it always amazes me when people talk about content delivery on the internet. They discuss it as if there is unlimited bandwidth available. Well guess what. There is not.
Sure, new bandwidth is being added onnetworks every day. But guess what, our ability to consume bandwidth is growing far, far faster than the speed at which it is being added. Call it digital gravity. The bigger and more powerful our PCs become, the more specialized processors that are enabled with internet connectivity, the more bandwidth we all consume.
There are some basic facts about the internet that remind me of driving on the 405 in Los Angeles. Traffic jams happen. There is no end in sight for those traffic jams. The traffic jams are worse at certain times of the day. Whether its the 405 or the internet.
Uless of course we add multiple tiers of service so that users, companies and applications that want to, or need to avoid those traffic jams have alternatives. We need HOV lanes and toll roads on the net as badly as we needthe HOV lanes on the 405.
Take this to the bank. The more we upload and download and share:
standard definition video, high definition video,home movies in DV and eventually HDV,multiple megabit photos the more bandwidth weconsume.The more PCs and servers we backup online, themoreWeb2.0 applications we use, the more new database applications come online, the more bandwidth we consume. The more bandwidth we consume, the more internet traffic jams we have. The more internet traffic jams we have, the worse our internet applications perform.
If the holy grail to some is any video content in high definition, anytime anywhere from the internet , well guess what, your 15mbs download speed is going to crawl like a 14k hayes modemduring peak use times. You think its bad now, having to wait overnight for that video to download. It can get a lot worse.
We can try all the tricks we want. Edge servers, peer to peer, it wont matter. Just like a 20 lane highway is still going to have gridlock if enough cars use it, so will the net.
But wait, theres more. That doesnt even account for the problems and hassles of network providers exchanging traffic.
Which leads to the foolishness of the information wants to be free movement. Video on the net is a nice to have application. Self publishing is a nice to have application, whether video or any other format. For our entertainment driven society, it seems to be the low hanging fruit that realizes the value of the net. Wrong.
The internet is a great enabler and equalizer not because if can do the easy stuff. Not because it can replace photocopied newsletter of the 70s, or provide an alternative to VHS cassettes, or provide an alternative to satellite or terrestial radio or tv, but because it can help people in ways that can change and savetheir lives.
Medical and home diagnostic applications require bandwidth. They also require a quality of service that cant be interrupted because little Johnny down the road is trying to download the entire NBC schedule for his freshman highschool class. To enable mission critical applications, you have to have mission critical reliability. And that mission critical reliability has to be able to reach any home that a broadband connection can reach. To do that you need multiple tiers of service.
I would rather have little Johnnys grandma getting priority for her video checkup with the doctor at the hospital over little Johnnygetting his bandwidth to upload the video of the prank he pulled onhis buddy.. I would rather make sure that information from life support or other important monitoring equipment, medical or otherwise is finding its way without interruption, and without the end user having to pay for an off the net solution. These are the applications that make the net great. These are the applicatins that offer equal opportunity to those who are disadvantaged.
I want the telcos and the cable companies and the wireless companies to work out a way to exchange traffic at multiple quality of service levels.
At that point, the internet becomes a viable means for important applications it cant support today. We will see a number of new applications developed, whether medical or otherwise, that cant be put in place today, simply because there are definable levels of service.
This doesnt mean that we all cant do all the fun things wecurrently use the net for. It just means that we will watch applications zip by us in the internet HOV lane. Yes, some of those applications will be commercial and geared towards non mission critical apps. If the networks are smart, they will account for this and prioritize applications within the “HOV” lane. Yes, it will mean some content will cost more if we want it faster, but that will be our choice. Just as
If information wants to be free. So be it. It will just take a little longer to get it.
Mr. Mark,
I\’ve always been a fan of watching you on the sidelines, both to see how exicted you get when the Mavs win and how pissed when the Spurs sink a three in the last minute… but c\’mon, this is a load of horsesh*t and you know it. This isn\’t about the weight of high-end internet applications as much as it is about gaining competitive advantage in the service provider marketplace. Bell has even attempted to argue that NN violates their right to free speech by allowing (by default on the current infrastructure) bandwidth to viewpoints Bell may disagree with! What a crock! It is a MISTAKE to put tiered service in the hand of those who have no concern for what type of information is proliferated… the internet will become TELEVISION! What\’s profitable remains. You make us believe that the goal of internet cultural progress, and the measure of its modernity, is the availability of bandwidth-heavy streaming data… as if being able to stream Mavs games in high-def is a greater cultural acheivement than a user-edited, user-maintained, and FREE encyclopedia. Just to make sure we\’re on the same page, Mark, it is NOT! No telecom is losing money providing internet service (imagine how much quicker Vonage would have gone under without net neutrality!), they\’ve just been gradually realizing that they\’re players in a game in which they didn\’t write ALL the rules. Stifling technological progress? Get creative. Build a new fibre-optic infrastructure and sell it to the highest bidders. Let the telecoms launch the \”new internet\” and see what it becomes. Seems like the people already have gotten creative, and spoken.
Comment by Josh the Sixers Fan -
How can you force a CLEC phone company to allow you to switch to another provider because the current provider is unable to provide internet or guaranteed phone service since I am the last phone on their cable get the worst service and they wont allow me to move. I need the big hammer technique to force my phone company Madison River aka Mebtel Telephone in Mebane, NC to allow me to switch to Bell South’s service which is available at my back property line and Bell South will at least be able to provide customer support and a guarantee.
HELP !!!!!
Comment by James Shivers -
Sorry. http://somaonline1.8.forumer.com
Comment by Fred -
Do you think there’s a market for high speed toll lanes on the freeways? Absolutely. Do you think it will ever happen? No way — too much political dynamite. So the folks who want to get quickly from one place to another, and can afford to pay a lot to do so, take to the air in jets and helicopters.
Comment by runescape money -
The rise in trafffic and the rise in cost due to tech advances are almost getting inversely proportional. A slowdown is most likely caused because the local network is slow- or because the data center is having network issues
Comment by wow powerleveling -
That is an interesting point.
Internet is still very free and open, nowdays there’re almost no problems with bandwith, there’re multiple free wireless access points and too many free services, but almost nothing is totally free, everyone wants to make money where its possible
– Mike
Comment by Mike -
Broadband over Power Lines is one of the most promising technology to deliver Symmetrical download/upload speed and is shown to support Hi-definition video streaming as presented by DS2 200Mbps BPL technology !!!
read more: http://broadbandoverpowerlines.blogspot.com/
Comment by TREXIA -
The larger question really is what are the market forces driving this? Over the last decade or so, the growth in bandwidth has largely kept up with and during short periods (e.g. end of the bubble), outstripped demand. In a situation like this, there is really no need for this to happen because the service degradations being discussed don’t happen enough to cause a mass outcry. Are we now at a place where applications are beginning to place more demand on bandwidth that can be supplied?
Let’s assume the answer is yes (although I’m not convinced of that myself). There are two ways to solve this problem technically. You can prioritize traffic or you can throw more bandwidth at it. To date, its been cheaper to throw more bandwidth at the problem. I’m not sure that that’s going to change anytime soon. However, if it becomes cheaper to provide traffic prioritization than it is to increase the bandwidth, prioritization will happen and those with money will get better service than those who don’t, regardless of what applications they are using.
http://fsboindex.wordpress.com/
Comment by FSBO Index -
This multi-tier thing could be good. If the differentiator of added cost/fees is volume, then voice traffic should be safe. Keep it free Skype, Gizmo, MS, Google, Yahoo…
Or will they somehow get to that sacred cow of the Bells business.
Comment by Jewelry Town Master -
I’d like to comment on two main points, 1) the implementation of multiple tiers in the Internet, and 2) the relevance of a tiered Internet to the Net Neutrality debate.
The debate around multiple tiers of service quality on the Internet is an interesting one.
Mark Cuban points out that the Internet is congested and needs an HOV lane (or more of a toll express lane from the sounds of it). But the thing is that QoS — a.k.a. Classes of Service, or tiers — does not in any way create bandwidth, it merely determines which packets get dropped when there’s not enough to go around. To use the freeway analogy, we are talking total gridlock between the hours of 3pm and 7pm. You may as well not even to get on the freeway because it’s completely unusable.
The reality of the situation is that any user on the network wants their application to work. If they pay any money at all for their connection (and perhaps even if its free) they want it to work. When I am in a hotel with free Internet Access, I expect it to at least be usable.
Global Crossing found this in it’s own network implementation. We implemented QoS in our network some time ago in support of our IP-VPN service, and reinforced it when we added commercial voice traffic to the network. The Internet services that run over the same pipes remain a best effort service, and reside in the lowest tier. We knew that our Internet customers would not be satisfied with degraded service, so the network is engineered provide high quality at all service tiers.
When it comes to Net Neutrality, the idea of a multi-tiered Internet is largely orthogonal to the issue. In the situation where no tiers are created Internet-wide, the local access carrier creates two tiers of service inside their network, one for their applications and one for everything coming in from outside of it. On the other side of the coin, Internet-wide service tiers are created so that other providers can send traffic in different classes to the access carrier, and the access carrier can still create one additional high-end class that supersedes all others.
The angle to be concerned about for net neutrality is where the access provider literally impairs the ability for their competition to provide their applications. Multiple tiers alone will not do this, because if there is no other application competing for performance in the bottlenecked areas (almost always the last mile) then the application in the lowest tier will perform acceptably.
