Towards Neutral Ground on Net Neutrality, Simon Wilkie, Jonathan Aronson, Francois Bar. But Wilkie wants to talk about what's going on at Annenberg (USC).

He notes that on Feb. 10 the USC Annenberg Center hosted a closed conference on Net Neutrality. Included content providers, network operators, equipment manufacturers, academics, Wall Street analysts, public interest. They developed a broadly supported list of principles. Paper explains that process and proposal.

Long and extensive dialogue. Emerged with a "split consensus." Almost a consensus, with a few holdouts. More agreement than disagreement.

Annenberg Principles (see annenberg.edu). This is the first in a series of processes Annenberg wants to have/host.

First point of consensus: Net Neutrality is a horrible term. Too value laden and politicized. Need better terms. Ambiguity re what broadband meant. Speed, QoS? fixed and/or wireless. We're going to take 1.25Mbps to the home as the standard.

NN components: End to end principle -- any end user can access any other content provider or end user, attach any non interfering CPE, point of interconnection.

Nondiscrimination: two types -- customer level (based on speed, access, quality and quantity of service); transport level (in terms of interconnection, many types).

Four categories for transport level discrim:

DS1 -- "neutral" eg a posted tariff; any firm can attain a priority level by paying posted price.

DS2 -- "non-neutral" exclusive deal eg a preferred or affiliated content provider receives access to end users on exclusive terms

F1 foreclosure Type 1: network operator forecloses access to its subscribers from certain other content providers sites/users

F2 foreclosure Type2: network operator offers affiliated exclusive content only to its own subs.

In the discussion, even consumer advocates thought tiering okay -- charging more for higher capacity fine. Neutral transport price discrimination D1 may be efficiency enhancing. But discrimination of D2 and F1 is likely harmful. Discrimination F2 is ambiguous -- take a "wait and see" approach.

Modified end to end principle: we looked at what's available to consumers now. If there are deviations now, consumers shouldn't lose those if they are useful. Speeds of 1.25Mbps ubiquitous today. Allows delivery of VHS quality video. So let's modify the principle. They should have this 1.25 available in the marketplace. But be able to pay more for more.

Economic theory behind this: costs a lot to build a network. Vertical integration can be good. We've got mix and match components, so we can have multiple equilibria. Equilibrium does not imply efficiency. So bottom line is that we need better information on the market parameters.

Sunk costs: investments in network upgrades are sunk. Low marginal costs of moving bits. Competition in 'dumb' pipes' drives prices to marginal cost. Fixed costs not recouped -- therefore people won't invest. There's a need for some type of price discrimination.

Internet is really a two sided market. Needs an intermediary to make the market. Multiple sources of revenue for network providers increases incentive to invest. There's a hold up problem: if intermediary can extract rents then that discourages innovation at the edge (content providers and end users).

So - what to do? Market reality is a duopoly likely for the short term. Some potential entrants -- aws, sprint, clearwire, power line. But long legacy of entrant failures. Notes that incremental value of adding a MVPD customer is around $2000. Can't justify the investment, so has to be another revenue sources.

Case study 1 -- videotext market. Look at minitel v. Bildschrimtext (DT). Nascent information services. Technology developed in UK in 1970s. Minitel deployed by France Telecom in 1982. Bildschrimtext deployed by DT. Minitel still functioning, Bildschrimtext is dead. Minitel was open to any information providers. Neutral but not open. Non discriminatory access (like DS1). Created private information services as "kiosks". Huge initial success. Limited by use of only FT services. eclipsed by Internet. DT was closed -- controlled content, limited both innovationa and usage. Killed off in 2001. [pretty gross case study!]

Market falure - case study 2 Australian cable TV market, similar cultrue to US, similar suburban population. Duopoly policy. No program access rules. Each franchise gained exclusive access to movie studio libraries. Take rate about 20% compared with 85% in US. Implies significant welfare losses.

These are real issues. We're concerned about high prices because of lack of competition and we're also concerned about low quality.

Principles:

Both operators and customers should win.

Any regulation should be national and uniform and with a light touch.

There should be basic access to broadband -- meaningful, neutral internet connectivity. 1.25Mbps. Beyond this, all bets are off. Baseline should be reviewed every four years.

Customers should receive clear, understandable terms and conditions of service re preferred deals/content discrimination. FTC should enforce.

Sascha Meinrath and Victor Pickard paper on The New Network Neutrality is next.  Rough paraphrase of Victor's talk follows.

