In the Executive Summary, Deloitte and Touche says: "Opponents [of net neutrality], however, argue that their business models are undermined by bandwidth-hungry applications particularly those with significant video content. Future growth will likely require considerable investment in new infrastructure. But infrastructure owners may believe that they are able to recover too little of the cost to justify investment."
That's just what Kerpen says it says. And in the section titled "The network neutrality debate needs resolution", the report says:
"Those who oppose creating [net neutrality] mandates argue that their business models are being undermined by Internet companies offering bandwidth-hungry services such as video and audio-streaming, heavily networked online games, video-based chat and peer-to-peer downloads. Many ISPs and telecommunications companies would like to start charging content companies, and others, a fee to provide access to their services. There are two primary reasons for this. The first is that ISPs and telecommunications carriers are seeing revenues stagnate. As penetration growth slows, competition drives down prices and rapidly rising Internet use among existing customers erodes margins. The second is that some of the largest Internet companies are enjoying bumper revenue growth and increasing profitability, and carriers would like to use their position in the value chain to participate in this growth.
"Internet usage and traffic are both growing rapidly. There is an increasingly urgent need for new revenues that could fund expansion of the infrastructure on which the Internet runs. For example, on several key intercontinental routes, such as that between Asia and Europe, backbone capacity has grown slower than usage (see Figure 1), and may increasingly struggle to keep pace with demand. Similarly, ISPs and carriers may have to invest in higher capacity infrastructure to continue to be able to provide genuine broadband speeds to consumers and business users.
"Balancing the two sides of [the network neutrality] debate is likely to remain challenging. Both sides have merit; both have their flaws. Clearly, something has to change in the economics of Internet access such that network operators and ISPs can continue to invest in new infrastructure and maintain service quality, and consumers can continue to enjoy the Internet as they know it today." (page 7, section titled The Network Neutrality Debate Needs Resolution.)
The report is about the future of the Internet; both network neutrality and congestion are part of it.
As you've pointed out before, Mike, I don't know where you stand in network neutrality because your position is so nuanced. You say you're in favor of it in principle, but opposed to it in practice or vice versa.
You're under no obligation to read every forecast of Internet usage that comes along, but prudence would dictate that before criticizing someone for misusing any particular one you should have at least a passing acquaintance with its content. You say you relied on Karl's misreading of the D & T report, yet you didn't quote him accurately and he didn't read the report either. So my problem with your post is that it piles ignorance on top of ignorance and contributes to a firestorm of misrepresentation.
When you dash off such seriously under-researched posts as this I wonder at your motivation. Are you paid by the word?
Your buddy Karl is offering the same back-pedal you've some to, this nonsense about the vast array of possible business models, etc. In fact, the argument over net neutrality has all along been limited to about three pricing schemes: the flat-rate system preferred by Google's minions; a tiered service model preferred by the telcos; and a QoS model preferred by various other people as a compromise position. Other pricing schemes may exist in textbooks, but they're not part of this debate and D & T doesn't suggest they ought to be.
Quite simply, net neutrality is a debate over the desire of the telcos to pay for infrastructure enhancements through tiered service revenues. It's not really about free speech, inviolate architectural principles, or free software.
The fact is that net neutrality in the proper sense of the term has never been a creature of law or architecture in the past is true but beside the point: we're under no obligation to leave any technical artifact as it was in the 1.0 release.
Broadband Reports doesn't say what you say they say. BR says: "but they make no mention of network neutrality law fears as the primary reason for the crunch" and you say they make no mention of network neutrality at all. Clearly, you're in error. The second level of "busted" is that BR misleads readers about the D & T report and fails to provide a link to it.
You also claim that D & T doesn't touch on the access network, another bald-faced lie. The report discusses both parts of the infrastructure.
So the question is simply this: where in hell do you get off calling somebody a liar for allegedly misquoting a report THAT YOU HAVEN'T EVEN BOTHERED TO READ????
Now that you've been busted, you're back-pedaling and denying the D & T report says what it clearly says, which is: "Clearly, something has to change in the economics of Internet access such that network operators and ISPs can continue to invest in new infrastructure and maintain service quality,"
Read that slowly a carefully. Unless something changes in the economics of the Internet, service quality will decline. That "change in the economics of the Internet" is exactly what Net Neutrality seeks to prohibit.
“Balancing the two sides of [the network neutrality] debate is likely to remain challenging. Both sides have merit; both have their flaws. Clearly, something has to change in the economics of Internet access such that network operators and ISPs can continue to invest in new infrastructure and maintain service quality, and consumers can continue to enjoy the Internet as they know it today.” (page 7, “The Network Neutrality Debate Needs Resolution”.)