Posted by: Dave Siegel at January 17, 2006 11:46 AM
I think that Mark Cuban has faced the future and seen the problems on the horizion. Just as VoIP was not feasable in the past for most people due to lack of bandwidth, he sees the next “level”, video applications being stuck in the mire of an internet that is too slow.
His idea of a multitiered internet is a good idea for a short term answer. It will get the bandwidth to those that want to pay for it quickly. The guys that are willing to pay $5 for a movie or $1 for a TV show will gladly fork out another $50 a month to be the only one on the block to be able to see a film at home the day it hits the theater.
BUT they will be the only one on the block. The rest of us, in my case the entire country, will have to wait until the internet becomes fast enough at a price we can afford.
Will Mark Cuban’s multitiered internet help us? NO. Will it help the independant video or movie producer? NO. IMHO it will hurt us.
Once the die is cast, there is no turning back. Having the ability to charge different rates for different service levels will stay with us forever, like having to pay more for calls to cell phones than landline phones (common outside the U.S.)
It’s a quick fix that will cause more problems in the future than any gain that will be realized.
As I say in my internet .sig, “The trouble with being a futurist is that when people get around to believing you, it’s too late”. Lets hope they believe you Jeff, not Mark Cuban.
Bobi Man
Comment by Mobile blog -
I think you have lost your mind. Lets take the example of the doctor checking on a patient at home. I happen to have set up and administered something along these lines. First, the nurse practitioner’s office paid a fee to an isp for the broadband connection. Then he paid a fee for the web server. This fee was paid to allow usage of the web server company’s broader pipe to allow multiple users to connect and work their way through to the point where they actually connected with the nurse practitioner. If that isn’t paying for his use of the internet throughway, I don’t know what is. Then the users of course had to pay for their own connections from their home computers into the webserver. And on top of all that money being tossed out on the table, the nurse practitioner’s office had to pay for the building of the website, the software server and client licenses, the maintenance agreements on both of those…. I don’t know where you get your information, but you are so way off as to be totally laughable. In order to upload or download, a fee must be paid. Currently google pays a fee just like everyone else for their ability to provide their service on the web. I’m guessing here, but I’d think they pay this in the form of maintaining servers and paying for their upload and download pipes. Please get a clue before spouting off your mouth and showing how stupid you really are.
Comment by Gertie -
Interesting. But its hard to trust you. Aren’t you heavily invested in HDTV? Seems like “tiered” service (caste system) for “emergency and medical” applications (HDTV?) would be great for you. Esp. if you could stop “little Johnny from downloading ABC’s primetime lineup.
Comment by John -
I think that Mark Cuban has faced the future and seen the problems on the horizion. Just as VoIP was not feasable in the past for most people due to lack of bandwidth, he sees the next “level”, video applications being stuck in the mire of an internet that is too slow.
His idea of a multitiered internet is a good idea for a short term answer. It will get the bandwidth to those that want to pay for it quickly. The guys that are willing to pay $5 for a movie or $1 for a TV show will gladly fork out another $50 a month to be the only one on the block to be able to see a film at home the day it hits the theater.
BUT they will be the only one on the block. The rest of us, in my case the entire country, will have to wait until the internet becomes fast enough at a price we can afford.
Will Mark Cuban’s multitiered internet help us? NO. Will it help the independant video or movie producer? NO. IMHO it will hurt us.
Once the die is cast, there is no turning back. Having the ability to charge different rates for different service levels will stay with us forever, like having to pay more for calls to cell phones than landline phones (common outside the U.S.)
It’s a quick fix that will cause more problems in the future than any gain that will be realized.
As I say in my internet .sig, “The trouble with being a futurist is that when people get around to believing you, it’s too late”. Lets hope they believe you Jeff, not Mark Cuban.
Drug
http://drugstown.com/blog.html
Comment by Drug Stown -
http://pub17.bravenet.com/guestbook/1460277761
http://www.zyworld.com/ringtone/free-ringtone.html
Comment by rene -
I can’t believe that there are those of you out there that think this can actually work to anyone’s advantage other than the telcos. To turn Mark’s own words back on him, the telcos WILL use this opportunity to screw over the VOIP providers and everybody else.
A tiered-service network will drive us right back into the 1980s with the modern equivalent of “calling plans”, on-peak/off-peak hours and worse.
I’m not ignorant of the fact that individual ISPs and transports are already doing things like MPLS, QOS and other forms of prioritization, but this is a far cry from what BellSouth and the rest are itching to do.
Anybody who thinks otherwise is either kidding themselves or a shill for Big Telecom.
regards,Spetra
http://www.spetra.ru/
Comment by Spetra -
I love it when people assert that Google, Yahoo, consumers, are using the “pipes for free”.
Newsflash:
Google pays for all of that bandwidth from its respective ISPs. The ISPs pay the telco for the pipes. That’s how this works.
The consumers pay for the bandwidth on their end.
Who, exactly, is getting free bandwidth? There is no such thing as a free lunch.
Creating prioritized traffic lanes on the net will be a grave mistake it would take years to recover from.
The internet is a success as it is. Let’s not go messing it up with, what I’m afraid is, a really bad idea on the part of the Telcos and Mr. Cuban here.
Mark, you usually have some great ideas, but you are way off on this one.
Comment by Dylan Carlson -
My two cents are that a multi-tiered system is like communism, a great idea in theory , but easy to corrupt once the powers that be start to “regulate” it. Just as we need car pool lanes, I also forsee that we will need “Voip express lanes”.
If these are regulated by the government, then they WILL be used against the VoIP providers by the telcos as the telcos have bigger lobbies.
Comment by PortR -
It was truly glorious and a testament to the power of community the lack of regulation on the internet, but it would seem, like Camelot, those glory days are coming to an end. For future regulation I liked the analogy of the ambulance getting guaranteed passage. I always like the idea of rationing beyond a minimum (why am I paying higher gas prices when I am a minimal user of an auto? If everyone drove as often as I do there would be no gas crisis. I feel like I’m being charged for other’s greed and insensitivity- but I digress).
In getting access to the internet through many networks (ISPs, hotels, etc.) there are companies that provide “bandwidth control” or “equalizing” appliances and have “fairness algorithms.”
That would seem to be the future. At least then it would not be based solely on money.
Comment by Leland -
Hello
I think that in the aggregate, same applications by different end users must have equal priority at the CO unless explictly chosen otherwise (and thus priced the same), but bigger users of bandwidth must pay more – which is the case even today – but not to the extend that accounted for priority differentiation at all endpoints. There’s a lot of bandwidth in the carriers’ pipes, but not at the last mile for all the subscribers of a CO to do VOD at the same time. Even without common agreements by the ISPs, shaping the traffic at this last mile will achieve a great deal. The ISPs will be compared by how well an user’s highest priority applications perform.
A sticky point is how should applications be prioritized: by providers or by end users? The second option is available to enterprises (with MPLS VPN, etc.) but that doesn’t help the masses; the first option can be a business opportunity for ISPs and this may be what the telcos should be doing before adding more bandwidth. The problem is they probably have little motivation to do this because it costs them money too without reaping the reward — unless the problem is understood and they are able to charge accordingly.
On the other side of the picture, it would be unreasonable to charge content providers with higher rate since they already have to pay for the bigger pipes.
Comment by buy tadalafil -
Hello dear Mark,
Sorry but I’ve got to disagree with you on this one, from both a business and a technical perspective. When you get into SLA territory, you find that price goes up way faster than the payback for each “nine” you get (99% versus), and uptime is always a legally fenced-in concept which isn’t going to cut it if someone’s life is riding upon it
Comment by Sergej Dir -
Mark,
I agree with most of what you said. I’m not a Telecom guy but I have been using a PC to access BBS’ and the Internet since the days of the first 300 baud modems. We’ve come a long way but really our demand for bandwidth is just beginning! Others have mentioned that the capacity is still in the marketplace, we don’t need tiered services, etc. I disagree with many of the comments I’ve read…The capacity IS in the marketplace but unforuntately this marketplace doesn’t reach most of the homes in the US. I live in a rural area outside of the city and I’m right on the fringe of DSL. There is no Cable TV and no acceptable wireless solution yet so I’m stuck with Verizon’s DSL on very old analog lines which has daily outages. They have the ability to give me fatter pipes via fiber-to-the-home but they won’t because they’re too busy building out these services to communities with a much higher concentration of homes and businesses.
We would all benefit from tierd services by our service providers. I would be more than willing to pay more for my service, as would many of you, if it was more reliable and had greater bandwidth. Others who only use the internet for email or the occassional web browsing would like to pay less for service because they don’t need the extra bandwidth. That would allow more bandwidth for the rest of us.
Every household has different requirements but as our dependency on broadband grows, so will our appetites for the bandwidth that’s required and who will step up to provide it.
If you look at the trend toward this convergence thing we’ve been hearing so much about for so long, our Voice, Video & Data will all be IP which will require much fatter pipes than most of us have now. Even the popular software we use will eventually be web based so there will be no limitations on where or how you access your data.
Comment by David Ward -
Mark,
The contention that unmet demand for bandwdith continues to outstrip the supply of bandwidth is a key premise for the need of P2P systems.
Remember, P2P systems are able to localize data transfers and will keep 90% of the traffic off of the highways (i.e. if you need a file and somebody on the same LAN already has it, P2P makes sure you don’t need to go to the information superhighway to get it).
At the end of the day, P2P can take a vast majority of the content/data traveling over networks and route it orders of magnitude more efficiently. This would leave the pipes open for more “mission critical” applications that may not have P2P replication characteristics.
