Home / Blogs

No Free Lunch in Internet Peering or Transit

Like many of you, I am keenly following the Comcast-Level 3 dispute and am trying to make sense of it all.  The dispute confirms several universal principles about Internet traffic routing that have passed the test of time:

1)  Consumers pay Internet Service Providers (“ISPs”) a monthly subscription with the expectation that the fee covers access to available content, i.e., the conduit. As the World Wide Web evolves and content options diversify to include full motion video, consumers simply expect their ISPs to make sure the download distribution pipes are sufficiently robust to handle high bandwidth requirements and commensurately large monthly download volume. Cable modem service agreements may have a cap on downloading per month, but consumers generally assume “All You Can Eat” access rights, plus the expectation that video streaming will work, i.e., no blurring, frozen frames, or blue screens.

2)  Because upstream requests for content are narrowband and because the typical consumer downloads much more content than he or she uploads, ISPs serving end users, such as Comcast, typically will have a large traffic imbalance with more downstream traffic to deliver than upstream traffic that the end user serving ISP might want other ISPs, such as Level 3, to handle.

3)  Until such time as Comcast’s “Television Anywhere” takes off and generates lots more traffic that Comcast will need other ISPs to handle-whether on a peering or transit basis-Level 3 vastly contributes to Comcast’s download “surplus” delivery burden to end users. Of course Level 3 replaces another content distribution network so the total volume of Comcast’s downloading burden does not change in the short term. However, in the context of peering and transit between Comcast and Level 3, the traffic volume relationship changes with a greater imbalance resulting from the new Netflix traffic Level 3 now delivers to Comcast.

4)  The Comcast- Level 3 dispute distills to a disagreement over whether and how much either should pay in light of changed traffic patterns. Because the parties already have traffic agreements, modification of terms might require additional payments from Level 3 to Comcast, absent Comcast’s need for Level 3’s upstream transmission services. Of course Comcast does need the services of Tier 1 ISPs like Level 3, but until Comcast starts distributing lots more of its cable television video product over the Web, Netflix downloading to Comcast subscribers will predominate.

5)  Cooperative ISPs typically align inbound and outbound peering traffic with an eye toward creating a balance, but either or both ISPs might also want to expand transiting services as these paid arrangements are based on the unlikelihood of balanced traffic loads.  Digital Society Policy Director George Ou reports that Comcast and Level 3 have both peering and transit agreements. George lays blame on Level 3 for expecting Comcast to absorb the newly increased volume of traffic delivered to it by Level 3 without additional payment by Level 3, or the offer of additional free upstream capacity.

Reasonable people can disagree as to the mutual exclusivity or substitutability of peering versus transit. George considers the two types of traffic arrangements mutually exclusive and has chided me for thinking that the parties could recalibrate both to mitigate the traffic imbalance if they wanted to (read more).

The Comcast- Level 3 dispute confirms that there is no such thing as a free lunch. It also highlights disagreement over who has to pay when consumers’ download requirements increase with full motion video access. George considers it a nonstarter for Comcast to raise end users cable modem rates, despite a vast increase in the value proposition created by IPTV.  Some economists consider it a given that Comcast has the “right” to demand compensation from both sides of its market position, upstream from Level 3-and possibly the real instigators of greater bandwidth requirements Netflix and Google-and also downstream from end users, i.e., cable modem subscribers, co-conspirators with Netflix and Google.

Bottom line: one or more players in the Internet “network of networks” will have to pay for greater capacity. Early on in the Internet’s development, avoiding payment strategies were depicted as “hot potato routing.” Carriers unwilling to upgrade facilities to accommodate greater demand sought to hand off traffic as soon as possible. Level 3 has no such option of passing the packets off to several different carriers for the last mile to end users. Comcast knows this and true to form the company exploits its position to the fullest extent possible.

By Rob Frieden, Pioneers Chair and Professor of Telecommunications and Law

Filed Under


ISP vs CDN Dan Campbell  –  Dec 11, 2010 3:30 PM

Peering has traditionally been between large ISPs who exchange routes and deliver traffic to each other, usually of roughly equal “size” (and thus no need for transit fees), but more importantly without really being the exact source that is generating the traffic.  They are merely large intersecting highways rather than the shopping mall, stadium, housing development or other source/sink of traffic.  When Level 3 became the CDN for Netflix, they effectively became Netflix from the point of view of who is hosting the traffic source. They became the mall, the stadium.  And those who host content whether on their premise or at a hosting center tend to have to pay for Internet transit service in accordance with the volume of traffic they generate.

Comment Title:

  Notify me of follow-up comments

We encourage you to post comments and engage in discussions that advance this post through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can report it using the link at the end of each comment. Views expressed in the comments do not represent those of CircleID. For more information on our comment policy, see Codes of Conduct.

CircleID Newsletter The Weekly Wrap

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

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

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



Threat Intelligence

Sponsored byWhoisXML API


Sponsored byDNIB.com

IPv4 Markets

Sponsored byIPv4.Global

Brand Protection

Sponsored byCSC

New TLDs

Sponsored byRadix

Domain Names

Sponsored byVerisign


Sponsored byVerisign