Milke hasn't read the D & T report. It has two full pages on net neutrality, including advice to hotheads to settle down and be reasonable: All those involved in the net neutrality debate need to keep open minds in 2007.
Go here and read it for yourself.
Mike appears to be making stuff up.
Chris says: "the HyperText Tranfer Protocol operates on the Application layer of the OSI model, while the applications that make up the various internet protocols operate at the session level, which all operate over UDP and TCP at the transport level."
Totally. For the benefit of others, why don't you explain something about these session level applications, and how HTTP works without a presentation level protocol (or a session level protocol, for that matter.)
That would be groovy.
Then maybe you can tell why DARPA only decided to fund ARPANET years after it was started.
Rstr5105's comment has got to be one of the all-time greatest examples of a self-cancelling argument I've ever seen. But don't quote me on that, Al Gore might not approve.
Re: Re: Re: Re: Narrowly correct or incorrect, dep
So tell me where you differ from my position on NN.
* I say the Congress doesn't know enough about the operation of packet networks to regulate Internet routing today, and as proof I point to the provisions of Snowe-Dorgan and Markey that ban QoS.
* I say that every privately-funded infrastructure project in the US (electric power, natural gas, phone, and cable) has been funded through the sale of services. Consequently, it's permissible for the fiber optic network to be funded, at least in part, by the Telco having an exclusive right to sell at least some of the services flowing over it, at least for some period of time.
* I say that the only alternative to the services-pay-for-the-infrastructure model is for government to lay the cable, and I don't want that.
Where do you differ?
And the question of competition doesn't imply that we have five fully-developed alternatives to fiber optics today, merely that we have options in the pipeline that can be deployed in fairly short order if there's a need for them.
Your options in Silicon Valley haven't decreased, they've increased. Be honest.
You and I have exactly the same position on Net Neutrality, Mike, so if I'm a "telco supporter" then you must be one too. Welcome to the fold, give me your address so I can make sure you get your checks.
The point is that the viability of WiMax as an alternative to wire for the last mile isn't a function of whether it's running on licensed or unlicensed spectrum, so the error you picked up in the WSJ article is of no consequence. Like a lot of geeks, you get all excited every time a civilian gets one of our facts wrong, whether it's an important fact or not.
Actually, if the WSJ had reported on WiMax's regulatory status correctly, they would have had a stronger case. The last mile is increasingly competitive, and lots of people are working really hard to make that continue to happen.
Right now, the last mile is owned by DSL and cable. If the Barton or Stevens bill passes, there is a third wireline option, and if those three options are not managed correctly, there will be increased pressure from EVDO, WiMax, BPL, and Muni WiFi. And if that's not enough, we'll have 802.21 handoffs among all the players.
You may or may not like the way any of these networks is managed, but to say there is no competition for the last mile is to ignore reality.
Mike, once again you get yourself so torqued-up by a slight misstatement of fact that you miss the larger point. The WSJ says: "WiMax, meanwhile, operates in unlicensed spectrum..." which is literally true, just not the whole story. WiMax was intended for licensed spectrum where transmit power levels can be higher, but it will happily run in unlicensed spectrum too, just not as well.
The larger point of the WSJ piece is that increased competition for the last-mile market makes NN regulations obsolete, if not now then certainly within the next few years.
The principal source of increased competition for the last mile will be wireless, and it will be good enough to keep the Telcos honest. That's the point, don't get too hung up on the details.
Mike says: It may be completely absurd from your technical viewpoint, but it's not at all absurd from a business standpoint. It was that very "flat rate" system that made the internet what it is today. You can argue (as you are) that it caused problems for the network, but that's a double-edged sword. If we had kept charging by the minute (or by bandwidth) in the first place, the internet never would have grown to become as useful and integral as it is today.
You're assuming a lot that hasn't been proved about the why's and wherefore's of the Internet's rise to glory as the only world-wide packet network in town, and I don't care to go there. I would submit that we're not done innovating in the design of packet networks right now in 2006, and we need to allow the experiement to continue.
And I'd also like to point out that it's not for you and me to say what makes business sense and what doesn't - especially for new services that we can't even define right now - but rather that our job is to define what makes regulatory sense. The freedom to fail is a vital part of the efficiency of markets, and just because we think it's dumb for a company to price its services in a particular way doesn't mean it should be criminal.
