Oh, that’s fascinating. Do I correctly understand them to be saying that they found nothing to support the idea that backbone companies would need hardware upgrades to support traffic from high-bandwidth sites such as Netflix? I think I remember that claim being advanced as a rationale against net neutrality rules, but correct me if I’ve misremembered. Or does this result not bear on that point?
The paper seems to hint that those kinds of questions are part of the motivation for the study: the increasing frequency of peering disputes begs the need for better information about congestion in general, and specifically about traffic utilization at interconnection.
But… it’s very cautious about answering them. The only real conclusion seems to be that the vast majority of interconnects have significant spare capacity, even at peak. That’s interesting new information at least, and as I read it does at least suggest your conclusion, that Netflix-style traffic is probably not causing significant bottleneck problems at the interconnects.
Oh, that’s fascinating. Do I correctly understand them to be saying that they found nothing to support the idea that backbone companies would need hardware upgrades to support traffic from high-bandwidth sites such as Netflix? I think I remember that claim being advanced as a rationale against net neutrality rules, but correct me if I’ve misremembered. Or does this result not bear on that point?
The paper seems to hint that those kinds of questions are part of the motivation for the study: the increasing frequency of peering disputes begs the need for better information about congestion in general, and specifically about traffic utilization at interconnection.
But… it’s very cautious about answering them. The only real conclusion seems to be that the vast majority of interconnects have significant spare capacity, even at peak. That’s interesting new information at least, and as I read it does at least suggest your conclusion, that Netflix-style traffic is probably not causing significant bottleneck problems at the interconnects.
Thanks for the summary. I certainly see that they have to be cautious about only making claims they can support.