Document Sample
THE INTERNET Powered By Docstoc
LAWYERS                             SHOULD KNOW

The COMMONS Initiative: Cooperative Measurement and
Modeling of Open Networked Systems

CAIDA: Cooperative Association for Internet Data Analysis
    table of contents
                                                                              about the author
Table of contents, 01                       Point #6 07
Author biography                           How data is -08
         KC Claffy                           being used
                                                                     Kimberly Claffy
            Point #1 02                    Point #7 08               received her Ph. D.
        Updating legal              Normal regulatory -10            in Computer Science
          frameworks                                                 from UCSD. She is
                                     responses doomed
                                                                     Director and a prin-
           Point #2 02                                               cipal     investigator
 Obstacles to progress -03                 Point #8 10
                                 Problematic responses -16           for CAIDA, and an
           Point #3 03                                               Adjunct Professor
       Available data: -05                   Point #9 16             of Computer Science
        a dire picture                      The news is -19          and Engineering at UCSD. Kimberly’s
                                             not all bad             research interests include Internet
          Point #4 05                                                measurements, data analysis and vi-
The problem is not so -06                                            sualization, particularly with respect
                new                        Point #10 20
                                    Solutions will cross -23         to cooperation and sharing of Internet
                                            boundaries               data. She advocates the use of quanti-
         Point #5 06                                                 tative analysis to objectively inform
 An absurd situation
                                      Sponsors, Credits 24           public Internet policy discussions. Her
                                                                     email is
 Adapted from:

             Last year Kevin Werbach1 invited me to his Supernova 20072 conference to give a 15-
 minute vignette on the challenge of getting empirical data to inform telecom policy. They posted the
 video of my talk3 last year, and my favorite tech podcast ITConversations4, posted the mp35 as an
 episode last week. I clearly needed more than 15 minutes...
             In response to my “impassioned plea”, I was invited to attend a meeting in March 2008
 hosted by Google and Stanford Law School — Legal Futures6 — a “conversation between some
 of the world’s leading thinkers about the future of privacy, intellectual property, competition, in-
 novation, globalization, and other areas of the law undergoing rapid change due to technological
 advancement.’’ There I had 5 minutes to convey the most important data points I knew about the
 Internet to lawyers thinking about how to update legal frameworks to best accommodate informa-
 tion technologies in the 21st century. With a few more months of thought, here is my current top ten
 list of the most important things lawyers need to understand about the Internet.

                                All content in this booklet (including text, photographs, graphics, and
                                any other original works), is licensed under a Creative Commons
                                Attribution-No Derivatives 3.0 license.7

1             Kevin Werback bio,
2             Supernova 07 Conference archive,
3             “Closing the Interactive Loop”,
4             ITConversations,
5             “No Access To Internet Data”, Jun 2007,
6             Legal Futures Conference, Mar 2008,
7             Creative Commons Attribution-No Derivative 3.0, 2008,

                                   point one point two                                                      Science Foundation’s program for Internet secu-          Another caveat: most security-related studies
                                                                                                            rity research13 spends ~$35M/year on dozens of           are published or funded by companies trying to
     I. Updating legal frameworks to                     II. Our scientific knowledge about the             research projects, none of which have data from          sell more security software, their objectivity is
     accommodate technological ad-                       Internet is weak, and the obstacles to
                                                         progress are primarily issues of econom-           operational Internet infrastructure.                     also in dispute. Again, EOT factors render truth
     vancement requires first updating
     other legal frameworks to accommo-                  ics, ownership, and trust (EOT), rather                                                                     elusive.
                                                         than technical.                                    C. Not only is traffic data off limits, but sharing
     date empirically grounded research
     into what we have built, how it is                                                                     data on the structure of the network is forbid-         point three
     used, and what it costs to sustain.                Economically, network research is perpetually       den too — commercial ISPs are typically not
                                                        behind network evolution — basic instrumen-         even allowed to disclose the existence of peer-               III. Despite the methodological limi-
                                                        tation can increase in cost 10X with one net-       ing agreements, much less their terms. So when                tations of Internet science today, the
There is increasing recognition that various legal      work upgrade, while network research budgets        developing tools for accurate Internet mapping,               few data points available suggest a
frameworks (from copyright to privacy to wire-                                                                                                                            dire picture:
                                                        are lucky to stay even. But the ownership and       researchers cannot validate the connectivity
tapping to common carriage) need updating in            trust obstacles are even greater: policy support    inferences they make, since the information is
light of technological developments of the last         for scientific Internet research has deteriorated   typically intended to be secret.
