Interoperability as a Code

advertisement
Interoperability as a Standardbased ICT Competition Remedy
Dr Ian Brown (Oxford) and
Prof. Dr Chris Marsden (Sussex)
IEEE SIIT 24 September 2013
Standards-based interoperability
framework

We assess regulatory intervention
◦ according to the code solutions used.

Standards-based solutions involve both
◦ competition analysis and
◦ interoperability requirements
◦ in strategic communications sectors.

We conclude that such standards
frameworks are urgently needed
◦ to enable citizens to make most effective use of
the opportunities offered by ICTs
We analyse the regulatory shaping
of “code” standards
Technological environment of Internet
comprising hardware, software and their
interactions,
 notably in the protocols and standards used
to achieve interoperability
 to achieve more economically efficient and
socially just regulation.
 Acknowledging that “openness” in standards
is controversial


◦ Kretchmer (2008) Open Standards Requirements
History of competition policy for
open technology standards
Long predates the Internet (Kahin and Abbate 1995)
 Governments see success of open standards as

◦ solutions for the well-known entrenchment of
◦ dominant Internet commercial actors using network effects
◦ (Pitofsky 1998; Lemley and McGowan 1998).
Evidence of extensive network effects and
 innovation that can rapidly tip markets

◦ Bar/Borrus/Steinberg 1995; Cowie/Marsden 1999
focussed policymakers’ attention on interoperability
 solution to emerging competition & innovation problems

◦ (EC 1997, De Nardis ed. 2011, van Schewick 2011)
Many de facto standards created
outside formal SSOs

Producers and consumers coalesce around a
dominant product or service,
◦ E.g. Windows OS, Intel microproc architecture

Problems of interoperability and refusal to licence
by the de facto standards setters,
◦ heavily enmeshed and interdependent environment of
computer software and hardware.

De facto standards setters leverage dominance
into other areas (Coates 2011)
Regulatory constraints created
Internet innovation in 1980-90s


Cannon (2003): fundamental regulation
imposed on U.S. telecoms firms in 1980s
◦ open network architecture (ONA)
◦ 1985 Computer III inquiry
 U.S. Federal Communications Commission (FCC).
 (Computer II and III inquiries refer to investigations by
the FCC into the regulation of data transfer and the
conditions necessary to achieve an increasingly
competitive market for that data.)

European equivalent amounted to
◦ interoperability plus
◦ physical interconnection between networks
 (Coates 2011)
Microsoft lessons?

Microsoft crushed Netscape
◦ moment of “Schumpeterian emergency”
 (Bresnahan, Greenstein, and Hendersen 2011)

Mehra 2011: innovative upstarts outwit
clumsier behemoths?
◦ Android/Apple v. Windows Mobile
Interoperability was solution
imposed – or refuted

Microsoft competition litigation,
◦ beginning with U.S. antitrust investigation in 1991 prior to
the dawn of mass Internet adoption,
enforced interoperability and application
programming interface (API) disclosure,
 Intel settling similar investigation into interoperability
and anticompetitive practices.
 Interoperability adapted by the complainants in

◦ Google and Facebook investigations EC 2010 (IP/10/1624)
◦ Apple’s iTunes price discrimination settlement EC in 2007–
2008 (IP/08/22)
◦ preliminary antitrust investigation into Apple’s App Store
policies (IP/10/1175).
Law and Code; Code and Law

Interoperability is not a panacea or magic
bullet in all cases
◦ Understanding protocols/standards
◦ as well as legal regulation,

leads to a better understanding of how
◦
◦
◦
◦
regulation can shape “better” standards
support interoperability and competition
Where required?
Lessig 1999/2006, Zittrain 2008
Regulating Code
Good Governance and Better Regulation in the
Information Age (MIT Press)
@IanBrownOII
@ChrisTMarsden
#RegulatingCode
Empirical investigation

Five case studies and one ‘prior art’ (encryption,
anonymity, security)
◦ Multi-year empirical investigation
◦ Builds on various EC/other studies including
 ‘Self-regulation.info’ (2001-4), ‘Co-regulation’ (2006-8), ‘Towards
a Future Internet’ (2008-10), ‘Privacy Value Networks’ (200811), ‘Network neutrality’ (2007-10) ‘Internet science’ (2012-15)

Reassesses prior art in view of ‘hard cases’
◦ Topics with no organised regulation/self-regulation
◦ Due to lack of consensus over solutions
◦ Clash between market outcomes and human rights
Prosumers not super-users


Web 2.0 and related tools make for active
users, not passive consumers
US administrative & academic arguments
◦ self-regulation may work for geeks,
◦ but what about the other 99%?

