What the author misses here is that Internet bandwidth is not infinite, and networking bandwidth is not free. The Internet is a network of networks, each one called an Autonomous System (AS). The goal of AS’s is generally to connect themselves to other AS’s in a way to be able to communicate with the rest of the Internet. To do this, they MUST build connectivity to other networks. This means, they must PAY. There is no such thing as a free packet.
There are two types of relationships between AS’s. One is “peering” where both sides feel that the optimal financial relationship is to pay for their own connectivity (leased lines, routers, etc.) to the other AS, and not to bother trying to measure and make each side pay for the actual bandwidth used. The other is “transit”, where one AS acts as a customer and takes services, of what is called its provider, to get the connectivity to the rest of the Internet, where the customer AS pays for the bandwidth used as well as the connectivity required. Historically, “peering” was established between AS’s with fairly symetrically balanced traffic flows between them, i.e. the same amount of traffic goes in to an AS as goes out of it.
“Peering disputes” have always existed on the Internet. The earliest documented dispute was in 1994 when DANTE asked EUnet to build more of its own connectivity because it was using too much bandwidth on DANTE, with a veiled threat to cut them off otherwise. In 1997, UUNet decided that several of its peers were actually “leaching” traffic from it, and de-peered them. Researchers have documented at least 26 publicly known peering disputes (https://arxiv.org/pdf/1409.6526.pdf).
An interesting inversion has occured from the early days of the Internet. When “information was free”, the economic benefit was to an AS that took more traffic than it sent (“leaching”), and made its users pay for access to the free information.
Now that services like Netflix can reliably charge fees for information, and given the massive bandwidth of video, the situation has reversed. Now it is the providers of information who receive economic benefit from pushing their packets to other AS’s and collecting monthly subscription fees.
The Comcast/Cogent-Netflix dispute shows the problem. Should Comcast have to pay to provide the networking for Netflix to make money (at the expense of Comcast subscribers, many of which who do not subscribe to Netflix)? Comcast said no, Netflix had to step up to provide connectivity to reduce the network congestion. Netflix did that. Problem solved, like the other peering disputes in Internet history. (BTW, part of the Netflix solution now is placing content caches inside end-user ISP networks, reducing inter-AS traffic.)
There is a very different theoretical problem — if an end-user ISP decides to impede a content distributor competitor on the basis of competition for content distribution itself. Unfortunately, it would be hard to tell in the short-term externally if this is the case or it is just a proper rebalancing of inter-AS bandwidth economics. But the fact that the Comcast/Cogent-Netflix dispute was solved, and that all major end-user ISPs in the US continue to raise their bandwidth on the Netflix ISP Speed Index (to an all-time high, above 3 Mbps now), suggests that end-user ISPs feel that they can not trade off improving their content distribution business by permanently reducing Internet delivery quality.
At many points in the history of the Internet, activists have fought for Network Neutrality from Government, such as with the fight against the Communications Decency Act. We have to recognize that additional government regulatory rules on Internet traffic may lead to uknown worse problems that we can’t even imagine now. If “Network Neutrality” was trying to solve actual problems, that would be one thing, but these “problems” are completely illusory — the issues we have seen come from the very nature of the decentralized nature of the Internet using the free markets to self-organize the exchange of bandwidth into the amazing emergent behavior that we see today.