Stay informed about the acquisition of Public Interest Registry

by Ethos Capital

Home / News I have a News Tip

Google's Cerf Offers Alternate Strategy to Metered Broadband Billing

In follow up to Friday's FCC ruling against Comcast, Vint Cerf, Google's Chief Internet Evangelist questioned what a reasonable approach would be for broadband networks to manage their Internet traffic? "The real question," he says, is not whether broadband networks need to be managed, "but rather how." Cerf has also expressed concerns over metered billing and suggests "transmission rate caps" instead. He says:

At least one proposal has surfaced that would charge users by the byte after a certain amount of data has been transmitted during a given period. This is a kind of volume cap, which I do not find to be a very useful practice. Given an arbitrary amount of time, one can transfer arbitrarily large amounts of information. Rather than a volume cap, I suggest the introduction of transmission rate caps, which would allow users to purchase access to the Internet at a given minimum data rate and be free to transfer data at at least up to that rate in any way they wish.

Google's Cerf No Fan Of Caps, Metered Billing Broadband Reports
Vint Cerf at it again Broadband Politics
Cerf on managing networks & the need for industry discussion Technology Liberation Front
Google backs ISP-guaranteed minimum data rates Ars Technica
Set speed limits, not weight limits WebProNews

Read full story: Google Inc.

Follow CircleID on

If you are pressed for time ...

... this is for you. More and more professionals are choosing to publish critical posts on CircleID from all corners of the Internet industry. If you find it hard to keep up daily, consider subscribing to our weekly digest. We will provide you a convenient summary report once a week sent directly to your inbox. It's a quick and easy read.

I make a point of reading CircleID. There is no getting around the utility of knowing what thoughtful people are thinking and saying about our industry.

Vinton Cerf, Co-designer of the TCP/IP Protocols & the Architecture of the Internet

Share your comments

The main reason why access ISP don't jeroen  –  Aug 05, 2008 2:41 AM PDT

The main reason why access ISP don't want to go anywhere in this area is simply because they are selling oversubscribed services. When you now get a 20mbit service, you are actually only getting a 1mbit version, in the hope that only 1 out of 20 people actually use the fully speed. Because there are 20 people, it might be that you download at different times and thus that in general of those 20 people, only 10mbit is being used. Thus all seems fine and you think you are getting 20mbit.

Because of this oversubscription prices can be "low", which is needed as the competition is also selling their stuff at "low low" prices, thus if one ISP changes this setup they will be "expensive" and people will go to the other ISP.

Fortunately in the US it seems there is a lot of monopoly (from what I understand) thus ISP's can easily change pricing without having to worry that their users go somewhere else. In eg The Netherlands that won't easily work (then again in the end most ISP's in .nl are owned by KPN anyway).

The fun part about all this end-user pricing schemes is that they don't actually even remotely match what the ISP itself is paying to their transits. They could easily just do the following math trick: (hardware-writeoff + personel + transit-fee + other costs + profit wanted) / number of users = price per month that they need to pay.

But that seems to simple of course to be used :)

speed limits Larry Seltzer  –  Aug 05, 2008 3:22 AM PDT

Another analogy that works I think: speed limits on a highway. If a highway has a 65MPH or (insert metric equivalent here) limit, that doesn't mean that everyone can do that speed at peak traffic times.

Can it be done? Larry Seltzer  –  Aug 05, 2008 3:16 AM PDT

We can talk about theoretical models, but there are a lot of users installed already, a lot of older cable modems out there and a lot of equipment out on poles. What models can be supported without having to rip and replace it all?

just cable? Larry Seltzer  –  Aug 05, 2008 4:27 AM PDT

Let me follow-up my own question: Would I be correct in believing that this is basically a cable modem problem? I've had DSL from two different providers and in both cases always had exactly the rated speed, based on simple tests.

But cable modem is always built around local shared bandwidth on coax. Surely this limits what they can do to guarantee bandwidth.

It's also worth noting that cable modems are largely a US phenomenon. I remember reading several years ago that few countries have any at all, and broadband is mostly on DSL around the world. Feel free to tell me I'm full of it if you know better.

Cable == shared medium then onto ISP jeroen  –  Aug 05, 2008 4:37 AM PDT

Cable == shared medium then onto ISP network
DSL == own copper to the DSLAM (typically the closest phone exchange) then onto ISP network

This is why P2P 'hurts' for Cable providers, both upstream and downstream for the whole neighborhood are completely clogged by 1 user filling it up. If they where just downloading this was not so much of an issue, but because the upstream is full, TCP ACK packets and similar get delayed too, thus hurting performance even more.

