Net neutrality issues have been making headlines for almost a year now, and this February we have the satisfaction of seeing some action. On February 4th the FCC released a Fact Sheet outlining the content of a proposal to invoke Title II authority over Internet service providers (ISPs) and establish net neutrality rules. On February 26th, the commission will vote on the proposal and is expected to approve — no suspense there. The suspense is in waiting to see exactly how the ISPs will frame their bids to invalidate the ruling.
In the meantime, the rest of us have a chance to catch up. In a previous post about the way Internet traffic travels through the pipes, we saw that packets are queued for routing. Now we’ll take a closer look at the specific queuing practices that led the FCC to establish net neutrality rules.
Internet traffic management
The ISPs who own and operate the pipes invest heavily in devices and software to support smooth traffic flow. They use routers (and virtual routers) that run highly complex, specialized software to manage the traffic intelligently. As part of the process, packets are queued for optimal transmission. Equal treatment of all Internet traffic would require the traffic management software to assign queues neutrally.
Routers (and virtual routers) run highly complex, specialized software to manage Internet traffic intelligently.
In a perfectly neutral system, packet queuing would occur on a first in, first out (FIFO) basis. However, FIFO routing cannot handle uneven or heavy traffic flow; it’s not sufficiently intelligent to keep up with real Internet traffic.
In a perfectly non-neutral system, queuing could occur in any fashion the ISPs desired. To give one possible setup, a current patent describes a system in which packets are queued in three classes: premium, assured, and regular. Packets in the premium queue jump ahead of those in assured and regular queues. ISPs would charge a premium price to businesses that want their packets in the premium queue. Additionally, an ISP would route its own packets in the premium queue, for free.
Equal treatment of all Internet traffic would require the traffic management software to assign queues neutrally.
The popular press refers to this kind of class-based queuing with the metaphor of “fast lanes“; we can see there are no actual fast lanes, but rather premium queues. The FCC refers to the practice as paid prioritization and techies use the phrase class-of-service differentiation. Call it what you will, the FCC is in the process of banning it.
Inside the real Internet pipes there is not one perfect queuing system. The FCC can’t pursue neutrality by specifying a software program for use by everyone. Routing programs are heterogeneous, depending on many factors including pipe type, capacity, and location. As Internet users we all benefit from engineered optimization of packet queuing and switching.
Still, the net neutrality principle places a limit on optimization by banning traffic management systems with money on their minds. If all goes as planned on February 26th, the FCC will ensure ISPs are prohibited from offering paid prioritization.