We're not just here to promote pro NN.  We're trying to begin a conversation that assumes importance of current argument but tries to think about next steps in order to be proactive.  People are paying attention, broad bipartisan coalitions are here -- we need a springboard, and we're trying to work on one.  NN is a battle worth fighting, but it's short-term.  NN isn't alone as an issue -- we still are way down in BB penetration.  And lots of undemocratic trends.  So we're broadening parameters to include a number of democratizing proposals.

A number of factors -- reframing of debate, complexity, shifting allegiances -- have separated democratic principles that should be considered together.  We need new historicism.  Need to look at last century and common carriage.

Paul Starr, Robert McChesney etc. show incumbents exploiting market advantages -- to hurt universal service and innovation.

So we'd like to go beyond econ and tech to normative principles about the role of the internet in society.  Need a renegotiated social contract with telecom providers.  We have ten principles/policy points.

[Sascha takes over]

So we think status quo is lame and we're opening up even more - that explodes the notion of what neutral network is.  We talk about historical precedents.  What we've done wrong we know -- so let's not do it again!  We look at common carriage.  While BrandX going on, we were saying NN was the big battle for 2006 -- and we were right -- but it was hard to get traction.  And now we think NN is just the first step.

So ensuring that network operators must lease their lines to other operators, explicitly allowing municipalities to be players, is critical.  We can't allow that kind of behavior.  AT&T breakup etc. sets same conditions that we faced in 1908.  AT&T destroyed many local carriers.  Open architecture and open source development is important -- a way to take control is to prevent access to hardware that might be perceived as open.  We noticed that 80% of wireless hardware owned by one company....

In the wireless realm in particular, we're moving away from open protocols and standards.  This has enormous implications for interconnectivity.  End to end architecture fits into that.  Maintaining an architecture that can allow for interconnectivity is important.  Vital to notice that conspiracy theories are often right! see NSA and CALEA.

He notes that Royal Post allowed for de facto "deep packet inspection," and our post was founded in reaction to that.  If you allow people to inspect packets, and require a mandate that that be done, you'll make legitimate packet inspection much more difficult.

We know that Windows Vista has DRM.  We know that allowing discrimination here will allow blocking of VoIP.  

We must remove notion that it's okay to remove the net bias of creating circumstances - conflation - of paying for premium services v. discrimination against those who do not pay for premium service.  These are two different things.  We need required minimums.

Interoperability is important. Systems need to talk to one another, and not be proprietary.  Too many actors are creating path-dependency.

Business-model neutrality.  Sascha is concerned about discriminating against municipalities.

Internationalization.  Also important to avoid American-centric management (notes ICANN).

Asks for feedback.

Craig McTaggart (of TELUS) paper is Was the Internet Ever Neutral?  Here's a rough paraphrase of his presentation.  All mistakes mine.

He asserts that the internet of today has many forms of discrimination and preferential arrangements -- caching, peering, filtering, and traffic management.  Points to Christian Sandvig saying that Lessig view is a myth.

Since the time that the public has been able to access the internet, it's been evolving in response to consumer demands, and we shouldn't stop now.  Internet's user population has changed dramatically since Arpanet etc. days.  It wasn't designed to support realtime bandwidth intensive applications.  In McTaggart's view (points to Ohm paper re myth of superuser), a lot of literature assumes that internet users are sophisticated, run their own routers.  But that's not true.  But those superusers have far too much importance in the policy discussion.  That's a problem for the mass market view.

Commercial ISPs have always served customer needs.  Meeting those needs may require changes to the internet's architecture.

Third: a neutral internet won't best serve consumer interests.  Mainstream users today have interests that are more important than historical architectural issues.  NN authors don't know better than users, and they keep patronizing users, who can act for themselves.  Consumers won't accept changes to the way their internet works.  But they'll love differentiated offerings.  So a non-neutral internet may better satisfy user demand. But we don't really know yet.

So -- while law professors are extolling virtues of the original architecture, tech community is long past this, and knows that architecture needs to change significantly.  End to end is a sacred cow.  Points to newarch, funded by DARPA, including Dave Clarke.  See "developing a next generation internet architecture" and GENI.  They want to create new core functionality, etc.

He says that NN proponents have an "end of history" way about them.

Noam asks for policy prescriptions.

Peha:  You're vulnerable to regulatory action if, when you have market power, you (eg) charge more for VOIP than for video.

Bauer:  I don't have an explicit welfare component to my paper.  My suggestion is that it would be a bigger mistake to adopt specific network neutrality rules and then decide we don't need them.  Congress should enable the FCC to adopt rules if needed. And don't prohibit municipalities to invest in broadband.