few decades. Unfortunately, the light is too dim        along several dimensions10 since the National                                                         A. We’re running out of IPv4 addresses that can
to really understand Internet behavior, usage           Science Foundation left the scene in 199511, and                                                      be allocated (there are many allocated addresses16
patterns, architectural limitations, and econom-        further when DARPA pulled out of funding ac-                                                          that are not in observed use17 , but there is no
ic constraints, because current legal frameworks        ademic networking research after 9/1112. Some                                                         policy support (yet) for reclamation or reuse)18,
for network provisioning also prevent sharing of        data points exposing the state of “Internet sci-                                                      and the purported technology solution ( IPv619
data with researchers 8 to scientifically investigate   ence”:                                                                                                ) requires investment that most ISPs are not pre-
any of these questions. Even for data that is legal                                                                                                           pared to make20. Regardless of whether Internet
to share, there are overwhelming counter incen-         A. Two decades of Internet research have failed                                                       growth is supported by IPv6 or a concerted ef-
tives to sharing any data at all in the competi-        to produce generally usable tools for band-                                                           fort to scrape more lifetime out of the current
tive environment we have chosen — although              width estimation, traffic modeling, usage char-                                                       IPv4 protocol, it will induce growth of core
not achieved9 — for the network provisioning            acterization, traffic matrix estimation, topol-                                                       Internet routing tables relying on a routing
industry.                                               ogy mapping, or realistic Internet simulation,                                                        system that is increasingly inappropriate for
                                                        with progress primarily blocked on the ability                                                        the Internet’s evolving structure. So while it’s
So while I support updating legal frameworks            to test them out in realistic network and traffic                                                     fair to say that we need a new routing system21,
to be congruent with reality, I think we need to        scenarios. A few researchers who do manage to       D. OECD published a 53-page report: Measur- no institution or agency has responsibility for
first confront that we have no basis for claiming       get data via relationships of mutual trust (in-     ing security and trust in the online environment: developing one much less the global economic
what reality is yet.                                    cluding CAIDA) are not allowed to share data        a view using official data14. As you may have and political challenge of deploying it.
 “no aphorism is more frequently repeated…              with other researchers, inhibiting reproducibil-    guessed by now, the report about ‘measuring se-
 than that we must ask Nature few questions,            ity of any result. Compared to established fields   curity’ is based on no measurements from any B. Pervasively distributed end-to-end peering
 or ideally, one question at a time. The writer is      of science, it is hard to defend what happens in    networks, only survey data reflecting user per- to exchange information is not only threatening
 convinced that this view is wholly mistaken.           the field of Internet research as science at all.   ceptions of their own security, which other stud- the integrity of the routing system, but also the
 Nature, he suggests, will best respond to a logi-                                                          ies15 have shown to be uncorrelated with reality. business models of the ISPs22. Although it bears
                                                        B. U.S. (and other) government agencies con-
 cally and carefully thought out questionnaire;         tinue to spend hundreds of millions of dollars
 indeed if we ask her a single question, she will       per year on network research — with cyber se-       13        NSF       Trusted      Computing      program,
 often refuse to answer until some other topic          curity research being the most fashionable this     id=503326&org=NSF
 has been discussed.” Sir Ronald A. Fisher,                                                                 14        Measuring Security and Trust in the Online Environment, Jan 2008,
                                                        decade — funding researchers who almost nev-        dataoecd/47/18/40009578.pdf
 Perspectives in Medicine and Biol-
                                                        er have any data from realistic operational net-    15        McAfee/NCSA Cyber Security Survey, Oct 2007,
 ogy, 1973.                                                                                                 us/McAfeeNCSA_Analysis09-25-07.pdf
                                                        works. An illustrative example: the National
                                                                                                            16        ANT Censuses of the Internet Address Space, Jun 2008,
                                                                                                            17        Mapping The IPV4 Address Space, Mar 2008,
 8          “Toward a Culture of Cybersecurity Research”, 2008,         18        ARIN’s Number Resource Policy Manual, Aug 2008,
 cfm?abstract_id=1113014                                                                                    19        ARIN IPv6 Wiki,
 9         Telecommunications Act of 1996,                 20        ARIN & CAIDA IPv6 Survey Results, Apr 2008,
 10        “Measuring the Internet”, Jan 2001,      PDF/monday/IPv6_Survey_KC.pdf
 11        “Post-NSFNET statistics collection”,         21
 12        “CS profs and the DOD”, Sep 2007,        tem/
                                                                                                            22        Jan 2008,
02                                                                                                                                                                                                                       03
noting that the business models for moving            because the network is still pretty much work-        than succeed at its own mission40.                 data collected by researchers and providers.