You're making a strawman argument, Kirt. I never said "all regulation is bad" or that "abuse never happens." I say we have to study the problem we're trying to eliminate before regulating it, and then draft the narrowest possible regulation to prevent it from happening without killing other services.
Let's take an example. Killing is wrong, and we have laws against it. One might draft a law against killing so broadly that we ban such things as killing in self-defense or national defense, killing animals (and even plants) for food, criminal execution, mercy killing, and abortion. And we might treat all forms of impermissible killing the same, regardless of whether they were premeditated, emotional, careless, accidental, or delusional.
So a simple principle that has the most broad popular support ("Thou shall not kill") yields a mass of different laws and regulations depending on intent, circumstances, conflicting rights, and social norms. We don't have enough experience with the regulation of packet networks' routing policies to jump whole-hog into this area without doing great harm.
That's why I advocate collecting more evidence before drafting new regulations.
Re: Mike and Richard Bennett Discussion - clarific
You're overlooking the fact that there are at least two dimensions to broadband service: bandwidth and jitter. We're all used to bandwidth and we accept service tiering on that basis already, as you point out. I can choose dial-up, DSL, cable, or some form of wireless according to bandwidth and understand what I'm getting, how much it costs, and why I choose what I choose.
Consumers are less well-acquainted with jitter, the fundamental element of QoS. It's technically the variation in delay from one packet to the next in a stream of packets, and it's something that packet networks like the Internet don't handle very well, by design. The fundamental tradeoff we made in moving from the circuit-switched PSTN to packet switching was to surrender the jitter guarantee made in the design of the PSTN for the higher bandwidth available in packet-switched networks. That was fine as long as all we did on the Internet was file transfers and services built on the file transfer model such as e-mail and web surfing.
But now we want a network that reserves a little bit of bandwidth for low-jitter, real-time services such as VoIP and video multicasts and then gives us the traditional packet-switched service across its remaining bandwidth.
The telcos, understandably, want to sell the real-time, low jitter service for premium prices and the rest of the network in the traditional way, with the exception that they want to place a cap on each user's consumption of raw bandwidth in order to guarantee that each gets a fair piece of the shared tubes.
Most consumers are so naive about the nature of the Internet - and this extends to the consumer advocacy groups that are pushing net neutrality guarantees - as to think that a 6 Mbps link means each consumer can use all 6 Mbps all the time. That's possible in a circuit switched, PSTN network, but it defeats the economies of packet switching.
Packet-switched networks such as the Internet are built on the insight that traffic is bursty and users don't use all of the network's capacity all of the time. That bursty nature has to be controlled in order to make real-time services work, and it has to be controlled to stop greedy bastards from hogging the whole tube. This wasn't so important back in of the early days of the Internet, but we have different needs and different users today.
Kirt says: You mix comments on fee-for-QoS and QoS alone. I have no problem with QoS offerings on a flat-fee Internet. If we decide that certain services need specific packet handling, let's do it for all comers as part of the service. The proposed laws prohibit the fees, not the function.
He misses my point. I was saying that putting overload protection in TCP killed the TOS model, because it forced TCP-dependence on IP and killed precedence. There is no rational basis for a "one size fits all" pricing model in a system of mixed traffic for mixed use. It's a good thing you're not regulating the auto industry.
And he also makes several fanciful charges about things he imagines "organizations" have done. These charges being almost wholly without factual basis, I won't respond to them in detail, except to say that I'm not opposed to regulation in principle, I just want my regulations to be technically sound and not a bar to progress. The Snowe-Dorgan and Markey laws don't rise to that standard.
Well Mike, that's a nice strawman argument and I can see it makes you feel very virtuous. You've exposed me as a liar without even once proving your point or addressing the issues. What I actually said is that simplifying issues is standard practice in politics, but according to you that's the same as saying "it's OK to lie as long as your side is right." Having God on your side isn't the issue.
But what is the issue? A silly commenter above says "QoS is fine as long as it's free!" to which we simply say "then why not use if for everything at every time? It's not QoS if you do that, of course.
I think this misunderstanding is the essence of the Google position: everybody has to pay the same price for Internet access on the consumer side, and on the content side we should have a bandwidth auction that allows bigger players to get steep discounts and no charge at all for high-value services such as QoS. That's a joke.
