x
Net Neutrality

No Heroes in the Battle for the Net

The FCC has stirred up a hornets' nest with its decision to monitor network interconnection agreements as part of the new Open Internet ruling. Not only do a lot of people have a lot of money at stake, but the issue of financial responsibility at these exchange points is extremely complicated and fraught with both technical and business model implications.

A new study* by the group BattlefortheNet (as reported by The Guardian) has produced findings that accuse major Internet providers -- including AT&T Inc. (NYSE: T), Time Warner Cable Inc. (NYSE: TWC) and Verizon Communications Inc. (NYSE: VZ) -- of degrading content delivery service from certain network exchange points to broadband subscribers. As one piece of evidence, BattlefortheNet points to the difference in performance of content delivered over Comcast Corp. (Nasdaq: CMCSA, CMCSK)'s network from a network provider called GTT Communications Inc. to the performance of content delivered over AT&T's network from the same company. Comcast provided a median download speed of 21.4 Mbit/s, while AT&T recorded a median download speed of only .2 Mbit/s.

On the surface, the evidence seems clear. AT&T is at fault.

But in reality, ISPs are now treading into territory where laws haven't yet reached.

Sure, AT&T is delivering lower performance than Comcast in this specific case, but is it AT&T's responsibility to make sure that performance stays at acceptable levels? Or has the state of content delivery changed so much that edge providers like GTT should shoulder more of the financial burden?

First off, we have no way of knowing what deals ISPs like Comcast and AT&T have reached with edge providers. And that's part of the problem. Did Comcast somehow wring more money out of GTT, and that's why the company is showing better performance in this one example? Or did Comcast invest its own money here, while making up the difference by collecting fees somewhere else in its network?

Secondly, no independent party has effectively made the case for what's "just and reasonable" in today's interconnection agreements. The nature of Internet traffic has evolved. Perhaps the rules of interconnection should as well.

At a high level, the issue comes down to this: Should ISPs have to allow anyone and everyone to connect to their last-mile networks for free? And should they be responsible for maintaining the necessary capacity to sustain acceptable performance rates?

The argument in favor: Yes, this is what consumers are paying for -- access to content from across the entire Internet.

The argument against: No, edge networks are dumping far more traffic on to last-mile networks than last-mile networks are pushing back in return. Peering should only be free when both sides are passing on a roughly equivalent amount of traffic.

Most individual companies (minus Netflix Inc. (Nasdaq: NFLX)) don't generate enough traffic to make much of a difference to ISPs, but CDNs and other edge providers are a different matter. They carry a huge amount of the Internet's traffic, which is why there have been several public disputes between companies like Cogent Communications Holdings Inc. (Nasdaq: CCOI) and Level 3 Communications Inc. (NYSE: LVLT) with large ISPs like Comcast and Verizon. (See Interconnect Deals Bear Net Neutrality's Stamp.)

There's also a difference between interconnection agreements driven by the need to reach only subscribers directly on an ISP's last-mile network versus those being driven by the need to reach additional networks linked to the ISP at other exchange points. In other words, some interconnection agreements include transit to other networks, while some deal solely with the exchange of traffic on to an ISP's last-mile routes. (See First Net Neutrality Complaint Hits TWC.)

Confused yet? You should be because it's a confusing issue.

The question is, is this something the Federal Communications Commission (FCC) can sort out? It's already said it's planning to try, but the agency may not be aware of just how difficult the process is likely to be. Even if the FCC can get all of the data it needs, there will be a doozy of a battle over what it should be allowed to regulate. (See Hey, FCC: Cedexis Has That Peering Data You Need and FCC Vote Shows Net Neutrality Strains.)

Net neutrality fights? We're just getting started.

*Note: It appears the data cited by The Guardian was produced by using the online Internet Health Test, which in turn was developed with the help of a Network Diagnostic Tool created by M-Labs. It's not clear that any actual study offering the findings from this data was released.

— Mari Silbey, Senior Editor, Cable/Video, Light Reading

VicePres27474 6/30/2015 | 9:57:10 AM
2 News Items of Interest GTT and AT&T sign new interconnect agreement: http://www.marketwatch.com/story/gtt-and-att-enter-into-an-interconnection-agreement-2015-06-30

New Streaming Media blog post: http://blog.streamingmedia.com/2015/06/isps-not-causing-network-slowdowns.html
R Clark 6/25/2015 | 2:46:13 AM
Re: New doubts about The Guardian's story Really no surprise here (regardless of the accuracy of the Grauniad piece). Net neutrality is complicated and really does require some sustained network monitoring and reporting.  Personally I've always wondiered if it would be worth it - just how much monitoring in how many network points would provide enough awareness about the totality of a network?  To be continued.
mhhf1ve 6/24/2015 | 3:53:41 PM
Transparent Peering agreements... I'm wondering if any economists have an opinion on how this market would be affected if all peering agreements were transparent to the govt. I'm sure there are a lot of deals made behind closed doors that might not look so great if they were made public... but it might not be necessary to make peering agreements public -- as long as some agency was able to reasonably decide what was fair..?

 
msilbey 6/24/2015 | 3:51:56 PM
New doubts about The Guardian's story Analyst Dan Rayburn is now reporting that the study The Guardian references didn't really exist. Instead, it appears someone pulled data using M-Labs' Network Diagnostic Tool and cited that data as a study.

See: http://blog.streamingmedia.com/2015/06/the-guardians-story-is-bogus.html

And see the latest M-Labs blog post: http://www.measurementlab.net/blog/interconnection_and_measurement_update

Regardless of the data, the issues raised are still relevant, but it would be nice not to have this debate muddled with more inaccuracies. 
HOME
Sign In
SEARCH
CLOSE
MORE
CLOSE