This is a guest post. The views expressed in this article are solely those of the blogger and do not represent positions of IEEE Spectrum or the IEEE.
The U.S. Federal Communications Commission's proposal to roll back the previous administration's Open Internet Order has put network neutrality back in the news. The FCC's new order, titled Restoring Internet Freedom, removes the bright-line rules enforced by the previous order. If the FCC passes its new order tomorrow, Internet service providers (ISPs) will no longer be prohibited from blocking, throttling, and prioritizing traffic.
A widely expressed concern about the FCC's new proposal is that permitting ISPs to create fast lanes—in other words, letting ISPs charge content providers for delivering their traffic to users at a certain speed or quality— will jeopardize long-term innovation. While large content providers such as Google, Facebook, and Netflix can afford to pay ISPs for that service, a new company probably won't be able to. Allowing ISPs to throttle traffic if content providers don't pay up will make it more difficult for startups to compete with large companies.
However, an aspect of network operations that is often ignored in the popular network neutrality debate is this: It is, in fact, desirable that ISPs not always be neutral in handling network traffic.
As an example, consider two users whose Internet traffic goes through the same congested link. If one user is streaming video and another is backing up data to the cloud, a perfectly neutral network would slow down both transfers. Most people would probably agree that, to create the best experience for the most users, it would be best to slightly slow down noninteractive traffic such as data backups, and free up bandwidth for videos and voice-over-IP calls.
Both the Open Internet Order from 2015 as well as the proposed order on Restoring Internet Freedom recognize the need for ISPs to manage their networks. The difference lies in how the two orders account for network management techniques.
The Obama-era FCC administration required that ISPs be prepared to present evidence—for example, in the form of performance measurements of their networks—to prove that they are managing their networks in a reasonable way. In contrast, Trump's FCC administration believes that this regulation places an undue burden on ISPs.
Instead, the FCC's new proposal merely requires ISPs to be transparent and publicly disclose how they manage their networks, so that customers can choose to take their business elsewhere if they are put off by their ISP's practices.
The challenge in placing the onus on ISPs to be transparent is that external observers have no good way to verify an ISP's claims. For example, it would be reasonable to expect that an ISP would resort to throttling some types of traffic only when its network is congested. When there is plentiful bandwidth to spare, there's no reason to throttle legitimate traffic.
However, if an ISP violates this expectation, detecting such a violation would be hard. The research community has developed increasingly sophisticated measurement techniques over the years to make inferences about the Internet, but it's still difficult to accurately identify when a network link is congested.
Thus, we must answer a key question in order to resolve the debate on network neutrality: How can we legally define the permissible ways an ISP could throttle or prioritize traffic in a manner that does not place undue burden on ISPs, yet is verifiable by third parties?
Only then will we be assured of having an Internet that is neutral when it can afford to be and is nonneutral only to the extent it's necessary to improve a user's experience.
Harsha Madhyastha is an associate professor of computer science and engineering at the University of Michigan, where he focuses on networking, distributed systems, and security and privacy. He has developed software and apps to reduce latency, detect social malware, and measure performance across networks. He has received funding from the National Science Foundation, Google, and Facebook for research on topics unrelated to network neutrality.