We're re-building the Internet, and not for the first time. The original Internet that Kahn and Cerf designed collapsed within two years of going on-line, and was replaced with one that had congestion control in TCP. That one has been patched and re-patched for many years, but is near to collapse. It will be replaced by a network that can handle a broad mix of traffic types and services, some advertising-supported like Google, some by subscription like Skype, Vonage, and the NY Times editorial page, some by ISP subscription like ESPN 360, and some by means we haven't even seen yet. It costs money to run pipes, it costs money to build server farms, and it costs money to produce "content".
It's completely absurd to hit every Internet user with the same bill regardless of how they use the network. It's completely reasonable to separate users into access plans that fit their needs, whether they're on-line gamers, VoIP callers, file downloaders, or web surfers. And it's completely reasonable to allow content producers to buy bandwidth or QoS at wholesale so they can re-sell it to ISP customers who can use it to access special services or content outside their service plan.
Google, Yahoo, MSFT, and the others are arguing for continuation of a status quo under which they've been very successful to prevent upstarts from challenging them by clever adaptation of applications to a more flexible protocol and billing system. They're building massive server farms close by the major hydropower dams on the Columbia River with 100's of thousands of servers, because in the Internet of today whoever produces the traffic controls the flow. The potential stifling of innovation is right there in The Dalles, Oregon.
The Internet that we have today isn't the last word in networking, and it hasn't even kept pace with the LAN and WLAN and WPAN technologies that feed it. We should prepare for massive overhaul of the entire system, and dispensing with the "neutrality" and "end-to-end" foolishness is a good place to start. As the Internet become a richer and more robust playing field, it may hope to one day catch up with the sophistication and utility of the networks people use in their homes and offices today instead of being an albatross around the neck of progress.
Look, Mike, when you try to make this debate all about who's telling the most lies, you're essentially trying to legislate on the basis of moral virtue instead of the actual issues. The telcos may very well be money-grubbing ho's and Google may just as well be shiny and clean, but that doesn't tell us which side is right on the issues. Politics is full of stuff that technical people recognize as "lying" and "dishonesty", but for the most part it's benign. The fundamental problem is that people either don't care about such technical issues or can't appreciate their complexity, so every one of these debates becomes a battle of sound bites at a certain level.
Politicians have a code where they don't call each other "liar", recognizing that simplification is the watchword of politics. So just deal with that instead of getting so excited whenever somebody expresses a thought in a different way than you would.
McCurry is right that Google is in this fight in order to keep their costs down, not because they're the champion of the consumer and of all things good and beautiful.
I will take issue with one thing you've said just above: I have not seen any of the people I know from the protocol standards and design community supporting the net neutrality position, not a single one. In fact, they're all in agreement with me that these regulations are at best premature. And that goes across political lines and includes some Kos diarists. The so-called highly technical supporters of net neutrality are people like Cerf who retired from engineering a long time ago, or people like Berners-Lee who don't actually support the bills in question.
Now we know that bandwidth and QoS are not interchangeable. Sure, you can accomplish a certain level of QoS temporarily by adding bandwidth, but that's like adding memory to a PC, it's only good until the next generation of applications comes along, so we need virtual memory too. And no amount of QoS is going to solve the problem of slow file transfer speeds. But the problem that faces us today is re-engineering the Internet to carry a mix of traffic, far different in its delivery requirements than the stuff we had back in the 80s. So we have to teach the net some new tricks.
And I wasn't calling this a Lefty Blog, I was referring with that remark to real lefty blogs, like mydd.com and dailykos.com.
Don't be so hypersenstive dude, it's only the Internet.
I'm spouting telco talking points? Really? If by "talking point" you mean "inconvenient fact", I suppose that's true. So what? Very few people are capable of dealing with this issue at any sort of deep level, so alas, the public discourse is necessarily dumbed-down.
At least you don't find me saying things like: "As has been pointed out repeatedly, QoS is not the best solution to the problem. More bandwidth works much better."
That's a total non sequitur. How much bandwidth do you have to add to a link in order to ensure that low-volume VoIP always gets low latency? And what happens to the future load on the network when that bandwidth is added? And then how much more do you have to add? That's not a solution, it's prescription for somebody else's economic ruin. Bandwidth is not free.
Trying to make out that Gary Bachula, the public relations director for Internet 2, has given the final solution on the QoS question is to embarrass yourself. QoS has been a hot research area for 30 years, and Bachula's personal opinion is simply one data point in a field with thousands. The engineers who carried out the QoS trial for Internet 2 (five years ago, doncha know) don't even agree with his assessment of their work, and one has come out in favor of the Stevens bill.
It just so happens that on this issue the telcos are right and Google and the Lefty Blogs and the scare groups are wrong. Deal with it.