European regulatory space
◦ more fertile ground to explore prosumerism
◦ as both a market-based and
◦ citizen-oriented regulatory tool
Government and market failure



Industry capture of regulators & legislators
Incumbents introduce new barriers to entry
Continued exclusion of wider civil society
◦ tenuous chain of accountability of participants
◦ to voters, shareholders and NGO stakeholders.
◦ effectiveness, accountability and legitimacy of these
groups in representing the public interest?
EC Responses to Problems
Better open standards procurement?
 Decision No 922/2009/EC on interoperability solutions for
European public administrations (ISA)
 Better multistakeholder standard setting?
 Multi Stakeholder Platform

◦ Commission Decision OJ C 349, 30.11.2011, p. 4

Recognised in EC Regulation (2013) on European
standardisation,
◦ amending Council Directives 89/686/EEC and 93/15/EEC and
Directives 94/9/EC, 94/25/EC, 95/16/EC, 97/23/EC, 98/34/EC,
2004/22/EC, 2007/23/EC, 2009/23/EC and 2009/105/EC and
repealing Decision 87/95/EEC and Decision No 1673/2006/EC

Reinforces ICT standardisation efforts from Framework
Directive 2002/21/EC
Towards interoperability as
prosumer law

Solution for prosumers & competition
◦ enhance competitive production of public goods
◦ innovation, public safety, and fundamental rights

Key aspects:
◦
◦
◦
◦
Communications not competition policy
Ex ante intervention (ex post supplements)
Interoperability (incl. FRAND)
Fair and reasonable defined by govt procurement
 Not detailed rate of return regulation
 Note that IT software leaders make supra-normal returns
◦ Detailed software interoperability,
 not the general description offered by Gasser/Palfrey 2012
 Specifics in Gasser (2007)
What regulation teaches about
code
Ex ante + ex post intervention
 Interoperability

◦ Procurement policy + regulation/competition

A biased policy towards open code –
◦ Data open to mash-ups (government)
◦ Systems interoperable (procurement)
◦ Use of alternatives to market leader (e.g. Linux)
 Via competition remedies and sponsorship
Information regulation precedent

Must-carry/must-offer obligations,
◦ imposed on many market actors,
◦ including obliged to offer FRAND terms
 (common carriers, broadband access providers, cable broadcasters,
electronic program guides);

Interconnection requirements on telcos,
◦ especially those with dominance—
◦ And AOL/Time Warner merger requirement for
instant messaging interoperability

Application programming interfaces (API)
disclosure requirements,
◦ placed on Microsoft by EC upheld by ECJ
EC Mandated Browser Choice

2011: MSFT refused to allow browser
choice by default in Windows 7
◦ fined €561m March 2013,
◦ previously fined €497m 2007 €860m 2012.
 Browser “error” expensive line of code
Kroes’ promise post-Microsoft


Will “seriously explore
all options to ensure
that significant market
players cannot just
choose to deny
interoperability.
“The Commission
should not need to run
an epic antitrust case
every time software
lacks interoperability.”
Euro-Interoperability Framework

Response to multi-€bn competition cases:
◦ Microsoft saga (to 2009), Intel (2009), Apple
(2010), Rambus (2009)
◦ Google (2013?) perhaps Facebook....
◦ Coates (2011: Chapters 5-6)



Announced by DG Comp (CONNECT)
Commissioner Kroes 2009-2010
Bias in favour of interoperability in policy
Concerns are broader than competition
◦ Include privacy, IPR, security, fundamental rights
Economics and Human Rights
Open data, open code, and human rights
 Blizzard of Internet governance principles 2011

◦ Law/economics, or human rights, do not translate
◦ OECD/EC vs. UNHCR/OSCE/Council of Europe


This apparent dialogue of the deaf
◦ competition policy & corporate governance problem
Urgent task: dialogue between discrete expert fields
◦ ICT growth driver and transformative technology
◦ transformative role in communication and dialogue
 ‘arms trade’ in censorship technology; Twitter ‘revolution’ (sic)
EC Regulation (2013) Recital 41
“It is essential for the development of
European standardisation to
 continue fostering and encouraging the
active participation of European
organisations

◦ representing SMEs, consumers and
environmental and social interests.
◦ Such organisations pursue an aim of general
European interest
Article 5.1


European standardisation organisations
shall encourage and facilitate
 appropriate representation & effective participation
of all relevant stakeholders,
 including SMEs, consumer organisations and
environmental and social stakeholders

◦ in their standardisation activities.
 They shall in particular encourage and facilitate such
representation and participation through the European
stakeholder organisations receiving Union financing
Developing study of code regulation
Similarities and cross-over with
◦ complexity science
◦ network science
◦ web science/graph theory
 Match Internet regulation to complexity theory

 Longstaff (2003), Cherry (2008), Schneider/Bauer (2007)
Network science fusion of scientific/fundamental
elements from various components
 Internet Science? EC Network of Excellence

Many Research Questions?





