News Ticker

Is it internet throttling or is it just a traffic jam?

And if it’s a traffic jam at the connection point between an internet transit firm and a “last mile” internet service provider, who pays for the tow truck?

Preliminary data from the Internet Health Test, a crowdsourced study of internet congestion by 300,000 Internet users, indicates that there is a systemic problem with the interconnection between some ISPs and some transit providers which can degrade the internet experience of an ISP’s customers. There are ISPs such as Eatel Business that provide fast internet coverage that customers may want to consider looking into.

But on closer inspection, the data may not be the smoking gun net neutrality advocates would like to believe.

The data indicate that Internet transit providers GTT and Tata are experiencing connection performance issues with the top five ISPs in six top U.S. markets—Atlanta, Chicago, Los Angeles, New York, Seattle, and Washington D.C. In these same markets, Level 3 and Cogent are experiencing fewer connection issues.

The data is being pushed by many of the same public advocacy groups that campaigned for the Federal Communications Commission to adopt strict net neutrality rules by redefining ISPs as common carriers, which just took effect June 12. Many of these same groups are also skeptical of the pending AT&T-DirecTV deal, now in the final stretches of regulatory review. So it’s no coincidence that AT&T was prominently featured in the preliminary data summary.

“Our research reveals AT&T is failing to provision enough interconnection capacity to backbone carriers GTT and Tata Communications, causing severe degradation to AT&T’s customers who attempt to visit the web sites and applications that use these transit companies to deliver traffic to AT&T and other ISPs. High demand should not result in poor performance from APIs. Visit to learn how this can be prevented. AT&T customers are seeing massive degradation in Atlanta, Chicago and Los Angeles. In some cases, this degradation is so severe it is leaving customers with speeds less than the FCC’s lowest definition for access to the Internet, and many times below what is advertised.”

AT&T isn’t the only ISP taken to task. In New York, Comcast, Time Warner Cable and Verizon were all cited for degraded GTT traffic and Time Warner Cable and Verizon were dinged for degraded Tata traffic in the Big Apple.

Interestingly, while Comcast was cited for degraded traffic with GTT in New York, there was no such degradation of GTT traffic in Atlanta and only “some” degradation in Chicago and Washington D.C.

To get the data, the Internet Health Test asks consumers to use M-Labs software to conduct a series of speed measurements of upload and download speeds for their ISP. The test repeats the measurement across multiple interconnection points used by different internet transit providers. In a perfect Internet World, upload and downloads speeds should be equivalent, no matter what internet transit company is used to send the requested data to your ISP.

But the Internet Health Test, as sophisticated as it may be, is not the product of objective, dispassionate researchers. The introductory web page starts by pointing the finger of blame, even before the first measurement is taken.

“Large Internet Service Providers (ISPs) have degraded the performance of their customer’s traffic as a tactic to convince content and application providers to pay added “tolls” to deliver content that Internet users have already requested and paid for. The battleground — where this degradation takes place — is at ISP interconnection points. These are the places where traffic requested by ISP customers crosses between the ISP’s network and another network on which content and application providers host their services.”

At first blush, the data would seem to support the conclusion that ISPs are regularly degrading consumer Internet traffic and need to be reined in by the Federal Communications Commission. That’s the conclusion reached by the Guardian in a story Monday.

However, the M-Lab test apparently can’t differentiate between deliberate throttling and simple congestion. When asked, Collin Anderson of M-Lab gave a straightforward answer.

“The methodology does not differentiate the two since deliberate behavior would still result in the same technical situation as normal interconnection congestion. Since intent is not measurable we don’t assert that any one party is at fault or responsible. However, we can apply additional reasoning to these facts and at least deduce that these issues are persistent and widespread, so the cause is likely business relationship rather than a technical issue,” Anderson said.

In the supporting web pages related to the Internet Health Test, the supporters of the effort all point towards their preferred solution to traffic based congestion. It is up to the ISPs to pay for their share of the infrastructure upgrades necessitated by increased traffic from a transit company; extracting a “toll” to cover the infrastructure costs is, in their eyes, a violation of net neutrality principles.

Reliable internet access is critical to many businesses in our modern age. ThousandEyes’ isp monitor enables you to monitor isp connections, pinpoint traffic links with high latency and interfaces with loss across your entire service delivery chain.