Reading McCurry's statement in context, it's neither a lie nor a spin, it's simply a simplification. He says: It is only neutral if you are a company like Google and want to sell movies streaming over the Internet. If you're a consumer, it means you pay higher prices so companies don't have to. The "neutral" proposal that companies like Google are touting will ensure that they never have to pay a dime no matter how much bandwidth they use, and consumers who may only use their computers to send e-mail and play Solitaire get to foot the bill.
McCurry is clearly talking about streaming videos over new, QoS-enabled links. While Google may very well pay a lot for raw bandwidth (and do we actually know if their contracts are usage sensitive, as Techdirt Mike asserts? I don't) it doesn't contract with anybody for QoS.
So McCurry is fundamentally correct that Google's law will permit them to use QoS for free, shifting the whole bill to the consumer.
And as for the lies around this issue, the "Save the Internet" crowd tells about 10 for every simplification coming from the anti-regulation side. Really.
On the post: The Internet Will Collapse And It's The Fault Of People Pushing For Network Neutrality?
Net neutrality is all over the report
That's just what Kerpen says it says. And in the section titled "The network neutrality debate needs resolution", the report says:
"Those who oppose creating [net neutrality] mandates argue that their business models are being undermined by Internet companies offering bandwidth-hungry services such as video and audio-streaming, heavily networked online games, video-based chat and peer-to-peer downloads. Many ISPs and telecommunications companies would like to start charging content companies, and others, a fee to provide access to their services. There are two primary reasons for this. The first is that ISPs and telecommunications carriers are seeing revenues stagnate. As penetration growth slows, competition drives down prices and rapidly rising Internet use among existing customers erodes margins. The second is that some of the largest Internet companies are enjoying bumper revenue growth and increasing profitability, and carriers would like to use their position in the value chain to participate in this growth.
"Internet usage and traffic are both growing rapidly. There is an increasingly urgent need for new revenues that could fund expansion of the infrastructure on which the Internet runs. For example, on several key intercontinental routes, such as that between Asia and Europe, backbone capacity has grown slower than usage (see Figure 1), and may increasingly struggle to keep pace with demand. Similarly, ISPs and carriers may have to invest in higher capacity infrastructure to continue to be able to provide genuine broadband speeds to consumers and business users.
"Balancing the two sides of [the network neutrality] debate is likely to remain challenging. Both sides have merit; both have their flaws. Clearly, something has to change in the economics of Internet access such that network operators and ISPs can continue to invest in new infrastructure and maintain service quality, and consumers can continue to enjoy the Internet as they know it today." (page 7, section titled The Network Neutrality Debate Needs Resolution.)
The report is about the future of the Internet; both network neutrality and congestion are part of it.
On the post: The Internet Will Collapse And It's The Fault Of People Pushing For Network Neutrality?
Drawing Mike a picture
You're under no obligation to read every forecast of Internet usage that comes along, but prudence would dictate that before criticizing someone for misusing any particular one you should have at least a passing acquaintance with its content. You say you relied on Karl's misreading of the D & T report, yet you didn't quote him accurately and he didn't read the report either. So my problem with your post is that it piles ignorance on top of ignorance and contributes to a firestorm of misrepresentation.
When you dash off such seriously under-researched posts as this I wonder at your motivation. Are you paid by the word?
Your buddy Karl is offering the same back-pedal you've some to, this nonsense about the vast array of possible business models, etc. In fact, the argument over net neutrality has all along been limited to about three pricing schemes: the flat-rate system preferred by Google's minions; a tiered service model preferred by the telcos; and a QoS model preferred by various other people as a compromise position. Other pricing schemes may exist in textbooks, but they're not part of this debate and D & T doesn't suggest they ought to be.
Quite simply, net neutrality is a debate over the desire of the telcos to pay for infrastructure enhancements through tiered service revenues. It's not really about free speech, inviolate architectural principles, or free software.
The fact is that net neutrality in the proper sense of the term has never been a creature of law or architecture in the past is true but beside the point: we're under no obligation to leave any technical artifact as it was in the 1.0 release.
On the post: The Internet Will Collapse And It's The Fault Of People Pushing For Network Neutrality?
Wrong again
You also claim that D & T doesn't touch on the access network, another bald-faced lie. The report discusses both parts of the infrastructure.
So the question is simply this: where in hell do you get off calling somebody a liar for allegedly misquoting a report THAT YOU HAVEN'T EVEN BOTHERED TO READ????
