Light Reading
Comcast's characterization that dispute is about peering 'could not be more misleading,' Level 3 insists UPDATED 12/4

Level 3: This Is Not a Peering Dispute

Jeff Baumgartner
LR Cable News Analysis
Jeff Baumgartner
12/3/2010
50%
50%

Level 3 Communications Inc. (Nasdaq: LVLT) lobbed another missile at Comcast Corp. (Nasdaq: CMCSA, CMCSK) today, claiming, in the form of a lengthy Q&A-style statement, that the MSO's characterization that this dispute is about peering "could not be more misleading." (See Did Level 3 Know What It Was Getting Into? and Level 3: Comcast Erected Web Video 'Toll Booth' .)

In Level 3's view, the debate is centered on "interconnection" between communications networks rather than a peering between two Internet backbone networks. (See Level 3 Wants to 'Clarify' Comcast Dispute.)

"Unlike 'peering' in the Internet backbone, where competition abounds and prices have been declining steadily, Internet carriers that have content requested by Comcast subscribers have no choice but to exchange traffic with Comcast," Level 3 said. "Comcast is using this dominant position to demand payment for traffic delivered at its customers’ requests. You simply cannot 'route around' Comcast to provide requested content to Comcast’s subscribers."

Interconnection, the company adds, "is a general term that applies when two communications networks exchange traffic, regardless of the commercial terms that are agreed." To back up its original "toll booth" argument, Level 3 said Comcast "wants to change the rules of the game" by using its local access network as "leverage to force Level 3 to pay for traffic requested by Comcast customers that already pay Comcast for access to that same content." Level 3 also claimed that no other broadband US ISP is charging the type of fees the MSO is now demanding.

Level 3 did acknowledge that that it will be dumping more traffic onto Comcast's networks, but fears that the recasting of the debate all but guarantees that traffic will remain "out of balance" because a customer who requests to watch a Web video typically sends a small upstream request to receive a much larger amount of downstream traffic.

"Comcast stands to make many millions of dollars from Internet backbone carriers that bring requested content to Comcast for delivery to Comcast’s subscribers," Level 3 said, claiming that the dispute doesn't have anything to do with its new Netflix Inc. (Nasdaq: NFLX) deal, but is about the "precedent" that Comcast sets with respect to the handling of over-the-top video services.

Likewise, Level 3's also sticking to its guns that this debate is indeed a network neutrality issue, holding that Comcast's ability to set pricing and "unpredictable" tolls will "become a tool to reduce competitive content available on the Internet."

Comcast's take
Comcast, meanwhile, continues to portray this as "a good old-fashioned commercial peering dispute" that is not about online video, paid prioritization or the installation of Internet toll booths," according to a letter dated November 30 to Federal Communications Commission (FCC) Wireline Competition Bureau chief Sharon Gillett from Comcast SVP of external affairs & public policy counsel Joe Waz.

"In order to undercut its CDN competitors, Level 3 wants to avoid the commercial arrangements other CDN companies use to terminate traffic onto Comcast's and other providers' networks, and instead force Comcast to accept its CDN traffic for free, under a 'peering' relationship," Waz wrote.

The letter claims that Level 3 approached Comcast shortly after Level 3 reached a "low ball" Netflix deal about securing 27 to 30 new interconnection ports to support the delivery of more traffic to Comcast's network "for free." Waz said Comcast scrambled to provide it with six ports at no charge, and told Level 3 it would provide additional ports on a "commercial wholesale basis" as it does with other CDN providers.

Comcast also posted a video on November 30 featuring Comcast EVP of national engineering and technology operations John Schanz explaining "how Internet peering works":



Update: Comcast responded with a statement of its own Friday, claiming that Level 3's latest statement treads already-traveled ground and that its network-neutrality-related allegations are false:

Level 3 has said nothing new. The fact remains this is a business dispute regarding traffic ratios, commonly referred to as peering, between Comcast and Level 3 which we are committed to resolve fairly and consistently with established industry principles. Industry experts and analysts overwhelmingly agree, as their commentary has shown all week long.