Just a thought,
Publius
Comment by Publius -
Mark:
This is a Malthusian “failure of the commons” situation, and your analogy to highways is correct. Before a highway is built, people were still getting from point A to point B on the surface streets. Then the freeway gets built with projected capacity of X cars/hr, which is typically a multiple of what the designers believe the current demand to be. Then two things happen: 1) most if not all of the people who were going from A to B jump on the freeway; 2) the capacity for moving people from A to B induces people and businesses to cluster around A and B (the transportation system changes where people live and work). As time goes by, the demand for capacity increases to the point where the freeway begins to slow, and as a result some number of travelers dive off on the surface streets again or find A-Z-B routes which are faster than A-B
When the A-B commute gets painful enough, more lanes get added to the freeway, and quickly the people who had found alternative routes come back to the direct A-B route. The demand starts increasing again, the freeway slows down, people dive off on the surface streets and you start another cycle.
The same thing has happened with the Internet. At the first level is the supply/demand economic which says that as more suppliers enter the market, total supply will get ahead of demand, causing prices to go down. But increased capacity doesn’t just improve performance for current applications, it creates space for new applications, which in turn create more demand.
Your thesis is that the market would benefit from having a premium offering where one can get better service by paying a premium. I couldn’t agree more. Drive down the east coast of Florida sometime. I-95 and the Florida Turnpike run parallel to each other. The first is free and the second has a toll. Why does anybody pick the toll road? Because it performs better!
The economics of running a network is that you have enormous fixed costs and almost zero variable costs. So if you build one, you’re going to be desperate to generate revenue until you cover your: a) cash flow; and, b) fixed costs (including DEPR). You will absolutely sell on price and will be willing to “overload” the network as much as the customers will get you get away with.
If you can carve off a little of your network capacity and sell it at a premium in exchange for better performance, there’s no question there will be demand. If the market will allow it, you’ll also try to convince standard rate customers to use the network at a different time than the premium customers (in the same way airlines, hotels, etc charge you less if you are willing to travel on the weekend).
Someone said there is excess network capacity out there. That must be true because prices are still cheap. But measuring capacity is a tricky thing: the only point that counts is that instant when demand is at its peak. At every other time, there is idle capacity. It’s the reason cell phone carriers still have the notion of peak and off-peak minutes. They want to get as many incremental minutes as possible used during times when the network is idle, because there is no incremental cost. However, additional usage during the peak is VERY expensive.
Paul
VP/Network Engineering (ret)
CompuServe
Comment by Paul Lambert -
man, you seem to be a genius. Are you like the American Confucius or something. I hope your innovative ideas are executed.
Comment by Champ -
Mark,
I read all of your blog entries and I have to say that I agree with you the majority of the time. In this particular case, I have to say that I disagree with you. I have only been in the telecomm industry for about 5-6 years, but I believe that whether the cable or telco’s are using ATM or GbE to deliver service to their customers, the QoS for each of these types of services is clearly defined in a way so that “critical” applications get the priority that the customers are paying for.
If the coup de grace is to provide HD from anywhere on the Internet, then I think that it will eventually come along sooner than you think. As we all know, edge servers provide most of the SD and HD video services to consumers today simply because the access network is controlled by the telco’s and they can buildout whatever they want; xDSL, FTTx, etc. However, I believe that as the telco’s increase the speed to the consumer over time, the encoding technology will become available that can reduce that HD content to something manageable that the Internet can handle as it increases it’s backbone over time.
Comment by Rob Kruciak -
Mark, you’re simply wrong about the actual traffic patterns. In fact, the actual increase in traffic has been modest and manageable, with the cost of carrying it dropping rapidly. While video is creating increased bandwidth demand, the takeup has been slow enough the problems are more imagined than real. While the scenarios you mention are certainly possible, that just isn’t what we are seeing in actual networks. I write the DSL industry news, and have been reporting the actual traffic patterns and costs for several years. It’s instead telco propaganda to justify erecting a toll booth to collect from your video programs and their customers. We’re currently looking at less than $1 per month of the $15-50 consumers pay going to cover the internal and external costs of traffic, down significantly over the last two years for most major networks.
Bandwidth isn’t free, but it’s pretty cheap. Cost of carrying traffic is going down with Moore’s Law at 20-40% per year. Actual traffic growth, despite the incease in video, has been less than that on most networks, so has not raised costs. DSLAMs designed since 2003 have been non-blocking to a very high concurrency, while dropping in price. Switches from the DSLAM to the Internet connection point can be inexpensively upgraded to carry likely traffic loads. The fiber is in place almost everywhere, so the main cost of increased traffic is switches and other electronics, dropping every six months.
Real world traffic management can handle the likely peaks quite well for standard definition television without needing any of the QOS stuff. HD will probably work ok as well on Verizon’s 15 meg FIOS, for example, which is engineered so you should be able to send from HD Net servers a 9.3 megabit stream unless Verizon chooses to limit you. HD data rates are tricky today, but well on the way to being solved on the technical level.
I’ve emailed you directly as well. Someone at the telcos is feeding you flase data.
Dave Burstein
Comment by Dave Burstein -
I like your stories, but I can’t stand all the typos and punctuation and grammatical errors in your posts. Maybe you should run your articles through Word before posting.
Comment by arjun -
of course you want it, Mark. You can afford it. but the playing field wouldn’t be even anymore. No more little guy makes it big, because they won’t be able to afford to compete anymore.
Comment by brad -
The idea of multiple forms of connecting is appealing. To improve reliability, people may have both cable and dsl in the future. Competition between the two may improve service, including the creativity and effectiveness of solutions for bottlenecks. Real incomes may need to increase so people can afford more than one type of internet connection.
The highway analogy is interesting. It might be interesting to keep discussing it.
Some highway jams are due to design of on-and-off ramps and intersections (junctions) in addition to lack of enough lanes during busy times. Is a potential internet bottleneck due to switching stations or bandwidth or both? I am very not sure.
Also, society can do things to mitigate traffic other than prioritization: utilize work schedules at alternative times, use technology to be productive while driving, build complementary public transporation (trains), etc. There may be things the internet mgt people can do to balance-the-load. I am not sure.
A prioritization mechanism for internet traffice, if centralized, might be dangerously subject to abuse from dictatorships.
Comment by nate -
The current risk in this country is too much preference given to old people, with potential underinvestment in young people that results in lack of fair opportunities for young people.
In general, one might argue that old people get tons in this country at the expense of young people. The political system is skewed to favor older people because older people may vote more, have more money, tend to move less, have more to lose and thus fight harder, and have relationships to get what they want from govt. even if not optimal for society (influence costs).
Look at recent medicare expansion. Look at trends in labor-force participation rates by age. Look at history of age discrimination litigation vs. other forms of discrimination. Look at recent federal govt. budget cuts (possibly programs for education and lower income people curtailed, while energy and other bills may have giveaways and welfare for energy companies). Then tell me old people are getting shafted relative to young people – internet or otherwise.
If the govt wants to help society, the fed govt would make an investment in young people equivalent to the investment in the expansion of Medicare drug coverage. Ideally the fed govt reaches kids at a very young age. If the investment enables a prank or two along the way, so be it. There can be truth in jest.
So I have to think Grandmas will be at least adequately considered in the internet realm. Relatively poor people, and people with no “relationships”, will be ignored.
In reality, if Little Johnny has relatively tons of money (inherited or earned) and connections, he will get his prank over an ordinary grandma’s health application. It is this way today in non-internet ways. Why would it be any different over the internet?
And even if *rich* johnny’s get his check-up before grandma gets her internet, grandma may still be better off in the future than 50 years ago.
———————————————
“I would rather have little Johnnys grandma getting priority for her video checkup with the doctor at the hospital over little Johnny getting his bandwidth to upload the video of the prank he pulled on his buddy.. I would rather make sure that information from life support or other important monitoring equipment, medical or otherwise is finding its way without interruption, and without the end user having to pay for an off the net solution. These are the applications that make the net great. These are the applicatins that offer equal opportunity to those who are disadvantaged.”
Comment by anon -
I just commented on this one on my blog, and I stand by my comments:
I do expect that the flat-fee residential broadband connection will become a thing of the past. Most likely, it’ll evolve towards some type of pay-for-use scheme, not dissimilar to the pricing scheme for my cellphone, which charges me a flat fee for all “in network” connections, additional fees for the use connecting to other providers, and some discounts if that use is outside of peak hours. I don’t get worse service when I go out of network, but I do have to pay by the minute or megabyte. And the rates might be cheaper if I time my downloads to off-peak hours. (Expect a version of iTunes that allows you to program your desired download schedule to match your broadband provider’s “off peak” hours.)
Those who can deal with being lower in the queue will be happy most of the time and unhappy at others.
-btc
Comment by Below The Crowd -
While I don’t have the information I would need to be involved in the bandwith available increase rate vs bandwith consumed increase rate debate, I think there’s another big pitfall to the tiered set up.
Bad incentives.
I don’t disagree that there is a fair amount of intelligence behind prioritizing certain traffic – we have a system where ambulances put on their lights and other cars get out of the way for a reason. On the other hand, if supply and demand starts becoming a real issue in the bandwith market, its not going to be good for anyone of us (unless we are telco executives) to create a system that rewards an avoidance to infranstructure upgrades.
I don’t want my ISP to avoid putting down more fiber because it knows that the price it can charge and the number of people it can charge to ‘prioritize’ their traffic will skyrocket if it doesn’t.
