IETF Standards Could Mitigate Net Neutrality Fights, Experts Tell Group

Handley said developers had to fix the problem by giving ISPs effective tools to make economics work properly. “Otherwise we will end up either with ubiquitous deep packet inspection or bad regulation,” and both would lead to stagnation of innovation, he said. Handley said DPI, already used by some providers in the U.K., would become even more prominent once IPTV takes off because of additional congestion in the networks.

In its response to congestion, the IETF has been “designing stuff to not treat all packets equally,” Handley said. “Giving low latency [shorter delay] to packets using DPI is deeply flawed,” he warned. Not only has DPI created a problem with privacy, but it also set in motion an arms race between ISPs wanting to slow down P2P traffic and P2P developers making the technology “more port-agile, encrypted and stealthy.” DPI designers in turn could start prioritizing known “friendly” traffic only, stifling innovation, he said.

The IETF should consider “user choice [to be] a design principle,” since developments in Internet technology have shifted power to network providers and away from users, said Barbara van Schewick, a net neutrality expert at the Stanford University Law School. She and Handley pointed to alternative developments already started in the IETF. Van Schewick said the recent work on “multipath TCP,” which would shift traffic away from congested routes on the Internet, was an idea that might be promising. Handley applauded work on a tool in development that would make congestion points visible from a packet’s origin to its destination, which would give incentives for “sane economics,” he said.

A more radical way to ensure net neutrality might be mandating encryption or making obfuscation and randomization a design feature of IETF protocols, Handley said. This would disable DPI “middle boxes” designed to analyze data packets. Mandatory use of IPSec could be forthcoming under a new standard that is about to be finished on IPv6 residential gateways, said Apple engineer James Woodyatt: “Do we really want to disable all DPI?” Leslie Daigle, chief technology engineer for the Internet Society, said “it is important that the technical standardization stays focused on building specifications that are about structure and transmitting of packets for a global network that supports innovation and development and deployment of new applications.” But she questioned whether neutrality had been always enshrined in the network protocols. Some links in the network are better than others, and routers would make choices about which path a given packet takes, Daigle said.

Experts must “educate regulators and policy makers about the implementation of heavy-handed, rigid regulation that focuses on current network technologies,” Daigle said. “Forcing neutrality could be as detrimental as promoting bias,” and “would require rewriting the policy every six months.” It’s important to make regulators aware that “the definition of ‘good’ and bad’ behavior lies outside the technical realm and in the land of appropriate competition and fairness,” Daigle said.

Ted Hardie, director of Internet and wireless for R&D at Qualcomm and member of the Internet Society board of trustees, pointed to an earlier discussion on the demand from law enforcement to standardize wiretapping technologies. In the same way the IETF had resisted efforts to develop standards for government control, it should also not empower people who “pass the packets” over those who “initiate communication flows,” he said. The architecture board would consider further work on the net neutrality debate, Chairman Olaf Kolkman said after the debate.