The most important thing to know about this dispute is that Comcast will do absolutely nothing to impact our high-speed Internet customers, who can and will be able to access any Internet content they want, including streaming video from all sources.


— Jeff Baumgartner, Site Editor, Light Reading Cable

(15)  | 
Comment  | 
Print  | 
Newest First  |  Oldest First  |  Threaded View
Page 1 / 2   >   >>
OldPOTS
50%
50%
OldPOTS,
User Rank: Light Beer
12/5/2012 | 4:16:49 PM
re: Level 3: This Is Not a Peering Dispute


Found the video helpful.


I thought an example of people on both networks going to a bank on the other network would have been helpful to explain how there used to be a good balance of inter network traffic.


But the large amounts of video has now made this in-balance significant if one network is the source of a single video provider, both because of the significantly larger BW of video and the fact that it is almost entirely one way.


At the heart of these agreements was the expectation that in the long run things balance out or be of little consequence. But with a few video sources and the higher cost for video BW the impact of these agreements can significantly impact profits. Thus the maneuvering and charges and counter charges.


OP

pdecker
50%
50%
pdecker,
User Rank: Light Beer
12/5/2012 | 4:16:47 PM
re: Level 3: This Is Not a Peering Dispute




<div style="color: #000000; font-family: Verdana, Arial, Helvetica, sans-serif; font-size: 10px; background-image: initial; background-attachment: initial; background-origin: initial; background-clip: initial; background-color: #ffffff; margin: 8px;">

regarding peering...


i'm having trouble understanding how traffic flow -direction- results in the cost for that traffic being higher or lower for a particular ISP on its side of the interconnection. &nbsp;each bit has to flow through both networks for traffic flowing between them, so it seems like both networks have to pay for those bits since they travel through both networks. &nbsp;an increase in traffic flowing from Level 3 to Comcast results in an increase in total traffic for -both- Level 3 and Comcast, and both will have to pay for the infrastructure to support that traffic. &nbsp;it's not like a balance sheet of "bits in" vs "bits out"... both bits in and bits out contribute to energy and maintenance costs, as well as revenues from customers sending and receiving the bits.


now, i accept that through certain links on the network the cost per bit is higher in one direction than the other, due to network topology and technology limitations, a prime example being the last mile to cable internet subscribers. &nbsp;this would seem to suggest Comcast should be prepared to better able to handle inbound traffic than outbound for content delivery to residential customers, since that is what those customers want and what the technology is equipped to handle as Comcast also sends some of their own video-on-demand traffic through their network to customers. &nbsp;Similarly, Level 3 should be focusing on outbound from their customers' servers out to their customers's clients, because that is the direction of traffic flow that will be in demand. &nbsp;so the cost should be minimized within both networks in the traffic direction that is of concern with this dispute, since that is the direction in demand by either side's customers (Comcast subscribers and Netflix). &nbsp;both networks will have to carry each bit flowing between those customers, so the only way the traffic "imbalance" should hurt Comcast is if they didn't expand their capacity in ways to best provide for the most likely scenarios of traffic growth.


is this coming down to demand exceeding bandwidth faster than it can be expanded? &nbsp;if so this should be Comcast's problem and not Level 3's, as without demand from their customers the traffic would not be entering their network from Level 3. &nbsp;Neflix and Level 3 can't push more data onto Comcast's network than Comcast's customers demand. &nbsp;furthermore, if Comcast wasn't in touch with their customers' demands well enough to plan properly, the onus should be on them to fix and pay for the problem, and they will have pass that cost on to their customers, as their customers are the ones demanding the increased service. &nbsp;it seems to me like Comcast is asking Level 3 to pay for expanding the network on the Comcast side of the border as a result of Comcast's customers wanting to take advantage of the increased bandwidth on Level 3's side of the interconnection border. &nbsp;this is&nbsp;analogous&nbsp;to roads on either side of a county line: &nbsp;if the road went from 3 lanes to 4 lanes, it seems silly to ask the county with 4 lanes to pay for the expansion from 3 to 4 to reduce the congestion on the 3 lane side.


i must be missing something... i guess because i'm a physical layer guy ^_^


also, repeating "[it's] worked for over 10 years" is not a sufficient justification for continuing a practice. &nbsp;10 years ago people weren't trying to stream video to their home on the scale they are now.


that the speaker in the video basically ignored the question on net neutrality makes it reek of the "TRA-LA-LA I'M NOT LISTENING!!" style of rebuttal to hearing an accusation Comcast doesn't like.