Internet traffic around have long been suspect,       ing” discounts the fact that criminals using the                                                         Since the real obstacles have to do with eco-
                                                                                                            We don’t have a lot of data about the Internet,
since the network infrastructure companies            Internet need it to work just as well as the rest                                                        nomic, ownership (legal), and trust (privacy)
                                                                                                            but what little we have is unequivocally cause for
that have survived the bubble have done so by         of us. Although we admit we don’t know how to                                                            constraints rather than technology issues, this
spending the last fifteen years manipulating23        measure the exact size of botnets36 what we know                                        point four       catalog has been less utilized45 than we hoped.
the network architecture24 and the regulatory ar-     for sure is that millions of compromised (Windo-
                                                                                                                                                                    C. Recognizing that the data sharing problem consti-
chitecture25 away from the Internet architecture      ws) systems37 are taking advantage of network            IV. The data dearth is not a new prob-               tutes a threat to national security46 the U.S. Department
(smart endpoints) toward something they can           and host software vulnerabilities to support un-         lem in the field; many public and pri-
                                                                                                               vate sector efforts have tried and failed            of Homeland Security (specifically, HSARPA47) has
control (smart network) in order to more effec-       known (but underground estimates are many)
                                                                                                               to solve it.                                         spent 4 years developing a project — PREDICT48—
tively monetize their assets26. Since the Internet    billions of dollars per year of criminal activi-
                                                                                                                                                                    to facilitate protected sharing of realistic network data
architecture was originally designed to be a gov-     ties (or activities that would be criminal if law-
                                                                                                                                                                    that will enable cybersecurity researchers to validate
ernment-sponsored file-sharing network27 with no      makers understood enough to legislate against
                                                                                                             A. Information Sharing and Analysis Centers,           the network security research and technologies they
support for usage-based (or any) billing, its fail-   them) with no incentive framework to support
                                                                                                             such as those that exist for the financial services    develop. Unfortunately after four years the PREDICT
ure as a platform for a purely competitive telecom-   their recovery. Although ICANN is trying to
                                                                                                             industry41 have been attempted42 several times,        project has not yet launched, and when it does it will
munication industry28 is not surprising. But we       set policies to counter some of the malfeasance38
                                                                                                             but there is no research activity or channel to        not be able to include data on networks that serve the
are going to be so surprised.                         that arguably falls under its purview (domain
                                                                                                             share data with the research community, nor            public49, since the legal territory is too muddy for DHS
                                                      names and IP addresses), ICANN lacks the ar-
C. There are demonstrated vulnerabilities in the                                                                                                                    lawyers to navigate while EFF lawsuits50 have everyone
                                                      chitecture and legitimacy it needs to enforce any      any independent analysis of the performance
most fundamental layers of the infrastructure                                                                                                                       in the U.S. government skittish about acknowledging
                                                      regulations39, and continues to struggle more          or progress of such a group.
(naming29 and routing30) for which technologi-                                                                                                                      surveillance of any kind. Even the private networks that
cal31 solutions32 have been developed but have                                                               B. The National Science Foundation has spent           PREDICT can serve immediately, such as Internet251
failed to gain traction under the political33 and                                                            at least $1M on CAIDA’s Internet measure-              (the research backbone in the U.S. serving a few hun-
economic constraints34 of real-world deployment                                                              ment data catalog43 to support sharing of In-          dred52 educational, commercial, government, and inter-
In the meantime, over 98% of traffic sent to root                                                            ternet measurements, but as a science and              national partners) have lamented that the PREDICT
domain name servers is pollution35.                                                                          engineering funding agency, NSF could only             framework does not solve their two biggest problems:
                                                                                                             fund the technical aspects of the data sharing         sketchy legal territory, and fear of RIAA subpoenas
D. The common lawyerly assumption that “the                                                                  activity: developing a database44 to support cu-       and/or lawsuits. Meanwhile, other accounts53 (from
