with the performance you have on your residential broadband. If so that’s great. But there are a lot of people that want better performance. The only way to make performance better for approximately 99.97% of the subscribers is to clamp down on the selfish behavior of the 0.03%. To be truthful it may well be that some of those people do NOT even realize the impact they have. All it takes is ONE such subscriber on any given access network, and it will seriously impact the performance of everyone else.

To date I have yet to see a substantive discussion of the technical facts of shared use networks and the technology available to manage traffic.

I’m hopeful that in the future we might see tools that allow subscribers to control their own traffic preferences. This would be a great first step. If you could set your own priorities you might, for example, set your email as a higher priority than your mcafee update or windows update. You might for example, set your web browsing to be a higher priority than your video download. Now comes the tough question. If your neighbor is trying to get a map for directions, is that a higher priority than your video download (which is going to take 2 hours). Is it really so bad that your download might take 2h 1min 20 seconds instead so that your neigbor can get directions to go somewhere. Is that unreasonable.

If your response is that both should work. That’s great — you’ve missed the point entirely. It is not possible to support 100% bandwidth 100% of the time from all subscribers simultaneously. Unfortunatley, some protocols have specific behvaior that lend themselves to dominate available bandwidth. In particular, unattended long connection applicaitons have a distinct statsitical advantage (over short lived interactive applications) given the time they have to adjust their windowing.

I agree with some of the others and I have to say that for the first time in years , I am sorely disappointed at how superficially gigaom has treated this topic.

As much as oil is not a limitless free resource, so to is the multi-billion dollar telecommunications service that you use — not free to build or operate. Even if we ignore the capital cost — I am still disappointed that with a socialist leaning you do not comprehend the concept of sharing.

If you and your neighbors waste water in your community (or if you are in California — electricity — rememeber the blackouts) — what happens when the reservoir goes dry. Mandatory water restrictions. It wouldn’t happen in the first place if you were more careful about your water consumption.


I’m sure this is just another way to squeeze more money out of customers. In order for this to really catch on, more companies have to do it. This would only upset customers so that is another reason why I doubt this will catch on. If enough people were willing to pay more for better ‘net, you know that would do it.


Stacey – so are you and OM running a 2 person relay on broadband business models of operators whose billions are at stake? Stick to covering what you understand (and it is not your billions in their business either). I just finished beating OM with the funny stick a few days ago for his ignorant ramblings (Kevin Martin post) – and you don’t have a clue how it works either. And last but not the least – quoting the “Free Press” does not add any validity to your observations.

Metered services are inevitable – get it through your heads – and find another topic to self-indulge.

Comments are closed.