</div>



opticalwatcher
50%
50%
opticalwatcher,
User Rank: Light Beer
12/5/2012 | 4:16:46 PM
re: Level 3: This Is Not a Peering Dispute


Comcast customers pay Comcast for access to "The Internet". Most customers (excluding people doing p2p) are looking at things on the net: searches/maps/blogs/youtube videos/government forms/discussion groups (reddit, etc), looking at friends pictures (Facebook,Flicker), and Netflix videos.


ALL of Comcast's traffic is heavily one sided--toward the customer. So Comcast could make the same argument for everyone and their interconnecting network provers, including those serving Google, Facebook, Amazon, New York Times, even the Government.


Shouldn't they charge them all? What if they don't pay? Does Comcast cut them off? Then Comcast's customers no longer are connected to "The Internet" as they typically understand it.

opticalwatcher
50%
50%
opticalwatcher,
User Rank: Light Beer
12/5/2012 | 4:16:44 PM
re: Level 3: This Is Not a Peering Dispute


OldPots,


In your example, would B be Netflix, and C be Level 3? I believe Level 3 has a large trans-continental network.


Which means A is Comcast. Don't A's customers pay for access to various B content providers? Isn't this what home internet connection is all about? Isn't that bandwidth pretty much the same (or at least capped), and therefore unrelated to B's pipe size? If they use their bandwidth to connect to one content provider, that just means they are not connecting to another.


This couldn't have anything to do with Comcast having their own paid movie download service (in fact, on their website they've even announced a DVD by mail service!) This looks more like A is making a decision that only one particular 'B' should pay.


What if Level 3 (or any other 'B' or 'C') refuses to pay? Does Comcast cut them off? How would Comcast's customers, who think they are paying Comcast for access to all Internet content, react to this?


If I were Level 3 I'd call their bluff.

pdecker
50%
50%
pdecker,
User Rank: Light Beer
12/5/2012 | 4:16:44 PM
re: Level 3: This Is Not a Peering Dispute


i agree that the short BIG pipe is likely to cost significantly less than the many small and intermediate pipes forming the distribution net, &nbsp;if nothing else because of the number of truck-rolls. &nbsp;


however, this is an argument about the cost of upgrading the capacity of the network from one end to the other.


from what i understand from the video and your statements, "peering" is about the flow of traffic through the interconnect, not the capacity of the system on either side or upgrading that capacity. &nbsp;as far as i could tell, it was something like "we're agreeing to exchange ~ZZZ Gb/s in each direction across the interconnect, for $Y". &nbsp;How that traffic gets to and from the interconnect within each network was each ISP's own business and problem.


however, what i don't understand is why the "balance of traffic" was supposed to be why things were agreeable to both parties. &nbsp;increasing the total traffic through the link will cost both sides money independently of whether it's traffic sent from A to B or from B to A. &nbsp;it probably will cost either A or B more to upgrade to increase that total capacity for traffic, but it will cost both sides money to pass more traffic.


i think of this kinda like two resistors in series, which don't care what direction the current is flowing, either way the total power consumption will be I^2 * R. &nbsp;though i'd want to add some diodes, etc, to make this more closely model it.


my point is that peering is supposedly about traffic "balance" across the interconnection, not total traffic within each network. &nbsp;which still doesn't make sense to me, why the "balance" is important instead of the total traffic.


your intermediary connection e.g. longhaul company C serves to further illustrate this point, as ideally both A and B should have to pay C for every bit flowing across from A thru C to B or from B thru C to A. &nbsp;I suspect your point is that currently every bit A -&gt; C -&gt; B is paid by B, and every bit B -&gt; C -&gt; A is paid by A, i.e. paid upon delivery and not paid by sender. &nbsp;this is obviously a safe thing to do if the balance is roughly the same, but once it isn't then it breaks down, and should go to both paying half for each bit flowing across C in either direction, and not just A paying more or B paying more.


that said, comcast's video would suggest there is no C here, only A and B, and for that scenario it doesn't make sense for them to pay upon receipt to each other or pay upon sending to each other. &nbsp;thus i don't understand how peering works. &nbsp;which is my point.


also, i counted the guy on the left saying "worked for over 10 years" or the like as at least 5 times.