And I hope my comment doesn’t garner a lot of ‘oh but the free market will handle that, no worries’ sort of comments. Its telcos guys.
Comment by Sean Kehoe -
Sorry Mark, I’ve got to disagree with you on this one, from both a business and a technical perspective. When you get into SLA territory, you find that price goes up way faster than the payback for each “nine” you get (99% versus 99.9% versus 99.999%), and “uptime” is always a legally fenced-in concept which isn’t going to cut it if someone’s life is riding upon it.
How is BellSouth or Comcast going to guarantee me more reliability for my cable connection for an extra 20 bucks? How are they going to guarantee that Joe’s landscaper down the street doesn’t cut my cable line (this happened to my girlfriend’s family a couple months ago, thus causing their TV, Internet and Phone to go dark)? They’re not. The way you’re going to get extra reliability is the same way you’ve always managed to get extra reliability – shell out more cash for more connections to more companies – if I have a wireless connection provided by Sprint and a cable connection provided by Comcast, and a dialup connection provided by AOL, then I’ll be much more redundant, have more bandwidth, AND I can do this today.
There’s no need for too much tiering in service, other than what’s already available today with the DSL-light options which cap your downloads to 40k and the Speakeasy 6 down/1 up monster ADSL connections.
Plus, like some others have pointed out, our needs for bandwidth are not growing exponentially, just like our needs (for most people) for hard disk space aren’t keeping up with what’s available.
Maybe you were talking more about life as a content provider? In that case, I’m reasonable sure there’s already provisions to be able to site a mirror on BellSouth or Comcast’s network, which is why Windows updates go way faster than they’re supposed to. Plus, there are plenty of mirroring solutions provided by companies such as Akamai.
Comment by John Peebles -
Mark,
While I’m generally pretty fond of what you have to say, I strongly disagree with your post. One of the key factors that made the Internet sucuessful was the lack of any central authority or tiers of service. With QoS, why should RoadRunner customers be able to download Time Warner movie trailers faster than Sony Pictures trailers? Similarly, what if Warner refuses to cut HDnet a tiered service deal because you compete in one of their markets? Net neutrality allows users to decide what’s important to them, not large ISPs and corporations.
Your analogy comparing bandwidth to highway traffic seems somewhat flawed. While we can’t widen the 405 without bulldozing a rather large number of houses with all the associated political issues that implies, we can add as much bandwidth as the market needs/wants. Google is reported to have miles of dark fiber waiting for a use. If anything, the bandwidth problem lies in the last mile delivery from central office to end user, a problem tiered service will do nothing to solve.
Comment by Zach Lipton -
I don’t agree with the majority of this post (TBH it reads like an advertisement for Redswoosh)
The main bone of contention is “our ability to consume bandwidth is growing far, far faster than the speed at which it is being added”
That’s not strictly true, it’s like saying “well we’ve got 80GB hard drives now, why aren’t we transferring those across the internet”.
Cost Vs performance. Current fiber technology is still increasing at a steady pace, the OC level goes up uniformly. The issue is people want to jump from 1-2Mbps to 20Mbps almost instantly. Telcos don’t find that increase cost effective in rural communities (and urban areas don’t want telcos digging up the road every 2 years). Waiting a little while for telcos to catch up isn’t a big deal. There’s no need for them to spend billions upgrading the network for HDTV when 2 years later they’ll need to do it again for a new technology (or simply more users).
As far as introducing a priority system, NO!!! Any idea what that’d mean? it was discussed by the FCC and is widely seen as a big red flag. What we’d see is one big walled garden, with ‘priority partners’ being given …well priority to their users over other (likely smaller) internet services.
There is already a basic QoS level in place for peering. Why ‘fix’ something that isn’t broken? Perhaps I’m lucky but i’ve never experienced a connection speed drastically slower than advertised, my ISP’s job is to supply me with the bandwidth I’ve paid for and they’re doing a good job of it.
In short, there’s no need for a new QoS or your complaints about bandwidth, the market will right itself through competition. If your ISP isn’t giving you the bandwidth you desire, you’re going to have to pay more or change provider.
Comment by Adam -
Mark,
Let me correct one of the assumptions you make about us consuming more bandwidth than is being added….I worked in the telecom industry for the past ten years until ’93….I experienced the telecom meltdown, which partially happened because of overcapacity….Too many companies building out to much bandwidth and they eventually went out of business…This capacity it still in the marketplace, and we are starting to see demand catch up to it…But we still have a long way to go.
Comment by Jason -
The analogy of internet traffic and traffic in LA is excellent. I grew up in LA and saw the 101 go from 3 to 4 to 5 lanes – it never made a difference – people just drove more.
At the end of the day, whether you like it or not, the bells made the investment to put new fat pipes into the home. They did it based on the FCC telling them that if they did, they would own this new digital right of way free and clear. Now that people are getting fast internet connections is it such a surprise that they want it for nothing?
I ranted on this a few weeks back. Beware of the government retroactively withdrawing it’s promises.
http://www.nyquistcapital.com/2005/12/29/more-government-better-broadband/
Might also want to check out a piece on what I like to call the “Industrial accident” – and why unintended consequences are the lifeblood of technology.
http://www.nyquistcapital.com/2005/12/21/more-bandwidth-industrial-accident/
This open access nonsense could put us back where we were after the 96 Telecom act- with no one willing to put up risk capital.
Comment by Andrew Schmitt -
Remember Sprint ION? I believe it ultimately never made it to market because of the exact QOS issues you mention.
Comment by Nick -
Hi Mark,
Today you can sign up for prioritized traffic (at enterprise level) with almost any carrier, everywhere.
Most carriers will just create an MPLS VPN for your company, giving you precencence over other people’s traffic (only inside their backbone).
The problem with traffic priority is that priorities don’t cross carrier network boundaries.
Comment by Iñigo -
There is going to be a need for packet police on the high priority line. Then we are going to have some privacy issues while looking at those packets. It will also probably need to be a felony offense for interference by hackers.
I don’t see a problem with paying a price for more bandwidth at a faster speed.
That’s like buying the Fast Lane pass at Magic Mountain!! Hmmmm stand in line for 2 hours or pay a small price and wait 3 minutes. Time is money and I like to leverage it!
I see more of a problem keeping everybody in the lanes they need to be in.
Gotta go… my son is waiting for me to play “Rise of Nations”.
Darci
Comment by Wild Flower -
God what spin!
Once again, skewed double-talk straight from the darkside of capitalism: “We do this [tiered service, priortized content] all for YOUR benefit.”
We already have tiered service in our internet connections. Do we need to tier it again in the high-speed arena to further segregate people of different financial worth?
Throughout history the prospect of bulking up one’s own “corporate share worth” has always been able to put “blinders” on the struggle for justice and equality for all. The darkside of capitalism has been and will continue to be the root source of much, much grief.
Net neutrality is essential for the long term health of the whole. Stop make-believing that it’s not.
bill (in washington, dc)
Comment by bill bouslaugh -
I am gravely concerned about the promotion of a teired internet system. I believed such a system risks falling prey to an intended or unintended filtering of information, and potential selective access to information by only those who can afford it, not to mention a gold mine for a selected few corporate interests. If a free society and world freedom is to prosper, then it is obvious that the free, unrestricted flow of information is imperative. The internet is one of the few information options remaining that, so far, has escaped the influence of money, corporate influence, etc. I also resent being campared to a drone, and I do appreciate the assistance of Common Cause in facilitating my expressing of my opinions. I continue to believe that I am fully capable of chosing which Common Cause issues I wish to support, and have and will continue to exercise such descretion.
Comment by Jerry Bloss -
Mark,
You mention in your post the need for us to be able to service “mission critical applications,” but you don’t specify who determines what is and is not mission critical.
In your home diagnostics example, if SBC, Bellsouth or whatever other ISP is being used by “little Johnnys grandma” is the one determining the reliability of the connection, given the comments noted in previous posts (particularly #23) it seems clear that not only will her ISP charge her more money, they’ll charge the content provider (in this case, the hospital or health services provider responsible for providing the diagnostics) more money. That is a cost that then transfers to her health insurance, or a government program such as Medicare. This translates to higher costs for all of us, either in increased insurance premiums or in the tax burden of supporting those programs. Either way you approach it, it’s not really *just* the people receiving the benefit of prioritized content delivery…we’d all be subsidizing it.
Now, depending on your politics, it may or may not be in our best interest in that particular example to subsidize prioritized content delivery. However, there will be other examples of costs being shifted to those who aren’t benefiting from such a system, and that’s prior to considering how ISPs will handle traffic between their networks: it’s one thing for Comcast to charge Google Video for prioritized delivery of their content, but it’s quite another for them to charge AOL for delivery of content from their network, and for AOL to likewise charge Comcast for delivery of their content to AOL users. Do you really think they’ll net those cross-network charges when each ISP can pass them to their respective customer bases and boost their profits?
Comment by Justin -
The gap between bandwidth we consume and the added speed networks upgrade to be widen at scary rate, if those who set industry standards don’t manage that difference, I believe everybody would suffer freezing internet soon regardless of what type of connection is used.
Comment by Mag -
So they become a new breed of traffic switching center that routes packets based on IP$… the new protocol powering their vision of the Internet.
Comment by whales -
Hello,
I managed to pay my utility bills again this month. I am estatic about this. Next month is a different story.