Lehr:  All versions of NN rules I've seen are bad.  We need general legislation (if we can craft it) that shows intent of what we're trying to do, and then have agency (without full discretion) enforce it.  Ex post alone is inadequate.  Trying to make the problem go away will help -- reform spectrum, don't block creative efforts by end-users.  Trying to design a framework for future is important.

Nachbar:  I'm in favor of bad legislation (joke!).  This NN problem is different because usually we don't mandate design in access fights.  That's the primary risk that we face here.  With a product like this, design is everything.  You can discriminate much more easily here than you could have with railroads.  So I'm more concerned about forms of discrimination here that are possible.  I think a strong non-discrimination norm enforced by standards would be a good idea.  And ex post is better.

Question:  old problems, new players, Noam has said.  Why don't we see differences in pricing here?

Lehr: We do need to see more price discimination.  We have a coordination problem here. 

Peha:  We have a choice from a tech perspective.  We can discriminate based on packet inspection etc. or you could throw capacity at the problem.  We did that in the 1990s.  Maybe we'll do that in the future.

Question:  What about consumer switching? won't people just migrate ISPs if there are discrimination issues? (from Martin from BT)

Lehr:  If you have competition, sure.  But switching costs can still be there.

Question:  What about just paying by bits or by service?

Peha:  That's technically possible, sure.  If the question is how to set the price, that's a hard problem.  That's what we're trying to figure out.

Question:  Assumption here is that pipe people can tax content providers. Cable systems show us that they have to pay more each year to content providers.  Why do you assume this is different?

Question:  In Netherlands, program providers PAY cable for access to the network.

Question:  What about looking toward the future -- barriers to entry for future entrants that won't have capital?

Question:  Market power is a necessary but not sufficient element for regulation.  What's the market definition?  And for content providers, they have a national market? shouldn't we look at national market from broadband access? (this from Yoo).

Question:  Innovation from IP is the most important -- not at platform level, which is like ISDN.

Question: Do we also look at market power for content providers? Also -- it's clear we don't like Telus-like discrimination.  And we don't like Madison River.  Both of those were acted on very quickly.  Largest driver for network infrastructure will require deep discrimination. (from Morgan person)

Panelists respond: 

Nachbar:  Market power is not at the root of a lot of things we regulate.  We could have competing police forces -- but we don't.  We've traditionally been more concerned about concentration in content.  Have to worry about nature of the product here.

Lehr:  Comcast owns a lot of content -- making a separation is hard.  Market power is one reason to justify, but not necessary for regulation.

Peva:  There is a terminating monopoly here, and I think that's the driving issue.  Existing antitrust rules well-designed for eBay, but not well-designed for last mile network provider. I'm not sure Telus would be in trouble in the US.  Biggest problem we have is duopoly for broadband access -- which is local.

Noam:  One day Google will ask for discriminatory pricing.

Interesting paper here from Jon Peha.  He's listing all the ways in which network discrimination can be helpful and harmful, and suggesting that both sides of the debate need to be addressed in any NN rules.  Very worth listening to.

Co-authored paper here from Bill Lehr, Marvin Sirbu, Jon Peha, and Sharon Gillett.  End users have tools too -- they shouldn't be viewed as passive in response to carrier discrimination.  But the arms race has costs that shouldn't be ignored.  And the issue remains complex and unresolved (good for those writing about NN).  Lehr asserts that the welfare implications of NN are ambiguous. 

I'm not so sure about that, because society needs diversity (requires it) in order to create greater wealth (monetary and otherwise).  Diversity above the network layer will be better for all of us. 

Thomas Nachbar is talking now about pre-New Deal public works cases.  Looking at lines of argument (eg, apartments in DC during WWI, political exigencies).  Necessity is a great justification (eg railroad).  But we often have things that we don't provide open access to - like food, and medicine, even though necessary.  If hold yourself out as providing a common resource, held to that (but that's mostly about civil rights).  How about market power?

Well, Nachbar points out that those who hold market power tend to be politically unpopular.  These cases are really rhetorical.  This isn't a traditional justification for open access regulation.  Wagons didn't have market power! 

Nachbar points out that all open access provisions have to do with networks.  There isn't an open access regime for bottled water.  Roads have always been subject to these requirements.  Always a special treatment of them.  How does recognizing this role help us be principled?

Nachbar gently suggests that we should look beyond market power as justification for open access.