Internet security situation must not be so bad                                                               ration, indexing, and annotation of Internet           non-objective parties54, with no data sources) claim that

23         Telecommunications Act of 1996,
24         “The Walled Garden: Access To Internet and Digital Computer Communications Technology”,
Apr 1996,                                                          40         Feb 2008,
25         Telecommunications Act of 1996,                                   41         Financial Services-Information Sharing and Analysis Center,
26         IMS ‘Reality Check’,               42         Information Sharing and Analysis Center (ISAC), Jan 2001,
27                                                              press/2001/itsac011601.htm
28         “Consolidation Leads to ISP Shake-Out in Europe”, Jul 2001,      43          “Correlating Heterogeneous Measurement Data to Achieve System-Level: Analysis of Internet
consolidation.cfm                                                                                            Traffic Trends”, Aug 2001,
29         Oct 2006,        44         DatCat, indexed Internet measurement data,
30         “Tubular Routing”, Mar 2008,                 45         DatCat traffic summary, Aug 2008,
31         Domain Name System Security Extensions,                       46
32         Secure BGP Project (S-BGP),                                           modeling-internet-attacks/
33         “DNSSEC-The Opinion”,                              47
34         “Operational Requirements for Secured BGP”,       48         Protected Repository for the Defense of Infrastructure Against Cyber Threats (PREDICT),
35          Comparison of Traffic from DNS Root Nameservers in DITL 2006 and 2007, http://www.caida.         49         Privacy Impact Assessment for PREDICT, Feb 2008,
org/research/dns/roottraffic/comparison06_07.xml#concl                                                       privacy_pia_st_predict.pdf
36         Kanich, et al., “The Heisenbot Uncertainty Problem: Challenges in Separating Bots from Chaff”,    50         NSA Spying, cases and resources,                                                    51
37                                                   52         Internet 2 Members list, Jul 2008,
38                                               53         “Comments of NBC Universal, Inc., to the Federal Communications Commission”, http://www.
39         “ICANN Reform: Establishing the Rule of Law”, Nov 2005,
ICANN_Rule-of-Law.pdf                                                                                        54         “The Movie Industry’s 200% Error”,

04                                                                                                                                                                                                                        05
 the vast majority of traffic on the Internet is illegal by   construction with parameters arranged to prove         information they extract65 from the data. No one        for real data77 in the U.S. do support the claim
 current laws55 and ISPs should be held accountable56 for     the need for the kind of traffic management be-        is investing in technology to learn about net-          that the current Internet transit business model is
 preventing this traffic57. Given the exposure to copyright   havior that ATT lobbyists are trying to justify        works while minimizing the amount of privacy            broken78. Whether the growth in traffic is due to
 lawsuits for file-sharing (ironically, what the Internet     to regulators and their customers. As with many        compromised in the process. This inherent in-           http transport of user-generated video79, or radi-
 was originally designed58 to do), the counter incentives     other questions about network architecture, be-        formation asymmetry66 of the industry is at the         cally distributed80 peer-to-peer file sharing (also
 to sharing data on operational networks grow stronger        havior, and usage, there are valid (i.e., empirical-   root of our inability to verify claims67 regarding      often video81), there is strong evidence82 from
 by the day.                                                  ly validated) inferences to make regarding QoS         either security68 or bandwidth crises69 justifying      network providers themselves83 that the major-
                                      point five              versus the alternatives, which could immediately       controversial business practices70 that threaten an     ity of bytes84 on the network are people moving
                                                              inform telecom and media policy, but research-         admittedly fuzzy, but increasingly popular71 con-       files85 from machine to machine, often the same
  V. Thus the research community is in
  the absurd situation of not being able to                   ers are not in a position to make them.                cept of Internet access rights.72 Although the little   files moving from a few sources to many users.
  do the most basic network research even                                                                            data that researchers can scrape together73, most       Unfortunately, this evidence implies that the
  on the networks established explicitly to                   point six                                              of it from outside the U.S., do not support the         current network and policy architectures are
  support academic network research.                                                                                 “p2p is causing a bandwidth problem74” claim,           astonishingly inefficient86, and that clean slate87
                                                                 VI. While the looming problems of the Inter-        the press releases75 we see as a popular76 substitute   Internet researchers88 should be thinking about
 This inability to do research on our own research               net62 indicate the need for a closer objec-
                                                                 tive look, a growing number of segments
 networks leads to contradictions in our field of                of society have network measurement
 “science” that cannot be resolved, including on                 access to, and use, private network in-             65          Make extra money with a Google search box,
 the most politically relevant network research                  formation on individuals for purposes               html
 questions of the decade: what are the costs and                 we might not approve of if we knew how              66
                                                                 the data was being used.                            67          “Threats to the Internet: Too Much or Too Little Growth”, Feb 2008, http://www.internetevolution.