I live in Texas the home of Gas and Oil. We pay more for Gasoline than many people outside of the “Gas and Oil” state. Our utilities are as high or higher than many depressed areas, yet we have a lower average income than the same areas.
I think it is kinda funny that in such of economy, where we are constantly hammered about dirty skys, we are rapidly building dirty energy businesses, that generate more dirty skys. The state government seems to be in favor of this issue.
Imagine that.
Comment by Bob Brown -
Writing in regard to #10, anonymous.
At least I have the guts to sign my name.
Hey young overcharged guy, about being overcharged to subsidise and support older persons such as myself. Do your homework.
Medicare is far from a free ride for retired persons on a fixed income.
This year I received $16.oo a month increase in social security.
The cost of medicare went from $48 to $82 a month.
The cost of supplemental insurance went from
$105 to $185 a month, the reason? medicare cut far back on the benefit amounts paid, so the price went up to compensate.
Drug insurance is $40 a month, with a $5+ for co-pay depending upon the drug needed. which can be far more than $5 if a specialized drug is prescribed.
I am now 76, having worked for 50 years before I retired.
I only wish that I could be working yet as you are.
I can only feel sorry for you having to help support me, as I’ve helped support older persons myself for many years.
But there is a difference in us, I didn’t whine and complain about it. I can remember when social security and medicare did not exist.
What used to happen? it was done in my family.
When my grandparents were too old to care for themselves they came to live with us, and they were welcome. Maybe money wasn’t too plentiful but we all got by.
God, I hope you and people like you never have to depend upon someone to help care for you, from the tone of your letter you would surely make someone’s life miserable.
Comment by glenn koser -
Personally, I think the idea of a tiered Internet is *very* wrong. It fundamentally changes how the Internet was designed. This is nothing more than a power grab by Telco and media companies to squeeze out those who cannot afford the high bandwidth charges. Much like purchasing space in the radio spectrum from the FCC – it costs billions and only those with all the money can afford to do so. Only those who can afford to do so are allowed to “speak”. If a tiered Internet is put into place, innovation, creativity and new exciting online products will dwindle.
The development of the Internet and it’s resulting growth was based on the core value that once a packet hits the network backbone, it’ll get to where it needs to go without any prioritization or special treatment.
Correlate it to the idea that “all packets are created equal” and then once tiering is created it’s “packets from those who are wealthy are given more consideration, the rest of the packets can mire themselves in the slow lane”. Kind of paints an elitist picture in a sense?
Of course Mark Cuban is interested in segregating the Interenet. 1. He’s very wealthy and could easily afford to pay to have his content give the “star” treatment. 2. He’s a content provider – he wants to use the Internet to push his media out to users. I find it interesting that someone who created so much money off the Internet (the open – none segregated Internet) now wants to effectively lock out any competition.
To follow-up to my statement – I find it interesting how riled up American’s got when the UN made a power grab to “govern” the Internet. No way we cried! We don’t want anyone meddling with it. Well – that’s exactly what the Telco companies are trying to do – govern who can get on the Internet – by placing tollbooths onto the fast lanes and charging an arm and a leg to use it.
One other point – just to point out how fucked Telco’s are in their thinking. Look at municipal broadband, and the abject terror that put into Telco’s. They are pushing for *law’s* to disallow American citizens to roll up their sleeves and create their own networks. Many cities are sick and tired of subpar broadband offers, or if you are in the boonies – no broadband at all. So, with the American spirit in mind, they say “let’s build our own”, much like rural electric and phone companies were done so long ago (and extremely effectively too I might add). The Telco companies, respond to this by yelling “unfair” and run to congress for special treatment by putting laws into effect to stop this. How utterly un-American. It really makes me sick. Now you have these same Telcos, (who’d drop millions in smear advertising and overpriced lobbyists, rather than improving their networks) trying to tear down the highly successful framework that made the Internet what it is today, just for more money and control.
I am all for the government stepping back in and slapping these imbeciles back into line.
Comment by Ribald_Jester -
Great plan for the rich, but what about those of us who budget ourselve’s just to be able to afford an internet service, let alone have to pay for the right to receive email we have already paid for, or have to pay for the right to search for information that is included with our service package, which is already so high it drains my pockets just to keep it going.
Also, Going on line to read my email, surfing the web to find needed information is a joy to me, and the only means of exscape I have to relax and froget about the hard day at work, not to mention other issues, calming me me down to where I can better deal with them later on in the day or night, and now all this might be taken away because some very stupid rich foggies want to get richer by robing me, and others like me of more of our hard earned money. Heck, I already pay for AOL $24,99 pluss tax, and for my DSL $ 80.00 including tax,thats over a $ 100.00 a month pluss, now I have to worry about $ 1,000.00 for rent, electric, gas, phone, water, garbage, car payments, my families needs, and now they want to make my internet connection none affordable, not to mention that my computer which I have an investment of about 6,000.00 includeing equipment that I might be forced to push in a corner to collect dust, because I can’t afford to enjoy the internet any longer.
There are other people out there like me, and they will be forced to do the same thing, and everything they have invested in their computers will all be in vain if they can’t afford to use them. Also, do you realize that if everyone like me are forced to shut down our computers, how much money would be lost to not only ISP’s, but to computer manufactures as well, because if people like be can’t afford to use the internet, and there is a whole lot of people out there like me, computer sales would drop.
And this thing about controling spam, and priority mail is a bunch of crap, I have software to help control spam that works just fine, and if I pay for an email service, then I want to receive all my email, not what my ISP thinks I should get or recieve, and I don’t want them sticking their nose in my business. Also, why should we trust AT&T, SBC, Verizon, Comcast, AOL, when they have been in trouble and accused of crimes against their customes many times in the past, and now they want to play goody two shoes, and want us to trust that they wouldn’t try to abuse us in letting them add these new charges, I don’t think so….? besides, they are the biggest spamers out there, so is this their way of making their spaming priority over other spamers. or is it just a cover up to team up with the government to spy on us, keeping track of everything we do on the web, and who we talk to. I say that everybody who values their internet service, and who wants to protect their information from getting into the wrong hands should stand up and fight this, and stop this from happening before it’s to late.
George
Comment by George Brown -
Mark, in light of your current activities I understand your interest in quality of service guarantees for high definition content but I refer you to these links:
http://blogs.siliconvalley.com/gmsv/2006/02/the_united_stat.html
and
http://www.newnetworks.com/broadbandscandals.htm
Seems quality of service wouldn’t be an issue had the phone companies lived up to their part of the bargain and delivered on what they promised for the $200 Billion dollar tax break they got.
Instead of using the $200 Billion in tax breaks and incentives to deliver fiber to 86 million homes in the US by 2006 like they promised we got DSL instead. No wait, now we’re getting 2nd generation DSL… maybe that’s the answer…. NOT!
Instead of asking for the Baby Bells to deliver tiered service, how about asking them to give the $200 Billion back…. it’s probably not an issue for you but I could use the $2000…
Bob Jakuc
Comment by Bob Jakuc -
Mark, in light of your current activities I understand your interest in quality of service guarantees for high definition content but I refer you to these links:
http://blogs.siliconvalley.com/gmsv/2006/02/the_united_stat.html
and
http://www.newnetworks.com/broadbandscandals.htm
Seems quality of service wouldn’t be an issue had the phone companies lived up to their part of the bargain and delivered on what they promised for the $200 Billion dollar tax break they got.
Instead of using the $200 Billion in tax breaks and incentives to deliver fiber to 86 million homes in the US by 2006 like they promised we got DSL instead. No wait, now we’re getting 2nd generation DSL… maybe that’s the answer…. NOT!
Instead of asking for the Baby Bells to deliver tiered service, how about asking them to give the $200 Billion back…. it’s probably not an issue for you but I could use the $2000…
Bob Jakuc
Comment by Bob Jakuc -
No offense but this idea is ridiculous. A tiered service would just be an invitation for the big players to swindle more money from companies like Google AND every day people.
You can easily purchase more servers if you’re finding that the amount of traffic is overloading your current equipment. There’s no need for a tiered service.
Comment by Mr. Holla -
between the telcos aggresively fighting municipal wifi and the proposal of tiered networking it appears the new american way in communication is to take the power from the people and to activiely attempt to curtail any furture potential inovation from small startups. The power of the internet has been the perceived equality. Mark if this tiered system existed when you first got in the game do you think you would have gotten to where you are now?
Comment by bastion -
I think there’s a great deal of confusion as to the nature of the ISP marketplace. This is not a perfectly free market where companies can compete equally with each other. On the consumer end, 99% of Americans have at the most two sets of wires that can deliver broadband – their phone lines (xDSL) and their cable service. You may or may not have various providers that you can connect to through those links (such as AOL, Earthlink, Speakeasy, etc.), but their offerings usually don’t differ significantly from the offerings of the respective cable or telco monopoly.
On the hosting side there’s a great deal more freedom – you can locate your servers where it makes the most sense – a CoLo, your own datacenter, a managed distribution network like Akamai, or some combination of these.
However, unless you use a distribution network or a particularly well-connected data center, you have little to no control over which ISPs your packets travel before they reach your consumers. Nobody allows “source-routing” across the Internet (heck, most people don’t allow it in private networks) – it would be easy to abuse and impossible to secure.
Economics aside, let’s look at the technical logistics of the matter. They suck.