Now that you've been busted, you're back-pedaling and denying the D & T report says what it clearly says, which is: "Clearly, something has to change in the economics of Internet access such that network operators and ISPs can continue to invest in new infrastructure and maintain service quality,"
Read that slowly a carefully. Unless something changes in the economics of the Internet, service quality will decline. That "change in the economics of the Internet" is exactly what Net Neutrality seeks to prohibit.
Do I need to draw you a picture?
On the post: The Internet Will Collapse And It's The Fault Of People Pushing For Network Neutrality?
You're busted.
“Balancing the two sides of [the network neutrality] debate is likely to remain challenging. Both sides have merit; both have their flaws. Clearly, something has to change in the economics of Internet access such that network operators and ISPs can continue to invest in new infrastructure and maintain service quality, and consumers can continue to enjoy the Internet as they know it today.” (page 7, “The Network Neutrality Debate Needs Resolution”.)
Maslick, you're busted.
On the post: The Internet Will Collapse And It's The Fault Of People Pushing For Network Neutrality?
D & T has two full pages in net neutrality
On the post: Now Net Neutrality Will Ban Ad-Based Network Business Models?
Far out, man
Totally. For the benefit of others, why don't you explain something about these session level applications, and how HTTP works without a presentation level protocol (or a session level protocol, for that matter.)
That would be groovy.
Then maybe you can tell why DARPA only decided to fund ARPANET years after it was started.
Peace out dude.
On the post: Now Net Neutrality Will Ban Ad-Based Network Business Models?
The Brilliance of Rstr5105
On the post: WiMax, Net Neutrality And Basic Fact Checking
Re: Re: Re: Re: Narrowly correct or incorrect, dep
* I say the Congress doesn't know enough about the operation of packet networks to regulate Internet routing today, and as proof I point to the provisions of Snowe-Dorgan and Markey that ban QoS.
* I say that every privately-funded infrastructure project in the US (electric power, natural gas, phone, and cable) has been funded through the sale of services. Consequently, it's permissible for the fiber optic network to be funded, at least in part, by the Telco having an exclusive right to sell at least some of the services flowing over it, at least for some period of time.
* I say that the only alternative to the services-pay-for-the-infrastructure model is for government to lay the cable, and I don't want that.
Where do you differ?
And the question of competition doesn't imply that we have five fully-developed alternatives to fiber optics today, merely that we have options in the pipeline that can be deployed in fairly short order if there's a need for them.
Your options in Silicon Valley haven't decreased, they've increased. Be honest.
On the post: WiMax, Net Neutrality And Basic Fact Checking
Re: Re: Narrowly correct or incorrect, depending
The point is that the viability of WiMax as an alternative to wire for the last mile isn't a function of whether it's running on licensed or unlicensed spectrum, so the error you picked up in the WSJ article is of no consequence. Like a lot of geeks, you get all excited every time a civilian gets one of our facts wrong, whether it's an important fact or not.
Actually, if the WSJ had reported on WiMax's regulatory status correctly, they would have had a stronger case. The last mile is increasingly competitive, and lots of people are working really hard to make that continue to happen.
Right now, the last mile is owned by DSL and cable. If the Barton or Stevens bill passes, there is a third wireline option, and if those three options are not managed correctly, there will be increased pressure from EVDO, WiMax, BPL, and Muni WiFi. And if that's not enough, we'll have 802.21 handoffs among all the players.
You may or may not like the way any of these networks is managed, but to say there is no competition for the last mile is to ignore reality.
On the post: WiMax, Net Neutrality And Basic Fact Checking
Narrowly correct or incorrect, depending
The larger point of the WSJ piece is that increased competition for the last-mile market makes NN regulations obsolete, if not now then certainly within the next few years.
The principal source of increased competition for the last mile will be wireless, and it will be good enough to keep the Telcos honest. That's the point, don't get too hung up on the details.
On the post: Mike McCurry: Will You Pay Google's Bandwidth Bills For The Rest Of This Year?
Re: Re: Re: It's not actually a lie
You're assuming a lot that hasn't been proved about the why's and wherefore's of the Internet's rise to glory as the only world-wide packet network in town, and I don't care to go there. I would submit that we're not done innovating in the design of packet networks right now in 2006, and we need to allow the experiement to continue.
And I'd also like to point out that it's not for you and me to say what makes business sense and what doesn't - especially for new services that we can't even define right now - but rather that our job is to define what makes regulatory sense. The freedom to fail is a vital part of the efficiency of markets, and just because we think it's dumb for a company to price its services in a particular way doesn't mean it should be criminal.
