EU telcos defend UN Internet takeover plans
A group of European telcos is defending a controversial proposal to radically alter the architecture of the Web and its governance. But the rhetoric has done little to slow critics.
A trade association of 41 European telephone companies responded last week to mounting concerns over its controversial proposal to turn Internet traffic management over to the International Telecommunications Union, a regulatory body of the United Nations.
The European Telecommunications Network Operators Association, or ETNO, made its proposal in June, part of a known as the International Telecommunications Regulations. Changes to the treaty, which has not been revised since 1998, will be finalized later this year in Dubai at the World Conference on International Telecommunications.
Only national governments can vote on the new treaty. But non-governmental organizations like ETNO, who pay the ITU upwards of $70,000 annually (PDF), can participate in WCIT, including proposing changes to the treaty.
ETNO's proposal, WCITLeaks Web site, would dramatically alter the longstanding practice of settlement-free peering arrangements between major Internet backbones.following its disclosure on the
Under settlement-free peering, networks do not meter or charge each other for exchanged traffic. According to the Internet Society (PDF), settlement-free peering remains the most common type of interconnection among networks.
Instead, ETNO's proposal would require the governments of 193 signatory countries to establish and enforce a new "sending party network pays" (SPNP) model for Internet traffic. That approach would effectively tax high-volume Internet content providers -- most of which are non-EU companies -- to respond to information requests from local customers.
Such a system would be similar to the outdated and often corrupt practice of national telephone companies setting rates for incoming international long distance charges.
Why change one of the most successful features of the Internet's architecture?
ETNO operates in a tightly-regulated European Union market, where consumer pricing and network unbundling decisions are largely out of their hands. EU carriers are already finding it difficult to make money providing Internet access, or to make a business case for needed infrastructure investments.
The organization's members, including Telecom Italia and Deutsche Telekom, readily admit they must find new forms of revenue to fund the expensive transition to all-IP networks for voice, data, and video. ETNO's only hope, it now seems, is to conscript the UN to mandate a new system that would charge content providers to deliver network traffic requested by ETNO's own customers.
ETNO's defense grows increasingly incoherent
ETNO's new report (PDF) fails to explain why the solution to its members' financial woes shouldn't begin with EU regulatory reform.
Nor does it respond specifically to mounting criticism from a wide range of stakeholders, including the Internet Society, the parent organization for the engineering task forces that maintain Internet standards and protocols worldwide.
In June, ISOC issued a detailed technical analysis (PDF) that rejected the SPNP model as putting "at risk the diversity and competition of the communications services marketplace that gave rise to the commercial Internet in the first place."
That's because accounting for SPNP would be difficult if not impossible. For one thing, packets in the same transmission may be sent along different routings. From an engineering standpoint, according to ISOC, SPNP would "jeopardize the distributed nature of the current Internet infrastructure by reducing choices over how and where traffic is sent."
Worse, as ISOC points out, implementing the ETNO proposal would likely discourage content providers from responding to information requests from users in developing nations, who may not generate sufficient revenue through ads and other sources to offset the new tax.
This is a particular danger with government-owned communications providers, who may demand confiscatory rates to reach their citizens. Users in many developing countries could find themselves effectively cut off from the Web and other Internet services.
ETNO's defense of its proposal raises more questions than it answers. For one thing, it does not in any way address ISOC's critique of the technical difficulty of retrofitting a SPNP settlement regime onto the Internet's packet-switched architecture. Indeed, ETNO's report devotes less than a page to the SPNP portion of its overall proposal, even though the idea of a tax on content has proven to be the group's most controversial suggestion.
Instead, ETNO simply reiterates its view that "Perpetuating an 'unpaid peering' approach for IP interconnection that developed when traffic patterns were largely symmetric can hamper the incentive to invest in transport capacity and network quality." The group argues that over-the-top video providers including YouTube, Netflix, and Hulu are extracting more and more revenue from the network but "are not contributing to network investment."