OldPOTS
50%
50%
OldPOTS,
User Rank: Light Beer
12/5/2012 | 4:16:44 PM
re: Level 3: This Is Not a Peering Dispute


Let me give an extreme example;


If company A has only the thousands of home customers and company B has only the content providers. And there is a long haul/interconnect provider company C connecting them. Who pays for each companies (increase in)&nbsp; cost for BW for video streaming????


In today's networks both ISPs (A &amp; B) have roughly about the same number of customers, businesses and content suppliers to balance the load and consequently the cost and equal traffic between them over the interconnect.


But if one (B) adds a significant content (monopoly) supplier by adding a short BIG pipe $ (a short 4 lane highway) then the other (A) must expand much more of it's network with a lot of medium distribution pipes $$$$ (many 3 and 2 lane highways to interconnect to the distribution points towards the homes)


Then who pays (C) for the expansion of the middle network interconnect/long haul pipes $$$ ???? Before you answer, consider who is getting revenue!!! - Company A with many customers, or company B with revenue from/as content provider with a BIG pipe. Do they pay equally for interconnect/long haul??


Before you argue that the interconnect company (C) is 'the Internet', that may or may probably not be true. In many cases for high traffic volumes it is cheaper to pay for private interconnect lines than 'the Internet' connections (a business case). Then who pays for 'the Internet' and the BW expansion required in it to handle all the new video BW??


The previous agreements, like old POTS agreements, were based on an assumed balance. equal load types and traffic on both networks (A &amp; B). But withe only a few big video content/traffic providers there is an imbalance of load types and BW (highway) costs to the networks.


OP


&nbsp;

paolo.franzoi
50%
50%
paolo.franzoi,
User Rank: Light Beer
12/5/2012 | 4:16:42 PM
re: Level 3: This Is Not a Peering Dispute


pdecker,


There are Cs and they get paid a lot of money by small ISPs. &nbsp;In this case, there is no C because B is a CDN.


seven


edit:


tera,


Doesn't Netflix pay Level3 for bits to be delivered? &nbsp;Didn't Level3 cut of Cogent because it refused to pay?


&nbsp;

fgoldstein
50%
50%
fgoldstein,
User Rank: Light Sabre
12/5/2012 | 4:16:42 PM
re: Level 3: This Is Not a Peering Dispute


The question of how much traffic flows in each diretion, and how the FCC deals with this, is misguided.&nbsp; It is like arguing over how many angels are dancing on the head of a pin, and then asking the FCC, playing Pope, to setle the dispute.&nbsp; It's just wrong.


The key question is the nature of the Internet itself.&nbsp; Is it hte PSTN, where the FCC makes up a detailed set of rules (CABS) about how different carriers classify traffic and bill each other, subject to mandatory interconnection, or is it an information service -- computing -- that is left to market forces?&nbsp; While some argue the former, and Level 3 seems to be moving that way, one need merely think about what would have happened in 1993, when the commercial ISP business started up, if, before they could operate, ISPs had to file tariffs and wait for the FCC to create Rules for interconnection.&nbsp; That takes a while... the FCC opened a docket on VoIP in 1996 and it's still unsettled, and the overall question of intercarrier compensation was reopened in 2001 and is still totally unsettled.


The Internet privatized cleanly because there were no bottlenecks.&nbsp; Once NFSnet was gone, there was no dominant backbone provider, so a "big five" Tier 1 popped up, evolving over the years but still with no vendor having real market power.&nbsp; Level 3 is probably the big kahuna though.


Retail access was also not a bottleneck.&nbsp; Computer II required the telcos to permit ISPs to attack to their network, so in the early years, there were thousands of ISPs.&nbsp; In 1994, AOL wasn't even an ISP -- the Online Service Provider business was still bigger -- but they, MSN, CIS and others added Internet applications, mostly one by one (mail first, then netnews, tnen browsing) until by 1997 they were considered "ISPs" too.&nbsp; All free market, no legal status issues, no rules.