On the post: Mike McCurry: Will You Pay Google's Bandwidth Bills For The Rest Of This Year?
Re: Re: Re: It's not actually a lie
On the post: Mike McCurry: Will You Pay Google's Bandwidth Bills For The Rest Of This Year?
Re: Mike and Richard Bennett Discussion - clarific
Consumers are less well-acquainted with jitter, the fundamental element of QoS. It's technically the variation in delay from one packet to the next in a stream of packets, and it's something that packet networks like the Internet don't handle very well, by design. The fundamental tradeoff we made in moving from the circuit-switched PSTN to packet switching was to surrender the jitter guarantee made in the design of the PSTN for the higher bandwidth available in packet-switched networks. That was fine as long as all we did on the Internet was file transfers and services built on the file transfer model such as e-mail and web surfing.
But now we want a network that reserves a little bit of bandwidth for low-jitter, real-time services such as VoIP and video multicasts and then gives us the traditional packet-switched service across its remaining bandwidth.
The telcos, understandably, want to sell the real-time, low jitter service for premium prices and the rest of the network in the traditional way, with the exception that they want to place a cap on each user's consumption of raw bandwidth in order to guarantee that each gets a fair piece of the shared tubes.
Most consumers are so naive about the nature of the Internet - and this extends to the consumer advocacy groups that are pushing net neutrality guarantees - as to think that a 6 Mbps link means each consumer can use all 6 Mbps all the time. That's possible in a circuit switched, PSTN network, but it defeats the economies of packet switching.
Packet-switched networks such as the Internet are built on the insight that traffic is bursty and users don't use all of the network's capacity all of the time. That bursty nature has to be controlled in order to make real-time services work, and it has to be controlled to stop greedy bastards from hogging the whole tube. This wasn't so important back in of the early days of the Internet, but we have different needs and different users today.
On the post: Mike McCurry: Will You Pay Google's Bandwidth Bills For The Rest Of This Year?
Re: It's not actually a lie
He misses my point. I was saying that putting overload protection in TCP killed the TOS model, because it forced TCP-dependence on IP and killed precedence. There is no rational basis for a "one size fits all" pricing model in a system of mixed traffic for mixed use. It's a good thing you're not regulating the auto industry.
And he also makes several fanciful charges about things he imagines "organizations" have done. These charges being almost wholly without factual basis, I won't respond to them in detail, except to say that I'm not opposed to regulation in principle, I just want my regulations to be technically sound and not a bar to progress. The Snowe-Dorgan and Markey laws don't rise to that standard.
On the post: Mike McCurry: Will You Pay Google's Bandwidth Bills For The Rest Of This Year?
Re: It's not actually a lie
But what is the issue? A silly commenter above says "QoS is fine as long as it's free!" to which we simply say "then why not use if for everything at every time? It's not QoS if you do that, of course.
I think this misunderstanding is the essence of the Google position: everybody has to pay the same price for Internet access on the consumer side, and on the content side we should have a bandwidth auction that allows bigger players to get steep discounts and no charge at all for high-value services such as QoS. That's a joke.
We're re-building the Internet, and not for the first time. The original Internet that Kahn and Cerf designed collapsed within two years of going on-line, and was replaced with one that had congestion control in TCP. That one has been patched and re-patched for many years, but is near to collapse. It will be replaced by a network that can handle a broad mix of traffic types and services, some advertising-supported like Google, some by subscription like Skype, Vonage, and the NY Times editorial page, some by ISP subscription like ESPN 360, and some by means we haven't even seen yet. It costs money to run pipes, it costs money to build server farms, and it costs money to produce "content".
It's completely absurd to hit every Internet user with the same bill regardless of how they use the network. It's completely reasonable to separate users into access plans that fit their needs, whether they're on-line gamers, VoIP callers, file downloaders, or web surfers. And it's completely reasonable to allow content producers to buy bandwidth or QoS at wholesale so they can re-sell it to ISP customers who can use it to access special services or content outside their service plan.
Google, Yahoo, MSFT, and the others are arguing for continuation of a status quo under which they've been very successful to prevent upstarts from challenging them by clever adaptation of applications to a more flexible protocol and billing system. They're building massive server farms close by the major hydropower dams on the Columbia River with 100's of thousands of servers, because in the Internet of today whoever produces the traffic controls the flow. The potential stifling of innovation is right there in The Dalles, Oregon.