 benefits of using QOS to support multiple ser-                                                                      com/author.asp?section_id=592&doc_id=146747
 vice classes, to users as well as providers, and how                                                                68          “Cybersecurity measure unduly secretive”,
 should these service classes be determined? Two                                                                     25-cybersecurity-backlash_x.htm
                                                              To the extent that we are investing public or          69          “Does Bell Really Have a P2P Bandwidth Problem?”, Apr 2008,
 research papers on this same topic contradict                                                                       posts/841710_bell_p2p_bandwidth_problem/
 each other — Why Premium IP Service Has Not                  private sector dollars in trying to measure the        70          “Comcast opens up about how it manages traffic”,
 Deployed (and Probably Never Will59) from In-                Internet, they are not in pursuit of answers to        money/20080421/
                                                              questions related to the overall network infra-        71          The internet freedom fight goes to Washington.
 ternet260 (the U.S. research and education back-                                                                    72          “Bill of Internet Access Rights”, Jun 2006,
 bone) and The Evolving Internet - Traffic, Engi-             structure’s health, system efficiency or end-to-       73          A Day In The Life of the Internet: A Summary of the January 9-10, 2007 Collection Event, http://
 neering, and Roles61 from ATT — with neither                 end performance, or any of the questions that
                                                              engineers would recommend knowing about                74          See footnote 69
 paper offering actual network data, although                                                                        75          Minnesota Internet Traffic Studies (MINTS), 2002,
 the Internet2 paper claims to be based on data               a communications system. The measurements              76          “Unleashed the ‘Exaflood”, Feb 2008, 010084479-
 from the Internet2 backbone. The ATT paper                   happening today are either for national securi-        lMyQjAxMDI4MDIzMjYyMzI5Wj.html
                                                              ty63or business purposes64, which both have an         77          Internet Study 2007: The Impact of P2P File Sharing, Voice over IP, Skype, Joost, Instant Mes-
 uses unsubstantiated numbers from invalidated                                                                       saging, One-Click Hosting and Media Streaming such as YouTube on the Internet,
 sources on the web and a model and simulation                incentive to maximize the amount of personal           news_&_events/internet_studies/internet_study_2007
                                                                                                                     78          “Avoiding Net traffic tie-ups could cost you in the future”,
55           “Home Copying - burnt into teenage psyche”, Apr 2009,        79          Youtube,
apr/07/digitalmusic.drm                                                                                              80
56           “IFPI hails court ruling that ISPs must stop copyright piracy on their networks”, http://www.ifpi.      81
org/content/section_news/20070704b.html                                                                              82          “Comments of AT&T on petitions of Free Press, Vuze, et al.. before the Federal Communications
57           “AT&T and Other I.S.P.’s May Be Getting Ready to Filter for copyrighted content at the network          Committee in the matter of broadband industry practices”, Feb 2008,
level”,        trieve.cgi?native_or_pdf=pdf&id_document=6519841106
58                                                                     83          “Comments of Verizon before the FCC in the matter of broadband industry practices”, Feb 2008,
59           “Why Premium IP Service Has Not Deployed (and Probably Never Will)”,
wg/documents-informational/20020503-premium-problems-non-architectural.html                                          84          “Comments of Comcast Corporation before the FCC in the matter of broadband industry prac-
60                                                                                tices”, Feb 2008,
61           “The Evolving Internet - Traffic, Engineering, and Roles”,                85          “Comments of Time Warner Cable Inc., before the FCC in matters of broadband industry prac-
papers/2007/786/Evolving%20Internet.pdf                                                                              tices”, Feb 2008,
62           “Top problems of the Internet and what can be done to help”,         86          “Democracy Now! Saves $1,000 a month with bittorrent”, Feb 2008,
presentations/2005/topproblemsnet/                                                                                   blog/2008/02/huge-cost-savings-bittorrent-vs-http/
63           “Bush Order Expands Network Monitoring”, Jan 2008,            87
tent/article/2008/01/25/AR2008012503261.html?hpid=moreheadlines                                                      88          FIND (Future Internet Design), NSF research program whose goal is to consider requirements for a
64           “Should AT&T police the Internet”, Jan 2008,        global network of 15 years from now, and how we would build such a network if we could design it from scratch,
06                                                                                                                                                                                                                              07
how to create truly scalable89 inter-domain rout- to make trillion-dollar communication and                     approve of how it is measuring broad-
ing90 and policy architectures91 that are content- technology policy decisions in the dark.                     band penetration107. The FCC has no
centric92, leverage our best understanding of the                                                               empirical basis in fact nor apparent
structure of complex networks93, and still manage
                                                        point seven                                             authority108 in a conversation about
to respect privacy94. No easy trick, especially with                                                            traffic, structure, pricing, or vulner-
                                                         VII. The traditional mode of getting
no viable deployment path for such a new archi-          data from public infrastructures to in-                abilities on the network since it has
tecture, at least in the U.S. where we have jetti-       form policymaking — regulating its                     no access to data from Internet in-
soned the policy framework95 that allowed innova-        collection — is a quixotic path, since                 frastructure beyond what providers
tions96 like the Internet. It should be no surprise      the government regulatory agencies101 have             volunteer to provide. And yet little da-
if the status quo is unsustainable , since we are