DSL can only real get overloaded when the link from the DSLAM to the ISP core network/transit gets full (generally there is ATM @ 622mbit or so on that link, thus at 20mbit per user, you need only 30 odd of those to fill it up ;) but ISP's can easily deploy 10GE connects there too which makes it easier.

DSL is easier to manage though: the link speed limits the user and there is no interference with other users.

For Cable, the providers tend to heavily oversubscribe (we sell 50mbit! ;), add on to that the fact that it is a shared medium, that maybe the provider internal network is not too fast, and you can see those tube clog up pretty fast.

Indeed, the US is primarily cable orientated. Europe (from what I know) is mostly DSL based. In the end one is always sharing some very fat pipes which can also get congested. The question with speed is always "from where to where and over which path and sharing that with how many users and how much is it really doing" and other factors.

Indeed, the US is primarily cable orientated. Larry Seltzer  –  Aug 05, 2008 4:47 AM PDT

Indeed, the US is primarily cable orientated. Europe (from what I know) is mostly DSL based. In the end one is always sharing some very fat pipes which can also get congested. The question with speed is always "from where to where and over which path and sharing that with how many users and how much is it really doing" and other factors.

Right, and with cable modem the shared point is right down at the user's level, so it's more noticed.

So perhaps it's theoretically possible, if DOCSIS management can handle it, to allow rate floors for cable modem users if the number of users on the local segment is limited. If this results in higher rates for everyone I bet they'd much rather have the old system.

That is not the problem at hand jeroen  –  Aug 05, 2008 4:51 AM PDT

That is not the problem at hand here.

The real problem is that the Provider is selling a service and stating "it is 50mbits" and it actually can only do 10mbit if actually used.

The problem thus is that they should not be allowed to market it as such.
With the sue-happyness in the US, I am actually surprised that nobody actually is sueing them yet. Then again, the contract probably contains in little teeny wee letters that it is the 'maximum obtainable' speed.

What is really wrong about their approaches against all these users is though, that they are going to *inspect* the traffic and then discriminate against certain traffic.

Their real solutions:
- lower the announced bandwidth (we do 1mbit)
- bandwidth caps
- let users pay more

advertising Larry Seltzer  –  Aug 05, 2008 5:09 AM PDT

So what do they actually say?

I priced out Comcast cable modem at my house (I use FiOS now) and as offered "Performance with PowerBoost® (Speeds up to 12 Mbps)" and "Blast® (16 Mbps)". The fine print below says: "...Many factors affect speed. Actual speeds vary and are not guaranteed....

You're right, that's too vague. I bet if you dig down far enough you can find an accurate description from Comcast, but that doesn't matter. They need to find an up-front way of conveying better what they are actually selling.

Revolutionary! or not... Chris Snyder  –  Aug 05, 2008 4:09 AM PDT

This made me laugh out loud: rate caps are the way Internet service has been marketed and sold since the days of dialup. Every American ISP I've ever used has sold me a plan with some number of bytes per second up and down the wire.

God bless Vint Cerf for actually suggesting that ISPs live up to that promise! All fine print and hand-waving about over-subscription aside, they should be legally bound to do so.

If over-sub is really a big deal, then at least they have a solid business target for improving their infrastructure: add enough capacity to support x customers at 1.5MB/s. Can't get much simpler than that.

Common practice in Australia The Famous Brett Watson  –  Aug 05, 2008 4:31 AM PDT

Over the years, most Australian broadband ISPs have settled on fixed price plans with data limits and throttling (although not exclusively so). A low-end (but not bottom-end) service might offer 3GB of data per month, plus a bonus 3GB in certain off-peak hours. If the data allocation is exceeded, the connection is throttled back to dial-up rates until the end of the period. The cost per month is fixed. To see the state of domestic broadband in Australia at a glance, see Whirlpool's Broadband Choice.

And actually, the Australian system is pretty jeroen  –  Aug 05, 2008 4:43 AM PDT

And actually, the Australian system is pretty good, as you pay for what you get.
The problem for .au is though that you don't have so much transit links down under to the rest of the world, and the links that exist are very pricy. (And I think to recall that some monopoly down there also has a role in higher prices). If one would take those out of the equation, thus one would apply the model to e.g. Europe, the prices would be much lower (I guess ;)

To post comments, please login or create an account.




Sponsored byThreat Intelligence Platform

IP Addressing

Sponsored byAvenue4 LLC


Sponsored byVerisign

New TLDs

Sponsored byAfilias


Sponsored byWhoisXML API

Domain Names

Sponsored byVerisign

DNS Security

Sponsored byAfilias

Brand Protection

Sponsored byAppdetex