Creating a “HOV” lane on the Internet requires either secure source routing that would have to be negotiated for every ISP (that you care about), or a micropayment system that is billed through your primary ISP. Either method would require some very good security to prevent abuse – IP networks are not presently designed to do things like this on the scale of the global Internet. Every core router and switch on the Internet would need to be replaced – this is tens or hundreds of billions of dollars of CapEx – because they would require several orders of magnitude more memory and processing power to handle the authentication and billing (even if you get clever and manage these on a per-source-host or per-connection rather than a per-packet basis – and these shortcuts would increase the opportunities for abuse). Much of the existing gear is running tight on memory and CPU power just delivering the packets without having to weigh and consider and authenticate and bill each one.
Even if you found the money, you’d have to develop standard RFCs (Internet standards) to manage these arrangements, which would take a minimum of 2-3 years. You’d have have routing and switching vendors develop products and sort out interoperability problems, which would take an additional 2-3 years minimum (even considering that they would start development during the RFC process). In 5-6 years, we’ll probably have added enough bandwidth through the natural evolution of things to deal with this issue in a brute-force manner.
Historically, simple network standards have nearly always trumped more “elegant-but-complex” or “efficient-but-complex” standards in the marketplace. Ethernet is not the most efficient or elegant standard ever conceived, but it’s slaughtered every other technology that it’s faced because it’s simple, easy, and it handles imperfect conditions rather well. It Just Works, and it Keeps It Simple, Stupid.
QoS is similar. It’s a cool idea, but it gets really ugly and complex to implement when there’s more than one network involved – and the ugliness and complexity tends to be even more political than it is technical. And the potential for err… inevitability of abuse (remember, this is not a free market!!!) has already been discussed extensively.
So sorry, Mark, it just ain’t gonna happen. The time, the money, the politics, and history are all against it. This isn’t a matter of being afraid to try something new, this is just planning, experience, and some simple forecasting and math speaking. We’ll beat this problem with simple brute force before we do with an efficient and elegant and complex solution. And that’s not a bad thing.
Comment by Erik Carlseen -
Mark,
You are strongly convinced this is a good idea, so I’ll presume that the ‘toll road’ thing happens.
Given that, how do you propose to force the Baby Bells to do the right thing? This is a hard question: but you’ve had a few good ideas here and there, so you might have one for this. I’m stumped about it.
Comment by Matthew Hornyak -
Mark,
Welcome to WALLED GARDEN 2.0! What you’re advocating is the biggest threat to the Internet today.
It’s quite simple really… the oligopoly ISPs want to run the Internet similar to the telephone network. They’ll charge customers extra for fast service (long distance) and they’ll charge companies that want to provide these services faster.
For YEARS these monolithic companies have been raking in billions of dollars on voice service. The first threat was the mobile phone when they offered free long distance calls. The second and final threat is the Internet. On the web voice is nothing more than an application. Nothing more than an application? Applications are free on the Internet… GOODBYE PHONE COMPANIES. In a few years the only phone companies around will be VoIP providers and the only reason they’ll exist is to terminate calls to stragglers still on the PSTN and to provide e911 service.
So what’s a phone company to do? Die quietly? They become ISPs… but being in the ISP business isn’t nearly as profitable as the good old days when people had crappy beige telephones. So they have to change their business model. And they can get away with it because they are a near monopoly with millions of subscribers that don’t have real choice.
So they become a new breed of traffic switching center that routes packets based on IP$… the new protocol powering their vision of the Internet.
What’s more troubling is that this isn’t just about charging for QoS. This is about the ISPs providing their own competing services. It’s THEIR network remember? Through QoS they’ll be able to provide you better VoIP, VOD, etc. service than any other provider unless of course the other providers pay for the privilege.
When you think about the next generation ISP you need to think WALLED GARDEN. In a perfect world ISP 2.0 wants a pipe going into your house with all these services coming from THEM. Think Caller ID, Call Forwarding, Call Waiting. Did the phone company give you those services for free?
The same thing will eventually happen to your phone. Do you think your mobile provider likes the fact that you can use google maps on your java based cell phone? Wouldn’t they rather sell you this type of service? In the mobile world this hasn’t happened yet… but it will. Think of your phone as a desktop and think of your GPRS, 1xRTT or EV-DO service as your DSL or cable service. Eventually the spectrum used for voice will be marginalized and most everything will be done over this “data circuit”.
You need to ask yourself one question Mark. Do you trust the phone company? No seriously… do you really trust the phone company?
You’re on the wrong side of this issue.
Joe
Dallas, TX
Comment by Joe -
Mark,
Welcome to WALLED GARDEN 2.0! What you’re advocating is the biggest threat to the Internet today.
It’s quite simple really… the oligopoly ISPs want to run the Internet similar to the telephone network. They’ll charge customers extra for fast service (long distance) and they’ll charge companies that want to provide these services faster.
For YEARS these monolithic companies have been raking in billions of dollars on voice service. The first threat was the mobile phone when they offered free long distance calls. The second and final threat is the Internet. On the web voice is nothing more than an application. Nothing more than an application? Applications are free on the Internet… GOODBYE PHONE COMPANIES. In a few years the only phone companies around will be VoIP providers and the only reason they’ll exist is to terminate calls to stragglers still on the PSTN and to provide e911 service.
So what’s a phone company to do? Die quietly? They become ISPs… but being in the ISP business isn’t nearly as profitable as the good old days when people had crappy beige telephones. So they have to change their business model. And they can get away with it because they are a near monopoly with millions of subscribers that don’t have real choice.
So they become a new breed of traffic switching center that routes packets based on IP$… the new protocol powering their vision of the Internet.
What’s more troubling is that this isn’t just about charging for QoS. This is about the ISPs providing their own competing services. It’s THEIR network remember? Through QoS they’ll be able to provide you better VoIP, VOD, etc. service than any other provider unless of course the other providers pay for the privilege.
When you think about the next generation ISP you need to think WALLED GARDEN. In a perfect world ISP 2.0 wants a pipe going into your house with all these services coming from THEM. Think Caller ID, Call Forwarding, Call Waiting. Did the phone company give you those services for free?
The same thing will eventually happen to your phone. Do you think your mobile provider likes the fact that you can use google maps on your java based cell phone? Wouldn’t they rather sell you this type of service? In the mobile world this hasn’t happened yet… but it will. Think of your phone as a desktop and think of your GPRS, 1xRTT or EV-DO service as your DSL or cable service. Eventually the spectrum used for voice will be marginalized and most everything will be done over this “data circuit”.
You need to ask yourself one question Mark. Do you trust the phone company? No seriously… do you really trust the phone company?
You’re on the wrong side of this issue.
Joe
Dallas, TX
Comment by Joe -
QoS Fee for Backbone and Tiered Flat Fee for ISP
Video, audio, and other rich media applications make Internet traffic increase exponentially; With technology progress including cable, FTTH, WiMax, BPL, …broadband at the edge for Internet access are becoming more than adequate. However, every backbone carrier is still losing money and has to be subsidized by enterprise voice and data.
Telecommunication incumbents want to recover their investment in infrastructures and are pushing for extra charge for QoS, which only a few backbone carriers, such as Broadwing’s HDTV distribution network using DTM with 100% QoS, can currently offer. This business model and practice cannot last forever. Someone has to pick up the tap.
Let’s imagine that the Internet went dark tomorrow. Who would stand to lose the most? How about those companies whose entire business models rely on a functioning, widely available Internet: Google, Amazon, eBay and their ilk. These three alone have annual revenue of $13 billion. You’d guess that these firms would welcome an initiative to put in place a framework that would allow carriers to set and receive fair-market payment for backbone services, in exchange for a commitment to deliver highly reliable connectivity.
On the other hand, end customers do not want to pay ISP twice and give ISP the right to control what they access although they are willing to pay more for higher bandwidth.
This will lead to a framework that content owners such as Google, eBay, Amazon, Yahoo, Disney, MTV, … to pay more for faster and better BACKBONE services with QoS while end customers pay a tiered (flat) monthly fee of $30 for 3 mbps, $50 for 10 mbps, $75 for 25 mbps, $100 for 100 mbps, … to access whatever they prefer as long as it is legal.
Since there are more than adequate bandwidth at the First/Last mile for Internet access, putting a tollbooth at ISP and charging extra for QoS at the gate is a pure excuse for discrimination and will only choke our economical growth.
Comment by Youlu Zheng -
I think a question one needs to ask is what does multi-tier service really means. I believe that there is a need for better QOS at the last mile as Mr Cuban argued for. (Based on work experience in this area for a number of years.)
I think that in the aggregate, same applications by different end users must have equal priority at the CO unless explictly chosen otherwise (and thus priced the same), but bigger users of bandwidth must pay more – which is the case even today – but not to the extend that accounted for priority differentiation at all endpoints. There’s a lot of bandwidth in the carriers’ pipes, but not at the last mile for all the subscribers of a CO to do VOD at the same time. Even without common agreements by the ISPs, shaping the traffic at this last mile will achieve a great deal. The ISPs will be compared by how well an user’s highest priority applications perform.
A sticky point is how should applications be prioritized: by providers or by end users? The second option is available to enterprises (with MPLS VPN, etc.) but that doesn’t help the masses; the first option can be a business opportunity for ISPs and this may be what the telcos should be doing before adding more bandwidth. The problem is they probably have little motivation to do this because it costs them money too without reaping the reward — unless the problem is understood and they are able to charge accordingly.
On the other side of the picture, it would be unreasonable to charge content providers with higher rate since they already have to pay for the bigger pipes.
Comment by Brian -
I’m still on the fence on this idea of tiered access – the market is still trying to find the right price for services (that’s why you still have people paying $25 for dial-up when they could be paying about the same for broadband). But government mandates like network neutrality would be a cure worse than the disease (see post 6 above).