                                                         as much reason to be reluctant102 as pro-              ta109 is needed to reveal that the Inter-
using the network quite differently from how it
                                                         viders103 regarding disclosure of how                  net’s underlying network architecture,
was intended . But if a new network architec-
               99                                        the Internet is engineered, used, and                  implementation, and usage is funda-
ture is needed, that’s a discussion that needs to        financed.                                              mentally inconsistent with almost ev-
include some validated empirical analysis of                                                                    ery aspect of our current communications and              Hence it should be no surprise if solutions to
                                                     For every other critical infrastructure in soci-           media policy architecture. The Internet sheds             measurement, like other persistent problems of
                                                     ety we have devoted a government agency to                 deep skepticism on current legal frameworks for           the Internet118, require engaging deeply with
                                                     its stewardship. The Internet was designed for             copyright, wiretapping, and privacy, as well as           economics, ownership and trust issues. Alas, In-
                                                     a cooperative104 rather than competitive policy            transforms or destroys dozens of industries that          ternet economics research is one of the few fields
                                                     architecture, so its designers did not consider            hold great economic and political power today.            worse off119 than Internet traffic or topology re-
                                                     regulatory aspects. But as a communications in-                                                                      search with regard to the ability to validate any
                                                     frastructure serving the public, most regulatory           The national security components of Internet              models or assumptions. (If you think tcpdump
                                                     aspects of Internet fall under the jurisdiction of         regulation, from wiretapping110 to disaster re-           and traceroute are replete with measurement er-
what we have already built. So long as the net- the agency who regulates the tubes105 it typically              covery111 to unstable112 leadership113 lamenting          ror, you should try analyzing the economics of
work infrastructure companies are so counterin- runs atop: in the United States that means the                  its budgetary and policy handicaps,114 inspire con-       network infrastructure companies. And if you
cented to share data100, we will continue having     FCC. Unfortunately, the FCC is not completely              cern115 than hope. That over 1% of observed web           think packet header and internal topology data
                                                     up to speed106 on the Internet, and does not even          pages are modified in flight without our knowl-           is hard to get, you should try to get financial
                                                                                                                edge116 is no source of comfort117 either.                numbers from the same companies broken out
89          “On Compact Routing for the Internet”,
routing/                                                                                                                                                                  by service offered so you could see how the ec-
90          Routing Research Group,
91          “Retrograde Inversion”, Apr 2008,
92                                             107
93          “Self-similarity of complex networks and hidden metric spaces”,       108         “Comcast: FCC lacks any authority to act on P2P blocking”, Mar 2008,
tions/papers/2008/self_similarity/                                                                              news.ars/post/20080318-comcast-fcc-lacks-any-authority-to-act-on-p2p-blocking.html
94          Electronic Privacy Information Center,                                             109         See point three
95          National Cable & Telecommunications Association et al. v. Brand X Internet Services et al.,         110         “Congress worries that .gov monitoring will spy on Americans”, Feb 2008, http://news.cnet.
2005,                                                      com/8301-10784_3-9882031-7.html
96           Common Carriers,                              111         “Internet Infrastructure: Challenges in Developing a Public/Private Recovery Plan”, 2006, http://
97          “PNSol: Answer to the broadband riddle?”, 
98          PEW /Internet Reports: Online Activities,             112         “Help (still) wanted: Cybersecurity czar”, Jul 2006,
99          Inventing the Internet, Jul 2000,       curity-czar/2100-7348_3-6094055.html
Abbate/dp/0262511150/                                                                                           113         “Cybersecurity czar will have hard road ahead”, Jun 2005,
100         See footnote 8                                                                                      news/11194
101         Bureaucracy: What Government Agencies Do And Why They Do It, Jan 1991, http://www.ama-              114         “RSA: Chertoff Likens U.S. Cyber Security to ‘Manhattan Project’”, Apr 2008, http://www.infor-                                 
102         “Can the FCC Handle The Truth?”, Apr 2008,                 115         “Will the next U.S. president lead on cybersecurity?”, Oct 2007,
103         “If Comcast Shares Broadband Deployment Data, The Terrorists Win”, Mar 2008, http://www.            _3-9807450-7.html                    116         “Detecting In-Flight Page Changes with Web Tripwires”,
104         “Tward A Cooperative Network Of Time-Shared Computers”, Nov 1966,             full_papers/reis/reis_html/index.html
files/toward-coop-net.html                                                                                      117         “The Digital Imprimatur: How big brother and big media can put the Internet genie back in the
105         Senator Ted Stevens talking about the Net Neutrality Bill: “Series of Tubes”,   bottle”, Sept 2003,
com/watch?v=f99PcP0aFNE                                                                                         118         See footnote 62
106         “FCC to start collecting more detailed broadband info”, Mar 2008,      119
pbcs.dll/article?AID=/20080321/SUB/669906587/1005/allnews                                                       i-have-learned-so-far/
08                                                                                                                                                                                                                             09
omics are actually evolving120.)                       capital, as well as realignment of incentives133 to      fundamentally related143 a larger number of ad-          rol of not only the U.S. government but the U.S.