Book published 22 March
2013
‘Regulating Code’ in
proceedings of ICIS, April
2013
‘Prosumer law’ paper for
EuroCPR (March version
now on SSRN)
‘Interoperability as a CodeBased Competition Remedy’
in Proceedings of IEEE SIIT
Comments welcome
Developing 2013 case studies
1.
Data protection in Social Networks
◦ Enforcement failures, Privacy by Design
◦ Dominance, network effects, corporate social
irresponsibility
2.
Search Neutrality
◦ Net neutrality argument
◦ Code-based solution to competition problem
◦ Prosumer focus cf. Microsoft
Google FTC and EC cases

Competition investigation
both sides of Atlantic since
2010:
◦ Settled with US authorities 3
Jan 2013
◦ Settlement proposal to EC 1
Feb 2013


Experts have severely
criticized timing and content
of FTC settlement
Grimmelman argued: “If the
final FTC statement had been
any more favourable to
Google, I’d be checking the
file metadata to see whether
Google wrote it.”
Source: Google proposal leaked to SearchEngineLand, 25/4/13
Grimmelman argued:
“If the final FTC statement had been any
more favourable to Google,
I’d be checking the file metadata to see
whether Google wrote it.”
4 lines of complaint
Search bias: Google favours own products over competitors in results
Vertical Search Opt-Out –
◦ Google don’t let websites opt out of particular uses of pages it indexes.
◦ complete opt-out giving up all Google traffic, a significant driver of traffic –
◦ especially Europe: Google has 90% search market in UK, Nl, France, Germany
Restricted 3rd party use of AdWords:
◦ “API Client may not [function] copies data between Google and 3rd Party.”
◦ Companies can advertise on Google and Bing,
◦ but cannot use a program to copy Google AdWords campaigns over to Bing.
◦ dropped by Google as token interoperability sop to FTC’s investigation;
Injunctions against standards-essential patents,
◦ including those by Google-acquired Motorola Mobility
 see Posner’s now–famous judgment in June 2012
◦ FTC concluded (4-1) unfair competition, Google agreed not to engage in it
◦ fires a shot not just at Google, but also at rivals –clever concession by Google!
Google and competitors routinely
privately regulate other’s code
points 3 and 4,
 Google claimed the right to regulate
others’ use of code,
 to use the AdWords API or
 to use Motorola Mobility’s patents..

“Prosumer law” approach
interoperability and content neutrality
1. Google to reinforce search neutrality

1. NOT bias results with search algorithms
2.
relatively trivial (by Google standards)
amendment to its code
1. allow websites more flexibility in listing,
2. rather than complete opt-out via the
existing robots.txt convention.
We do not make strong normative
claim that Google should adopt
neutral perspective
◦ (nor do we adopt approach to net neutrality),
We advocate truth-in-advertising
 Any search engine (or ISP using search)

◦ claiming verifiably neutral results
◦ produce the same
Or prominently advertise its product as
1. commercially driven,
2. affiliate-biased
3. selective search engine.
Requirement does not impose
significant regulatory burden



reinforces the brands of search providers of
integrity.
would not apply to selective search
providers if labelled such
‘a search engine which selectively provides
you with search results according in part to
its commercial affiliations’
◦ (or equivalent wording)

prominently displayed above search results
in that case.
Code-based solutions lighter than €1b
fines or structural separation

In book, we suggest similar approach
◦ to network neutrality violators

could not advertise their services as
allowing end-users’ choice
◦ in accessing the ‘Internet’

when in fact it is a commercial Intranet
◦ to which full access is provided.
Social networks: US solutions
instead of EU non-enforcement
Facebook’s 400m European users
 27 national regulators of personal data.
 Facebook chose regulator relocated in 2006

◦ from Dublin to Portarlington, Co. Laois, Ireland,
◦ Google is also regulated from Portarlington.
While German state and federal regulators and
others may rattle sabres at Facebook,
 Irish regulator audited Facebook spring 2012
 insisting on remedial action on nine counts

Prosumer law: direct intervention
Abusive dominant social networking sites
 prevent Facebook, Google+ any other

◦ from erecting a fence around its piece of the
information commons:
◦ ensure interoperability with open standards
 Which lowers entry barriers (in theory!)
 Enforcement of privacy law even in Portarlington
50 ways to leave Facebook

Not sufficient to permit data deletion
◦ as that only covers the user’s tracks.

Interconnection and interoperability,
◦ more than transparency and
◦ theoretical possibility to switch.

Ability for prosumers to interoperate to
permit exit
◦ Lower entry barriers tend to lead to increased
consumer welfare
US FTC constant audit




Class actions:
2011 $8.5m Google Buzz.
Jan 2013, Facebook $20m
Nov 2012, FTC Google settled for $22.5m
◦ tracking cookies for Safari browser users

2012, both agreed to settle privacy complaints
◦ FTC privacy audit of products for a 20-year period.
◦ That’s until 2032!
Sector-specific regulation of social networking already
exists de facto in the United States,
 Europeans wring their hands on the sidelines.

◦ proposed new European Regulation
◦ unlikely to be implemented before 2016.
Download