Enough alarmism on peering disputes from all sides
Nate Anderson of Arstechica posted this article today slamming the latest study from A.T. Kearney on the A Viable Future Model for the Internet commissioned by European broadband providers, but it’s led to more unnecessary alarmism on all sides. The Kearney report make the argument that an impending infrastructure crisis is upon the Internet unless they can charge content providers for bandwidth, but this always comes off as accusing the content providers (especially the ones named in the report) of free-riding. The content distributors then act like they’re victims of attempted extortion and then go on to cite the myth that all web and content providers should have the same capability on the Internet under “Net Neutrality” when the reality is that content providers have always had to pay more for superior bandwidth to some network carrier (not necessarily the broadband providers) in one form or another. But these alarmist arguments from either side do nothing to advance Internet public policy and they pose a real risk of unwanted government intervention on both sides of the Atlantic.
Why even imply free-riding?
The A.T. Kearney report doesn’t make any specific charges of free-riding against the content providers, but the alarmist tone gives the content providers all the ammunition they need and Nate Anderson went out of his way to play up the free-riding charge. The alleged proof of the ISP’s evil intentions cited in Anderson’s story is this excerpt from the Kearney report:
“Most Online Service Providers pay a fee to their Connectivity Provider(s) to be connected to the Internet, which is generally based on the bandwidth they require, while the largest ones act as if they were Connectivity Providers in their own right and connect to others via peering agreements. In both cases these charges are generally flat fees, not linked with usage and they form a very small part of their total expenditure/cost structure. In effect, Online Service Providers are paying to connect their services to the network but are not paying for downstream service delivery.”
While the report’s intension may not have been to suggest that content providers were free-riding, I can see how it came off that way. If the online service providers are paying their connection charges which may be going to a transit carrier, a CDN provider, or the broadband provider, what difference does it make if they’re paying for “downstream service delivery” or not? If the broadband provider wants to compete against the transit providers or the CDN providers for the content providers’ business, just offer a better service at a more competitive price to the content providers. Why imply that the content providers aren’t paying for something that they should be paying when they should be arguing that the content providers might be paying too much and that broadband providers can offer a better service?
It’s baffling why the report would imply that the content providers are currently paying small “flat fees” but they should be paying broadband providers usage fees instead. This played right into the hands of the critics of broadband providers. The fact is that the content providers would often end up paying less to the broadband providers for higher quality bandwidth than the connection charges they currently pay to transit providers. Why even make these charges when broadband providers don’t have the business or political clout to block the content providers using alternative methods for delivery, but come off looking like the bad guys? If there are bad content players like Netflix and Level 3 demanding hundreds or thousands of Gbps of additional private peering capacity from broadband providers for free, then call them out but don’t just blame all content providers.
Scaremongering on usage based billing
Unfortunately, Anderson’s article interpreted the the Kearney report in the worst possible way. Anderson argues that the ISPs want to “extend usage based billing” from consumers to the content providers. That is an effective scare tactic because the audience is generally only familiar with consumer Internet billing which is largely flat rate based in the US, and they hate the idea of usage based Internet billing which is negatively associated with massive bill shocks. But in the commercial internet space, there is no substantive difference between “usage based billing”, flat rate billing, or 95th percentile of peak bandwidth billing. At the end of the day, all three forms of billing works out to roughly the same dollar figure and to suggest that usage based billing for content providers is somehow new and evil is unwarranted scaremongering.
It’s easy to show why there’s nothing fundamentally bad about usage billing. If a content provider is billed $1/Mbps/month for very large quantities of bandwidth on peak 95th percentile usage, and their average traffic level is 1/3 of the 95th percentile levels, then it works out to roughly $0.01 per GB transfered. What difference does it make if the content provider is billed on $1/Mbps peak 95th percentile usage or $0.01 per GB transfered? Why so much alarmism over an analyst’s estimate of ISPs offering a €0.05 per GB peering service when it really isn’t all that different from a peak connection rate charges? The only reason to be alarmed is if the broadband providers threaten that they’ll block or unreasonably degrade all other methods of content delivery not involving payments to the broadband provider and that hasn’t happened. So long as there are no unreasonable threats, the broadband providers should be able to offer any service at any price they want and the content providers are free to accept or reject them as they see fit.
The danger of government intervention on the Internet
The content provider powerhouses in recent years have been tempted by their newly found success on the Internet and new powers in D.C. and Brussels. They are the new face on the political block which comes with an aura of innocence untainted by the evils of politics but that gets old quickly if it hasn’t already. It won’t take long for these companies to realize that any early lobbying success quickly turns into an ugly and never-ending quagmire when their attention is better spent on continuing innovation if they want any chance of warding off more nimble startups.
The major Internet content providers (who are largely based in the United States) like Google and Netflix are having some success in getting the kind of government intervention they want and they are working to get the FCC to eventually ban all types of paid peering, but they should be mindful that French politicians are looking to do the reverse and intervene on behalf of European broadband providers to charge content providers more. But both extremes are wrong because it’s taking a free market system responsible for the innovation on the Internet and turning it into a partisan and nationalistic political battle. For both sides of dispute, they would do well to tone down the alarmism a few notches and turn their effort to more pressing business matters.
George,
You are totally misrepresenting A.T. Kearney’s position. They propose a $0.05 per gigabyte tax on all traffic to access network owners. In no way does A.T. Kearney propose that access networks in any way position themselves to compete more effectively with transit providers. As the proposed solution is an unavoidable tax, set unilaterally by the access network owner, content providers are NOT free to accept or reject them as they see fit.
