A Workable Analogy for Assessing Net Neutrality

A Workable Analogy for Assessing Net Neutrality Rob Frieden Professor, Penn State University 102 Carnegie Building University Park, Pennsylvania 16802 (814) 863-7996; rmf5@psu.edu web site: The concept of network neutrality means different things to different people often leading to misunderstanding. The blog entries, news accounts and other comments I have seen evidence a discussion that does not compare apples to apples and oranges to oranges. For issues like this I look to a shared frame of reference through the use of analogies. They work for me even as most folks hate them: soup is to sandwich as horse is to carriage; ebony is to ivory as Stevie Wonder is to Paul McCartney—most analogies seem “Old Skool.” So which analogy, if any, works in the net neutrality debate? Net neutrality is to the Internet as common carriage is to telecommunications? Net Neutrality is to airline classes of service as Southwest airlines is to all coach class service? Air traffic control is to aviation as Net Neutrality is to Quality of Service? Net Neutrality is to Internet One as multiple Internets are to Internet Two? Perhaps airline/aviation analogies might work here. Advocates for a two-tiered Internet, comprised of plain vanilla, “best efforts” versus designer, “better than best efforts” have pointed to multiple classes of service provided by most airlines. However the first, business and economy class analogy falters somewhat, because all classes of service arrive at the same time. On the other hand as to the possibly critical first and last minutes of ingress and egress, at least some airlines attempt to provide premium customers preferential early access to the aircraft and first off the plane opportunities. Net neutrality involves quality of service, but the key measure 2 of quality involves speed of delivery and all classes of service pretty much equally arrive at the destination. Well then how about a distinction between the now abandoned Concord supersonic transport option versus plain old jet transport, or the difference in travel time between propeller aircraft and jets? Surely the Concord delivered a faster transit for the long haul, but absent opportunities to “jump the queue” the Concord had to slow down and join the destination approach line of planes regardless of top speed. Perhaps supersonic aircraft operators could negotiate a preferential runway access opportunity, just as “trusted travelers” may soon be able to pay for a speedier transit through airport security checkpoints. But for the time being air traffic controllers operate from an “open skies” neutrality principle that appears not to discriminate against foreign carriers, slower planes, or in which frequent flier program a particular air traffic controller has the most miles. Perhaps Network Neutrality advocates can use air traffic control as an example of why non discrimination makes sense even if a market could develop for buying priority access to runways. Do we really want to balkanize access to runways into a up to the minute “spot market” as opposed to a prepaid—perhaps even auctioned off to the highest bidder—landing slot? Are there public interest rationales for preventing multiple access pricing models in the Internet? In a recent blog entry Mark Cuban suggested that the Internet should have multiple tiers of service just as High Occupancy Vehicle and toll roads offer a faster transit time for a price. See We should acknowledge that already Internet access pricing falls into a dichotomy of carriers who pay other carriers for network access and those that do not. Small carriers and most content providers pay for network access through paid peering or transit. Only Tier-1 ISPs who 3 have the subscriber base, embedded network and traffic volume can successfully demand and receive “Sender Keep All,” “Bill and Keep” zero transfer payment access to other Tier-1 ISP networks. Might not this dichotomy justify charging customers for preferential and superior network delivery? Carriers can squeeze out additional payments for preferential network access if and only if they can abrogate and negate existing network peering and transit agreements. While these agreements typically are shrouded by non-disclosure agreements, even carriers such as SBC-at&t offer quasi-public general terms and conditions under which they will peer with other ISPs. See set out fully what interconnecting ISPs can expect from each other. A fundamental element in advertising these agreements includes the reasonable expectation that a carrier providing transit— routes in the vernacular—can and will deliver traffic on a best efforts basis. It reasonably follows that once having accepted payment from a transit paying ISP, the transit providing ISP cannot seek a second bite of the apple by hitting up customers of the transit paying ISP for an additional charge. Internet interconnection agreements typically are priced on a connection or capacity basis and not customer-specific, minutes of use terms typically used by telecommunications carriers when settling accounts for routing other carriers’ domestic and international traffic. Internet peering and transiting agreements already build in net neutrality. Demanding additional compensation from particular end users of a carrier already provided peering or having paid for transit would appear to violate the terms of the existing traffic routing agreement. I do not understand how a carrier providing transit, such as SBC-at&t or BellSouth, could single out some other carrier’s customer and demand additional compensation. I have seen no peering 4 or transit agreement that reserves to the carrier providing transport the option of singling out specific customers of another ISP for inferior or superior treatment. In the law vernacular BellSouth lacks privity of contract with Google unless and until Google contracts directly with BellSouth for transit or paid peering. Unless this occurs BellSouth has privity of contract with Google’s ISP and their existing peering and transit agreements apply. Practically speaking BellSouth would have something of value to offer Google if and when BellSouth could offer a superior and complete end-to-end routing “experience” for Google and its customers. To engineer a complete routing, of superior quality, for all Google traffic BellSouth would have to establish a globally ubiquitous network or secure superior service commitments from any and all networks over which Google traffic would traverse. While packet sniffing technologies have improved I am uncertain whether BellSouth or any ISP could guarantee such quality particularly when BellSouth does not provide all first and last network links from Google and to each Google customer. Until an ISP can deliver on its promise of a superior network connection, without violating existing contract terms preventing deliberate degradation of service, network neutrality stands as standard operating procedure.