In 2005, the FCC overturned Computer II, flouting the Telecom Act, and ended open retail competition.&nbsp; So the duopolists now have market power.&nbsp; This leads to misguided calls for content regulation ("network neutrality").&nbsp; However, the cost of carrying video is actually really high, so it's understandable for ISPs to want to either block it, cap prix fixe usage, or get compensated.&nbsp; This is the market at work.&nbsp; Even if there were open competition for retail ISPs, this dynamic would play out.


So Comast is being villified for abusing market power, but they're probably behaving exactly as they would behave in an open market.&nbsp; Level 3 is playing with fire, asking for the FCC to regulate inter-ISP contracts.&nbsp; Do they really want more CABS?&nbsp; They've literally got billions of dollars in "agree to disagree" disputes with Verizon and AT&amp;T over VoIP termination that the FCC could technically throw either way.&nbsp; That's hardly a system to want to emulate.

OldPOTS
50%
50%
OldPOTS,
User Rank: Light Beer
12/5/2012 | 4:16:41 PM
re: Level 3: This Is Not a Peering Dispute


Tera's (Hi) comment hit it on the head;


"What if Level 3 (or any other 'B' or 'C') refuses to pay? Does Comcast cut them off? How would Comcast's customers, who think they are paying Comcast for access to all Internet content, react to this?"


My importamt question was, who got what revenue???


All must win, get paid appropriately, at some level to stay in business.


If A and B have to increase their sub price to cover the expansion of their network, where do they get the revenue?? From those customers that believe they are paying A for their connection, including internet, to content from B, and/or a site like Netflix? &nbsp; Letting Netflix and/or B off from paying the real cost? So does A have to contribute to B &amp; C and how much? Even pay for the entire network to C?


B has the marketing advantage because customers believe that 'the Internet' is free to the content. And that the site, B is only selling content. While B is also paying for link (short) to intermet. But does B also pay less to C than A pays to C??


OP


Happy busyhour on 'the Internet' after Christmas toys light up


Sorry for the confusion, but I tried to explain it from the generic physical perspective inorder to abstract it from this particular case. As fg pointed out the model has left the FCC without settled answers. Been there and done that allocation problem in a very large enterprise network. Hell every year as everyone see it from their perspective and think they should pay less.


In POTS the "busyhour" related/calculated directly to BW capacity from sub to sub. There were very limited "contrnt providers. The wirless network charges for minutes/BW used at both ends even thogh the customers don't prefer it, but accept it.



&nbsp;


pnni-1
50%
50%
pnni-1,
User Rank: Moderator
12/5/2012 | 4:16:41 PM
re: Level 3: This Is Not a Peering Dispute


If I was Comcast I would try and&nbsp;send all of my Intercontinental traffic over those under utilized egress links to L3.&nbsp; :)&nbsp;

Page 1 / 2   >   >>
Flash Poll
LRTV Custom TV
Distributed NFV-Based Business Services by RAD

7|18|14   |   5:38   |   (0) comments


With the ETSI-approved Distributed NFV PoC running in the background, RAD's CEO, Dror Bin, talks about why D-NFV makes compelling sense for service providers, and about the dollars and cents RAD is putting behind D-NFV.
LRTV Custom TV
MRV Accelerating Packet Optical Convergence

7|15|14   |   6:06   |   (0) comments


Giving you network insight to make your network smarter.
LRTV Custom TV
NFV-Enabled Ethernet for Generating New Revenues

7|15|14   |   5:49   |   (0) comments


Cyan's Planet Orchestrate allows service providers and their end-customers to activate software-based capabilities such as firewalls and encryption on top of existing Ethernet services in just minutes.
LRTV Custom TV
Symkloud NVF-Ready Video Transcoding, Big Data

7|9|14   |   3:41   |   (0) comments


Kontron and ISV partner Vantrix demonstrate high-performance video transcoding and data analytic solutions on same 2U standard platform that is ready for SDN and NFV deployments made by mobile, cable and cloud operators.
LRTV Huawei Video Resource Center
The Evolving Role of Hybrid Video for Competitive Success