If, for instance, a Time Warner customer couldn’t download Sony content as fast as he/she would like, wouldn’t they just switch to Qwest or AT&T, who already offer DSL and are starting to offer fiber optic broadband?
On the other hand, if the government prevents content owners from striking deals of any kind, it could potentially harm the Internet’s growth potential. Net neutrality, as I understand it, would prevent Google from doing too much with those miles of unlit fiber, since that would put them in the network business rather than the application business.
Comment by Raphy -
“Google is not discussing sharing of the costs of broadband networks with any carrier. We believe consumers are already paying to support broadband access to the Internet through subscription fees and, as a result, consumers should have the freedom to use this connection without limitations”
Google’s Barry Schnitt
Comment by Matt -
“The little guy is at a huge disadvantage right now.”
I don’t see that being the case at all. True high-bandwidth customers get lower $/Mbps but the level of service is much the same.
I believe you’re talking about the ISP peering scheme, where Yahoo’s partnership agreements with SBC/BT/etc get them cheap bandwidth & peering, but they don’t get priority access to customers. Who would want to access the internet in a walled garden?
I understand the little guy is at a disadvantage because their $/Mbps is higher but that’s economies of scale, true in every business. I’ve yet to hear of companies getting priority access (in a peering sense, not a marketing sense) to ISP customers at the expense of other sites. Can I have an example? Then we can prove your theory.
Your complaint seems to be focused on HDTV, which is a nice technology but for 95% of the world is not neccessary yet. I realise you have a stake in HDTV and it’s in your interest for everyone to be able to download your movies in HDTV but HDTV has not reached critical mass yet, there is no need to incease bandwidth tenfold for everyone.
Though I will agree with you on “can we please include native multicast and peered multicast.”
Comment by Adam -
Mark while you are highlighting that the internet traffic growth is outstripping the growth of the underlying infrastructure, and that there’s nothing we can do about that – the fact remains that traffic overload does not cause any of the slowdowns. The rise in trafffic and the rise in cost due to tech advances are almost getting inversely proportional. A slowdown is most likely caused because the local network is slow- or because the data center is having network issues – There is a QOS already in place and any means to impose priority on content/applications in digital highway is NO-NO in the increasingly triple-play converging world. The internet’s success itself was based on the democracy & non-discriminatory approach adopted by it. Last mile problems can’t be solved by multi-tiering
Comment by Sadagopan -
Some companies also creating thier own Walled Garden Networks like Peer Impact and Kontiki and Red Swoosh(Mark is an investor) using p2p technology .
They use thier own QoS in thier Walled Gardens and take factors like node speed ,ISP and geographical location to determine how they will conect Peers .
If your Peer is on a node that has the same ISP as you I know that Peer Impact and Kontiki will attempt to keep as much traffic as possible within network and as close to your node as possible , I would think Red Swoosh would do the same .
Bittorent unfortuantly uses a peercache in the form of a tracker wasting alot of bandwith.
Comment by Matt -
Love the discussion on this. Some of the dissenters might want to look at reality. The little guy is at a huge disadvantage right now. All the companies you are worried about taking over bandwidth at the expense of equal access and little companies, well guess what, that is exactly what is happening right now. but instead of organized and defined programs, its done through bulk.
Buy enough bandwidth and things get tuned your way. Buy a DSL line, and you compete with the masses.
If there are established programs where its not about how much bandwidth you buy, but rather how you integrate into the network and the Service levels you acquire, then the network has the opportunity to be much more equal opportunity. Everything is published. (which i think should be a requirement for any tiers of service).
Then of course htere are the CDNs like Akamai. Akamai basically buys in bulk, adds their software and marks it up. Its an incremental tier of service, just once removed from the network. I would rather see an integrated bit torrent API with the client part of every customer install that any and all of us can write to. Pay the money to the network for the incremental software and service instead of the CDN if they can match or exceed the CDN service, why not ?
If the tiers are standardized, then we might even see some inter network quality of service rather than havig to have 10 zillion servers populated every where.
then mission critical apps can have some standard to write to. Hardware vendors and software vendors can create tools to monitor the results of those standards and their application interfaces.
and while we are at it, can we please include native multicast and peered multicast. Now that is a tier of service I would definitely pay for that would definitely improve the opportunities available to one and all across the net.
Comment by Mark Cuban -
I used to be involved with a site that was on S.W.Bell’s “Best Of” list. They included our site more or less as part of their service offering. Our content made them look good and helped them win customers, and in return they linked all their customers back to us from their web site. No money changed hands. It was a great arrangement. Had they tried to charge us for that, I would have told them where to stick it. Whoever was in charge back then had a clue.
Do you know that what you are proposing works in reverse too? Whatever HDNet pays for top QOS to each carrier, I will be able to go to each carrier and pay double that fee to ensure HDNet content sits at the very bottom tier with the lowest priority packets. Remember, if a Telco can charge extra to unblock the number of a customer paying for caller ID block, they certainly won’t hesitate to take my money to unprioritize your packets. This is why this proposal stinks.
You’re a very, very smart guy. What on earth are you thinking? It’s one thing to charge the consumer for priority access to anything they want to view (i.e. dial-up vs. broadband), but to handicap every non-paying content providers in this manner amounts to racketeering. What it means is that Indy producers will have yet another obstacle in their face in a medium that is supposed to be a level playing field. Artist revenues will now be stolen – yes stolen – from the revenue stream to pay the unnessesary overhead costs imposed by this proposal.
Lastly, as a consumer, I want priority access to everything. Big, small, sensible or really stupid. It’s my browser, and I’m going to use it. Why would I ever use a service provider that cannot give me priority access to everything. Bigger, faster, better.
This proposal is bad for the consumer. It’s bad for the artist/content creator. It’s bad for the content provider. It’s bad for the Internet.
Comment by Simon Higgs -
Petabox Specs
OVERVIEW
The petabox(tm) by the Internet Archive is a machine designed to safely store and process one petabyte of information (a petabyte is a million gigabytes). The goals– and current design points are:
* Low power– 6kWatts per rack, and 60kWatts for the whole system
* High density– 100 Terabytes per rack
* Local computing to process the data– 800 low-end PC’s
* Multi-OS possible, linux standard
* Colocation friendly– requires our own rack to get 100TB/rack, or 50TB in a standard rack
* Shipping container friendly– Able to be run in a 20′ by 8′ by 8′ shipping container
* Easy Maintenance– one system administrator per petabyte
* Software to automate mirroring with itself
* Inexpensive design
* Inexpensive storage
http://www.archive.org/web/petabox.php
Comment by Matt -
@Daniel Zappala – Re-read the quote in my comment #23. SBC stated that it wanted to charge content providers to allow them to send data over their pipes to SBC customers. Now, this quote did kick over a bit of a hornet’s nest and SBC may have issued a retraction or “clarification,” but the quote provides enough context to eliminate any ambiguity as to intent. Keep in mind, these are Fortune 500 companies that are typically run by accountants or lawyeres – all they see is a huge pile of CapEx that is someway involved with someone else’s business. They don’t understand the technology aspects or the ‘cultural’ issues involved – they just see “hey, money is going over our pipes – we should get some!” Now, this would be as stupid as charging businesses for any deals consumated over a telephone conversation – even though both parties pay their providers for access – but that’s what you get for putting people who lack perspective in charge of companies.
@Matt – do you have any documentation for your claim that those are the Internet Archive servers? I think that makes more sense than portable data centers – if I were Google (and judging by my bank account, I’m not), I’d just lease space and capacity in CoLo facilities through one or two dozen shell companies and then one day yell “Surprise!” – but some backup would be nice before I repeat your claim.
Comment by Erik Carlseen -
Mark,
I disagree. In theory, your idea is correct, but we’re talking about the ‘Baby’ Bells here. The tone of their arguments is already starting to sound Mafiaesque — ‘wouldn’t want anything to happen to your data now?’ Let’s look at anything they’ve done: has it been implemented fairly? No. Has it been implemented well? No. Has it been implemented super-profitably? Hell, yes. Who wins? Seidenberg and crooks like him–and customers lose.
The Internet succeeded because of the level playing field that allowed anyone with a broadband connection to transmit their data to the rest of the world. Why would we expect the telcos to allow this to continue when they can gouge the likes of Google and MS to an extend that the little guy can’t afford?
Looking at broadband penetration in other Western countries, we see that the US is far behind. Why? Of course geography is part of it, but the telco cartel is a big reason. It will only get worse if they are allowed to charge customers for data-transmission extortion.
Your idea is good in theory, but do you really trust the telcos to do it right? And if the market hasn’t already corrected their bandwidth-gouging rates (compared to many Western countries), why would it correct further attempts at gouging?
I hope to hear your remarks on this. The practical implementation details are too important to ignore.
Comment by Matthew Hornyak -
Here’s what most of you don’t get: people with money always get their way. Continuing with Mark’s analogy, the US interstate system is the most easily accessible and commonly used transportation system on the planet. But it’s not fast enough for some. So people with a little more money choose instead to buy an airline ticket. But if you have even more money, you can bypass the inconvenience of crowds and restrictive schedules by taking a private jet, as Mark does.
Mark is saying he’d like to have a couple of lanes on the interstate system which gives him the right to drive at unlimited speeds in exchange for paying a toll. By raising and lowering the toll, you could control the number of people who choose the boogy lanes.