The bulk of the ETNO report is more concerned with denying that its proposals amount to new regulations in the first place. "ETNO is not asking for increased governmental interventions," the report insists. Rather, the trade group repeats that it merely wants the treaty to establish a new "reference for commercial negotiations" through new "principles" that ETNO "wishes to see introduced" into the treaty.
In fact, ETNO now implausibly claims that its proposal is simply aimed at ensuring the revised treaty prohibits member nations from blocking private interconnection agreements based on SPNP. As the report concludes, "ETNO, through its proposal fosters, commercial agreements between parties on a voluntary basis, therefore, the sending party principle would not be imposed but commercially negotiated." (sic)
U.S. and other governments increasingly skeptical
But the actual text of the proposed changes betrays that gloss. ETNO's amendments clearly state that member nations "shall ensure" that "operating agencies" (including private network operators) establish SPNP to replace existing settlement-free peering. And the commercial agreements ETNO has in mind are hardly voluntary:
"[T]o ensure an adequate return on investment in high bandwidth infrastructure, operating agencies shall negotiate commercial agreements to achieve a sustainable system of fair compensation for telecommunications services and, where appropriate, respecting the principle of sending party network pays." (emphasis added)
The difference between voluntary "principles" and mandatory treaty obligations is more than mere semantics. Amending the treaty as ETNO proposes would require signatory nations ("shall ensure") to mandate SPNP agreements for its own carriers.
It would also oblige local carriers to negotiate similar agreements with foreign networks that send IP traffic to its end-users. U.S. carriers would be forced to negotiate these agreements, even if the U.S. rejected a revised treaty. An Internet traffic trade war could result.
ETNO has been repeatedly called out for misrepresenting the true nature of its proposal. At the annual Technology Policy Institute summit in Aspen this summer, for example, former U.S. Ambassador David Gross and others that its proposals would not mandate any new role for the ITU or member nations in enforcing an SPNP regime.
In response, ETNO's Chairman, Luigi Gambardella, suggested that since English was not the first language of his organization's members, it was possible the proposal had not been perfectly drafted.
But when asked if the repeated use of the word "shall" was an error, and whether ETNO was willing to change it to "may," Gambardella demurred. He later"We're open to improve our proposal. We have asked our U.S. friends: 'Let's work on it, let's improve it. If you feel that we can express it in a better way, let's do that.'"
The U.S. delegation, however, is unlikely to compromise. Despite an already-bitter election season, hostility to any expansion of the ITU's role in Internet governance is perhaps the most bi-partisan idea to hit Washington in years.
In June, for example, the U.S. House of Representatives unanimously passed a joint resolution (PDF) reiterating the "unequivocal policy of the United States to promote a global Internet free from government control and preserve and advance the successful multi-stakeholder model that governs the Internet today."
And last week, the Senate Foreign Relations Committee voted to send the resolution for a full vote in the Senate. Sen. John Kerry (D-Mass.), chairman of the Committee, indirectly condemned the ETNO proposal. "We're facing off against those who want the world's approval," he said, "to balkanize the Internet into countries where governments can censure speech or impose new taxes on the transmission of information across national boundaries."
The White House, the State Department, and Republican and Democratic commissioners of the FCC have all voiced similar concerns.
ETNO is now facing growing dissent from representatives of developing nations, who are suspicious of the organization's self-serving claim that new settlement regimes can "offer additional sources of revenue necessary to finance their infrastructure." Similar justifications were long advanced for charging discriminatory rates for incoming international long distance calls, but little of the collected funds were actually used to improve network facilities.
Despite repeated promises to improve its proposed amendments, ETNO has so far stood firm in its original wording. In a letter provided to CNET by ETNO, Gambardella last week told ISOC's European Council that after considering ISOC's objections, "ETNO is all the more convinced of its position and will continue to defend it in the public discussions on ITRs revision, as well as in the appropriate institutional and public fora."
The letter made scarce mention of the SPNP proposal, other than to repeat that ETNO's proposal merely states a new "founding principle for IP interconnection."
But the actual text still says otherwise.