7|4|14   |   4:09   |   (0) comments


At Huawei's Global Analysts Summit in Shenzhen, China, Steven C. Hawley from TV Strategies speaks to us about the evolving role of hybrid video for competitive success.
LRTV Huawei Video Resource Center
How CSPs Leverage Big Data in the Digital Economy

7|4|14   |   4:48   |   (2) comments


Justin van der Lande from Analysys Mason shares with us his views on how telecom operators can leverage customer asset monetization with big data. His discusses the current status of big data applications and the challenges and opportunities for telecom operators in the digital economy era.
LRTV Huawei Video Resource Center
Accelerator for Digital Business Future Oriented BSS

7|4|14   |   3:08   |   (0) comments


Mobile and internet are becoming intertwined; IT and CT are integrating; and leading CSPs have begun to transform to information service and entertainment providers. How should the BSS system evolve to enable this transformation? Karl Whitelock, an analyst at Frost & Sullivan, shares his views.
LRTV Huawei Video Resource Center
Orange Tunisia Discusses Multi-Band Antenna With EasyRET Solution

7|4|14   |   2:45   |   (0) comments


As new site acquisition becomes more difficult, Orange Tunisia has requested multi-band antenna to support UMTS and LTE innovation. Some things considered include reducing the cost of antenna maintenance and having high reliability antenna and EasyRET solution.
LRTV Huawei Video Resource Center
How Telefonica Spain Considers Antenna Selection for LTE Network Deployment

7|4|14   |   2:19   |   (0) comments


Tony Conlan, Global CTO of RAN, Telefonica, shares his opinion on antennas in LTE network deployment: Tower space is the premier requirement on antennas; reliability is important to reduce OPEX; and EasyRET solution will be helpful for antenna maintenance.
LRTV Huawei Video Resource Center
dtac Thailand: Multi-band Antenna & Capacity Solution for a Better MBB Experience

7|3|14   |   3:45   |   (0) comments


With the development of LTE, tower space and load are limited for new antenna, but users' capacity requirements are growing fast. To provide a better MBB experience, Panya Vechbanyongratana from dtac Thailand shared his experiences and antenna requirements.
LRTV Documentaries
BTE Panel: Network of the Future

7|2|14   |   1:00:57   |   (0) comments


Full-length video of the ATIS Panel Discussion: 'How Far Away Is the Network of the Future & What Does It Look Like?' from the Big Telecom Event (BTE) in Chicago.
LRTV Custom TV
Redknee Supports BH Telecom With Redknee Unified

7|2|14   |   6.14   |   (0) comments


Lucas Skoczkowski, CEO of Redknee, and Amir Orucevic, Director BH Mobile, discuss how the benefits of the Redknee Unified suite of products provide BH Telecom with innovation and leadership in the market, with the flexibility to launch services faster to the market, provide new and compelling promotions and pricing models, and combine services in order to drive ...
Upcoming Live Events!!
September 16, 2014, Santa Clara, CA
September 16, 2014, Santa Clara, CA
October 29, 2014, New York City
November 11, 2014, Atlanta, GA
December 9-10, 2014, Reykjavik, Iceland
Infographics
Allot's latest MobileTrends Charging Report shows that value-based pricing plans are up from 35% in 2011 to 85% in 2014.
Hot Topics
Microsoft to Axe 12,500 Ex-Nokia Employees
Sarah Reedy, Senior Editor, 7/17/2014
GM: 10 Car Models on Road With AT&T's LTE
Dan Jones, Mobile Editor, 7/18/2014
Have IBM & Apple Partnered Their Way to Cloud Leadership?
Mitch Wagner, West Coast Bureau Chief, Light Reading, 7/18/2014
The Municipal Menace?
Jason Meyers, Senior Editor, Utility Communications/IoT, 7/22/2014
FiOS: Slower Sub Growth, Faster Upload Speeds
Alan Breznick, Cable/Video Practice Leader, 7/22/2014
Like Us on Facebook
Twitter Feed