Do you think there’s a market for high speed toll lanes on the freeways? Absolutely. Do you think it will ever happen? No way — too much political dynamite. So the folks who want to get quickly from one place to another, and can afford to pay a lot to do so, take to the air in jets and helicopters. I remember being at a Super Bowl in Miami when there was a holding pattern for access to the helicopter pad. Definitely out of my league.
So will the big morass of the Internet ever have high speed toll lanes like Mark wants? Who knows. But one thing I do know — if the public internet doesn’t find a way to answer the need, the folks who can afford it will just find another solution. Maybe that means a shadow Internet backbone that you have to pay a lot to use. Do you think the richest million people in the country would pay 10x their ISP bill to get great performance all the time? You bet they would.
Comment by Paul Lambert -
Funny, I was just writing about this topic at about the same time, but from the point of view of what a disaster this would be.
What concerns me is that if you head down this path, you deliver the carriers the very tempting leverage tool of letting their existing best-effort networks degrade just enough to make many of today’s services unusable (be it VoIP, bitTorrent, online gaming, ipTV, or podcasting) — in turn letting them strong arm the app and content providers into paying fees for priority service. How many carriers do you think will be able to resist doing this? Hint: read Ed Whitacre’s quote from business week again.
Comment by Steve Smith -
Actually, I already pay my ISP. I have a choice of purchasing connection different speeds.
But what you are proposing is that I (I run several sites) pay Cox (as an example) extra money to ensure that my content gets delivered in a timely manner to one of their subscribers who is already paying for the bandwidth? Sounds like a double dip to me?
I like this option better….
Since I have to pay to host my servers and the bandwidth they use shouldn’t I be charging Cox a fee for allowing their subscribers to access my services?
I think my way sounds more fair…
Comment by Steve Ciske -
An important distinction to make: SBC says it wants to charge content providers, and give them preferential service. Consumers say this is exactly what they don’t want. On the other hand, there is a long history of QoS standards (some available from Cisco for years) that allow ISPs to charge users for preferential service. I believe this is something that consumers may actually be willing to pay for, since it is an extra fee for an extra service. This would be perfectly appropriate for the utopia Mark describes, in which high definition video is available on demand (for those who pay).
The problem with Internet QoS standards has always been getting global deployment and interoperability. If I pay my ISP $X/minute, I need my ISP to communicate my service requirements to its provider, along with a portion of my fee to guarantee those requirements are met. Until the large ISPs get together and decide to do this, we are left with individual ISPs trying to make a buck by charging the providers instead.
Comment by Daniel Zappala -
The highway analogy isn’t quite right. Adding 20 lanes to a highway isn’t feasable because its too difficult to exit from 20 lanes over. Data doesn’t have that same limitation.
Comment by jeff -
Bob Cringley’s rumours about the Shipping Container in the Google carpark is actually the Internet Achive’s servers.
Comment by Matt -
Bob Cringley’s rumours about the Shipping Container in the Google carpark is actually the Internet Achive’s servers.
Comment by Matt -
the bells are not talking about making a new path to put these extra charged content providers on, they elude to artificially slowing down the rest of the internet. since their customers are paying to be on their network wouldnt they expect to be able to get to all sites without henderence from the isp they pay?
Comment by jerry scroggin -
http://internet2.nd.edu
Also, what about Google’s super-caches. All these rumors of a shipping container with all the super caching capability (http://www.pbs.org/cringely/pulpit/pulpit20051117.html)?
Sorry Mark, this one seems a little alarmist. People always come up with either a better transport/caching system or bigger pipes. They start by charging an arm and a leg as they get started (creating your HOV lane) and then eventually move it to the masses.
Comment by Joe S -
For many people reading these stories, all they see are telecoms trying to get paid twice for the same bandwidth.
When I download something, the content company isn’t using BellSouth’s network. I, as a paying customer, am using the BS network.
I have an upper “speed limit” and a monthly thoughput cap. As long as I remain under those limits, that BS has set and I have agreed to … I don’t see where they can demand another dime.
If I go over the limits – hit me, the consumer, or have the rules kick in and throttle my usage.
No, these guys are trying to wring out more money from bandwidth that they’ve already sold to consumers.
It’d be like the city and state charging The Mavericks road tax specifically for me because I drove to see a game.
Comment by Mel -
Bitorrent acually performs better when more people are trying to access a file (as long as you have a few stable seeders at first, but that’s usually not a problem with popular files).
Comment by Michael G. Richard -
Mark,
I would suggest taking a look at Teletruth and what they have to say about the Bell’s.
http://www.newnetworks.com/breakupBadBellMergers.htm
Comment by Frank Muto -
I think this whole problem would be avoided if firms with “mission critical” needs would hire Internap(www.internap.com). They are the best for mission critical QoS. Since they but transit from all of the backbones, traffic carried by Internap is guaranteed the highest priority.
Comment by Just Me -
I think this whole problem would be avoided if firms with “mission critical” needs would hire Internap(www.internap.com). They are the best for mission critical QoS. Since they but transit from all of the backbones, traffic carried by Internap is guaranteed the highest priority.
Comment by Just Me -
End-to-end QoS was one of those Internet concepts that sounded like a great idea at the time, but died an appropriate death – in reality, it’s a horrible idea. For evidence, I present the following quote from SBC CEO Edward Witacre, in a BusinessWeek Online article (link: http://www.businessweek.com/@@n34h*IUQu7KtOwgA/magazine/content/05_45/b3958092.htm – sorry for the mangled link but TinyURL is blocked) on Nov 7, 2005:
Q (BusinessWeek): How concerned are you about Internet upstarts like Google (GOOG ), MSN, Vonage, and others?
A (Witacre): How do you think they’re going to get to customers? Through a broadband pipe. Cable companies have them. We have them. Now what they would like to do is use my pipes free, but I ain’t going to let them do that because we have spent this capital and we have to have a return on it. So there’s going to have to be some mechanism for these people who use these pipes to pay for the portion they’re using. Why should they be allowed to use my pipes?
The Internet can’t be free in that sense, because we and the cable companies have made an investment and for a Google or Yahoo! (YHOO ) or Vonage or anybody to expect to use these pipes [for] free is nuts!
My answer to Witacre: Your customers are paying the for access, you stupid, greedy [multiple expletives deleted].
A CxO at Verizon has made similarly stupid, bleating noises (sorry, I don’t have a link for that one). The problem with differentiating traffic on the Internet is that the large ISPs don’t know when to quit – they’ll try to turn every peering point into a virtual tollbooth – a Y2K extortion scheme for businesses that would make the mafia proud (“Those are some nice packets you’ve got there… it would be a … shame … if anything should happen to them”). I’ll leave the “sleeping with the fishes” joke as an exercise for the reader.
Even if you didn’t have the problem of idiots who don’t understand the Internet running telcos, it’s still a ‘non-trivial’ (read: damned near impossible) problem to decide what the ‘fair’ priority is as packets move from network to network – everyone has different agendas and priorities. The best thing you could hope for is an ungodly complicated micropayments system. You think carriers bitch and moan (and occasionally shut down access) over peering now? Take that about eight orders of magnitude higher (no, I know what you’re thinking – that they could work out a method for peering that fairly compensates everyone – history says they won’t).
With regards to distributing media, the answer will be in distributed technologies like bittorrent, or the rumored “Google Shipping Containers” (link: http://www.pbs.org/cringely/pulpit/pulpit20051117.html ) (which, like all rumors, must be taken with an appropriate amount of salt) or similar caching technology close to the endpoints – you could, for instance, sell people a “lockbox” that they place on their home LAN that downloads encrypted content before it can be unlocked – once the release date / time hits, they allow the content to be accessed – kind of like a preemptive TiVo. Those are just examples, people who actually care will probably come up with better ideas.
Also on the ‘end-user’ side, your example of Grandma’s medical stuff being prioritized over Little Timmy’s pr0n downloads (or whatever) is best handled at end-user’s router / firewall. These things will become more intelligent and easy for users to configure – actually, they need to get there now, as anyone who has tried to talk to a friend’s Vonage line while said friend surfs the net already knows. And now and in the future, anyone who trusts the Internet for life or death traffic deserves whatever they get. That’s what private networks are for.
Comment by Erik Carlseen -
Following up on post #20 – the next wave of networking will be about prioritizing traffic based on application and data on a per user basis. The question is will you want your service provider looking at your health information to see if its important enough to prioritize over someone else’s health information. Do you even want your service provider(s) doing this ?
Comment by Shake -
The importance of data changes over time. A doctor’s message stating that someone is healthy is less important than an emergency call for help. So, an automated prioritized traffic system is going to need statistics and user input to figure out what traffic is important. It currently is hard coded to planned out quality levels.
Comment by David Skowronski -
Mark I absolutely love your style of writing and the post makes alot of sense!! However, I agree with a few others who have said that the internet’s success was much do to the freedom of the marketplace. People can pretty much do whatever they want (to an extent of course). I wouldn’t want to see any regulations put into place. If these larger companies need to have constant lightning speed access to the web then they won’t mind paying high-dollar for premium access through comcast or charter or whomever. They also won’t mind paying 10 grand for just ONE database or application server when in fact they may be needing many of each. There are hundreds of small companies and several very large companies like Level 3 communications who have an interest in expanding the speed and capabilities of the web. Give them time…. they’ll get to it. They know they are behind the ball and they are working to fix the problem. Until then, all you freaks out there will have to wait to download all your hard core extreme bondage porno!!
Comment by John Pearson -