                                                       promote cooperation among competitive play-              dresses actually exacerbates the routing problem,        military150. And it still took a couple of rounds of
Unfortunately (again) understanding the eco-
                                                       ers. Where does that capital and incentive to            getting us closer to the wall that seemed further        threats to cut off funding to attached sites who
nomics of the system is not where spare private
                                                       cooperate come from?                                     away. In the meantime, the current IPv4 routing          did not upgrade!) Instead, operators are busy ex-
or public sector capital is going. In the 1990’s the
                                                                                                                table is already splintering into smaller pieces as      perimenting with business models to try to figure
telecoms spent their capital suing each other121
and the government122 over laws so vaguely writ-
                                                       point eight                                              network operators engineer finer-grained con-            out how to make a profit on IP transit, e.g., fancy
                                                                                                                trol over traffic patterns. So, while IPv6 exists        QOS services that customers aren’t asking for151,
ten123 as to defy consistent interpretation124, much
                                                          VIII. The opaqueness of the infrastruc-               as a set of technologies, many experts are grim          metered pricing152 (known to have its own prob-
less measurable enforcement, across any two con-          ture to empirical analysis has gener-                 about its future, since it doesn’t solve144 the funda-   lems153), or giving up and getting rid of the part of
stituencies125 in the ecosystem. This decade we           ated many problematic responses from                  mental routing scalability problem.                      the company that moves IP traffic around154. They
are spending our capital suing the telecoms126 for        rigidly circumscribed communities
                                                          earnestly trying to get their jobs done.                                                                       have also recently experimented with reforming
not suing the government after 9/11 when the                                                                    B. Most network operators145, especially for-prof-
                                                                                                                                                                         their industry trade meetings155 to be more useful
government asked them to break laws127 that are                                                                 it ones, cannot justify the investment to deploy
                                                                                                                                                                         given that they aren’t authorized to share any sig-
                                                                                                                IPv6 when their customers are not asking for it146,
                                                       A. To its credit, the IETF134 acknowledged and                                                                    nificant information about their own networks.
                                                                                                                and their customers won’t ask for it until they
                                                       endeavored to solve the technical limitations of                                                                  In the meantime, if they have one, they heavly
                                                                                                                can no longer get IPv4 addresses. Large network
                                                       the current IPv4 protocol135, primarily the insuffi-                                                              subsidize from the magnificently profitable wire-
                                                                                                                operators continue to remind IETF engineers
                                                       cient number of addresses136 and the inherent scal-                                                               less156 side of the company while they build the
                                                                                                                that they didn’t solve the problem the network
                                                       ability limitations of137 the routing architecture138.                                                            case for more deregulation.
                                                                                                                operators really need solved147. Operators do real-
                                                       To its chagrin139, the IETF learned that neither
                                                                                                                ize they are all in this together, but they aren’t       C. Thinking about the health of the Internet ten
                                                       the philospher king nor rough consensus-based140
                                                                                                                institutionally structured to think longer than          years out or longer should theoretically happen
                                                       approach would yield an architecture that made
                                                                                                                five years out148. They also lack the capital, legal     within the stewardship missions of ICANN157
                                                       progress on both problems at the same time. So
just as outdated as the copyright laws128. Thomas                                                               framework, and incentive to develop an alterna-          and the ICANN-rooted address registries158, who
                                                       the IETF punted on the routing problems141 since
Jefferson129 would no doubt recommend rewriting                                                                 tive replacement, even in partnership with their         lease Internet address space based on demonstrat-
                                                       they seemed further away, and focused on build-
all of it from scratch130. Unfortunately the timing                                                             suppliers. (The last time we upgraded the network        ed need159 . The ICANN and registry communit-
                                                       ing a new network architecture that had a larger
is bleak: these developments are occurring at a                                                                 architecture149 the network was under the cont-
                                                       number of addresses, and some other142 stuff
time when sustaining Internet growth131(which,
                                                       most people don’t usually mention. But because
no, we still do not have good ways to measure132..)