The Internet that we have today isn't the last word in networking, and it hasn't even kept pace with the LAN and WLAN and WPAN technologies that feed it. We should prepare for massive overhaul of the entire system, and dispensing with the "neutrality" and "end-to-end" foolishness is a good place to start. As the Internet become a richer and more robust playing field, it may hope to one day catch up with the sophistication and utility of the networks people use in their homes and offices today instead of being an albatross around the neck of progress.
On the post: Mike McCurry: Will You Pay Google's Bandwidth Bills For The Rest Of This Year?
Re: Re: Re: Re: Re: Re: It's not actually a lie
Politicians have a code where they don't call each other "liar", recognizing that simplification is the watchword of politics. So just deal with that instead of getting so excited whenever somebody expresses a thought in a different way than you would.
McCurry is right that Google is in this fight in order to keep their costs down, not because they're the champion of the consumer and of all things good and beautiful.
I will take issue with one thing you've said just above: I have not seen any of the people I know from the protocol standards and design community supporting the net neutrality position, not a single one. In fact, they're all in agreement with me that these regulations are at best premature. And that goes across political lines and includes some Kos diarists. The so-called highly technical supporters of net neutrality are people like Cerf who retired from engineering a long time ago, or people like Berners-Lee who don't actually support the bills in question.
Now we know that bandwidth and QoS are not interchangeable. Sure, you can accomplish a certain level of QoS temporarily by adding bandwidth, but that's like adding memory to a PC, it's only good until the next generation of applications comes along, so we need virtual memory too. And no amount of QoS is going to solve the problem of slow file transfer speeds. But the problem that faces us today is re-engineering the Internet to carry a mix of traffic, far different in its delivery requirements than the stuff we had back in the 80s. So we have to teach the net some new tricks.
And I wasn't calling this a Lefty Blog, I was referring with that remark to real lefty blogs, like mydd.com and dailykos.com.
Don't be so hypersenstive dude, it's only the Internet.
On the post: Mike McCurry: Will You Pay Google's Bandwidth Bills For The Rest Of This Year?
Re: Re: Re: Re: It's not actually a lie
At least you don't find me saying things like: "As has been pointed out repeatedly, QoS is not the best solution to the problem. More bandwidth works much better."
That's a total non sequitur. How much bandwidth do you have to add to a link in order to ensure that low-volume VoIP always gets low latency? And what happens to the future load on the network when that bandwidth is added? And then how much more do you have to add? That's not a solution, it's prescription for somebody else's economic ruin. Bandwidth is not free.
Trying to make out that Gary Bachula, the public relations director for Internet 2, has given the final solution on the QoS question is to embarrass yourself. QoS has been a hot research area for 30 years, and Bachula's personal opinion is simply one data point in a field with thousands. The engineers who carried out the QoS trial for Internet 2 (five years ago, doncha know) don't even agree with his assessment of their work, and one has come out in favor of the Stevens bill.
It just so happens that on this issue the telcos are right and Google and the Lefty Blogs and the scare groups are wrong. Deal with it.
On the post: Mike McCurry: Will You Pay Google's Bandwidth Bills For The Rest Of This Year?
Re: Re: It's not actually a lie
Lie #1. "A new law pending in Congress gives control of the Internet to the telcos."
Lie #2: "The Internet has always been regulated"
Lie #3: "Network neutrality is fundamental to the architecture of the Internet."
Lie #4: "We're grass-roots, they're astroturf"
Lie #5: "Google has never given any money to Moveon.org"
Lie #6: "Google doesn't want a free ride."
Lie #7: "Common carrier regulations enabled the Internet to flourish."
Lie #8: "The last mile has always been governed by common carrier law"
Lie #9: "Ted Stevens doesn't understand the Internet."
Lie #10: "Google speaks for the consumer."
How's that?
On the post: Mike McCurry: Will You Pay Google's Bandwidth Bills For The Rest Of This Year?
Re: Re: It's not actually a lie
On the post: Mike McCurry: Will You Pay Google's Bandwidth Bills For The Rest Of This Year?
It's not actually a lie
McCurry is clearly talking about streaming videos over new, QoS-enabled links. While Google may very well pay a lot for raw bandwidth (and do we actually know if their contracts are usage sensitive, as Techdirt Mike asserts? I don't) it doesn't contract with anybody for QoS.
So McCurry is fundamentally correct that Google's law will permit them to use QoS for free, shifting the whole bill to the consumer.
And as for the lies around this issue, the "Save the Internet" crowd tells about 10 for every simplification coming from the anti-regulation side. Really.
Next >>