If you do not believe me, please re-read pages 33-34. You will also find that A.T. Kearney proposes to change all settlement free peering agreements to paid peering agreements.
“If a content provider is billed $1/Mbps/month for very large quantities of bandwidth on peak 95th percentile usage, and their average traffic level is 1/3 of 95th percentile, then it works out to roughly $0.01 per GB transfered.”
This sentence makes no sense. 1/3 of 95th percentile? Please edit for clarification. $1/Mbps/month equals about $0.003125 per GB. You probably want to say something about peak to mean ratios to conclude an average cost of $0.01 per GB.
Furthermore $0.05 per GB is truly something to be alarmed over. It is on average 10-100 more than the going price for IP transit. In other words, it would be a substantial tax, dwarfing all other costs.
Again, please don’t take my word for it. Refer to the following web sites for typical IP transit and commercial peering costs.
http://he.net/
http://www.nl-ix.net/products/pricing/#nlix_public_peering_products
You will note that global IP transit is quoted at $1/Mbps/month and domestic IP transit as low as $0.20/Mbps/month.
“The only reason to be alarmed is if the broadband providers threaten that they’ll block or unreasonably degrade all other methods of content delivery not involving payments to the broadband provider and that hasn’t happened.”
How is this different from what Comcast is doing to Level3?
[...] This post was mentioned on Twitter by AdamThierer and jbsay, GeorgeOu. GeorgeOu said: http://bit.ly/eTS7F0 Looks like another Ed Whitcre moment for ISPs. Ars overdrive on scaremongering over expansion of usage based pricing. [...]
@Zed
Let me first clarify what I mean by an average traffic level 1/3 that of 95th percentile levels. If the absolute peak is 500 Gbps but their 95th percentile is only 300 Gbps, the customer is billed at 300 Gbps and not 500 Gbps. But to compute the actual GB usage per month, we must base that on average bandwidth and not the 95th percentile level. If that average is 1/3 of 300 Gbps, then the average bandwidth is 100 Gbps can be used to compute the amount of data transfered per month.
You claim that “$1/Mbps/month equals about $0.003125 per GB”, but thats assuming that average equals peak 95th percentile levels which is a flawed assumption.
1 Mbps/month is the billing rate applied to 95th percentile peak load.
1 Mbps * 3600 seconds/hour * 24 hours/day * 30 days/month * 1MB/8Mb * 1GB/1000MB = 324 GB
However, I’m using the example where average traffic level is 1/3 of the 95th percentile which means average bandwidth is .3333 Mbps. That means on average we’ve only transfered 108 GB per dollar which works out to roughly 1 cent per GB a I stated in the post.
@Zed
“You are totally misrepresenting A.T. Kearney’s position. They propose a $0.05 per gigabyte tax on all traffic to access network owners. In no way does A.T. Kearney propose that access networks in any way position themselves to compete more effectively with transit providers. … If you do not believe me, please re-read pages 33-34. You will also find that A.T. Kearney proposes to change all settlement free peering agreements to paid peering agreements.”
First, you’re misinterpreting the Kearney paper’s position when you interpret that they’re proposing an access tax to “all traffic” which would include the transit links and best effort traffic. Kearney proposes an alternative to the best effort transit model where content providers switch to paid peering circuits. Kearney is not calling for a “tax” or an additional fee on existing traffic coming over transit.
Second, A.T. Kearney or even the EU ISPs can propose whatever they like so long as it is a voluntary agreement. Until they threaten to block or actually implement a blockade on the transit link, or conveniently forget to carry some destinations on their transit link, there is nothing to be alarmed about.
Third, I criticized the Kearney paper for not positioning their proposal as the more cost/service competitive model. I also criticized them for their tone which lent itself to misinterpretation and alarmism from the content provider side.
@Zed
“Again, please don’t take my word for it. Refer to the following web sites for typical IP transit and commercial peering costs.
http://he.net/
http://www.nl-ix.net/products/pricing/#nlix_public_peering_products
You will note that global IP transit is quoted at $1/Mbps/month and domestic IP transit as low as $0.20/Mbps/month.”
First of all, the accurate price in the second link you provided is $1.9/Mbps/month for a 1 Gbps interface with 3 year contract. That’s a fairly steep commitment.
Second, there’s different quality of bandwidth and some of those lower price tiers are obviously for shared transit since these look like relatively small hosting packages.
The price for dedicated tier 1 backbone transit as of early 2010 is around $3 to $10 per Mbps per month depending on commitment level. Prices are continually declining rapidly, but usage is going up as well.
Leave your response!
Twitter Feed
About Us
Digital Society is a digital think tank that believes culture and commerce are inseparable, that the digital economy flourishes when people are free and rights are secure, and that free markets free people.
Digital Society is an independent 501(c)3 non-profit organization, funded by donations from Jon Henke and from Arts+Labs. We advocate for a pro-culture, pro-commerce digital society through research, analysis and debate on emerging technology issues.
Reply Comments
Transparency and interactivity are trademarks of the Internet era, and we aim to foster them here at Digital Society. It is inevitable that some people will disagree with the technology policy positions we take. We want to have that constructive debate.
The Reply Comments feature gives our critics a chance to respond to our viewpoints and the Digital Society audience convenient access to competing arguments. Any time we directly challenge the views of an individual or a group on this site, the party in question may substantively respond in a guest post.
Please contact executive director Jon Henke by e-mail.
Subscribe
Daily Digest Email
Recent Posts
Recent Posts
Most Commented
Most Viewed