                                                       today’s addressing and routing architectures are
will require extraordinary investment of                                                                        143         “RFC4116 - IPv4 Multihoming Practices and Limitations”, Jul 2005,
                                                                                                                144         “Storm clouds looming for Internet, experts say”, Apr 2008,
 125         Reciprocal Compensation,                           article/08/04/17/Storm-clouds-looming-for-Internet-experts-say_2.html
 126                                                                             145         Nanog (North American Network Operators’ Group),
 127                              146         ARIN & CAIDA IPv6 Survey Results, Apr 2008,
 128                         PDF/monday/IPv6_Survey_KC.pdf
 129         The Thomas Jefferson Hour,                                           147         “RFC4116 - IPv4 Multihoming Practices and Limitations”, Jul 2005,
 130         Thomas Jefferson Hour episode: Constitutional Convention,           rfc4116.html
 Show%20672%20Constitutional%20Convention.mp3                                                                   148         The Art of the Long View: Planning for the Future in an Uncertain World, 1991, http://www.gbn.
 131         “Eight Bold Steps To A National Broadband Strategy:”, Jan 2007,        com/BookClubSelectionDisplayServlet.srv?si=361
 baller-lide_8Steps_NatBBStrategy.pdf                                                                           149
 132         “IPv6 Deployment: Just where are we?”, Apr 2008,       150         U.S. Department of Defense,
 html                                                                                                           151         See footnote 61
 133         “CAIDA and ARIN Release IPv6 Survey”, Apr 2008,         152         “Time Warner Cable to experiment with Internet access charges”, Jan 2008,
 arin_ipv6_survey/                                                                                              time-warner-cable-to-experiment-with-internet-access-charges/
 134         Internet Engineering Task Force,                                 153         “Why Metered Broadband Slows Internet Innovation”.
 135         Internet Protocol version 4 (IPv4),                              11/194749.shtml
 136         IPv4 exhaustion,                                      154         “Time Warner to unplug cable business”, Apr 2008,
 137         “Models of policy based routing”, Jun 1989,                  dd-bbfc-0000779fd2ac.html
 138         NSFNET routing architecture, Feb 1989,                       155         “NANOG Evolution-What’s Next?”,
 139, 140 Process for Organization of Internet Standards Working Group (POISED), Jun 1994, http://www.          156         “Led by Wireless Unit, AT& T Reports a 21.5% Gain”, Apr 2008,                                                                                 wp-dyn/content/article/2008/04/22/AR2008042202921.html
 141         “Lessons from IPv6”,                        157         See footnote 40
 142         Internet Protocol version 6 (IPv6),                              158         The Address Supporting Organization:
                                                                                                                159         ARIN Number Resource Policy Manual,
10                                                                                                                                                                                                                        11
ies recognize the limitations of IPv6, and by now        impede the uptake of IPv6, so the bottom-up              rket evolution of IP economics185 is the root cause186   developing lots190 of technology191, even legal tech-
also the limitations of the IETF160. IETF experts        registries are inherently conflicted regarding the       of the mess.) When the dizziness subsides, we            nology192 to promote data retention193 and shar-
are similarly astute161 about the problems with          problem they’re trying to solve.                         will have to acknowledge that the carriers are           ing194, but the agencies and the taxpayers they
ICANN162. And of course both communities                                                                          right: it would be a disaster if the government told     represent get little in return. A related problem
                                                         D. Meanwhile, over in the media policy173, re-
are aware of the pressure on the current address                                                                  carriers how to manage congestion on their net-          is that the lack of experience with data sharing in
                                                         form174, passionate activist175, and well-inten-
space. Since IPv6 is the only existing solution,                                                                  works187, which is why the endgame must be —             an admittedly quite young field of science means
                                                         tioned legal scholar176 corner of cyberspace, it is
they both promote IPv6 deployment163, although                                                                    as it has always been with essential facilities and      that there is no established code-of-conduct195 for
                                                         as if Eli Noam177’s warning about the imminent
they lack reliable methods164 to measure IPv6                                                                     common carriage — that carriers do not have              protecting196 user privacy197 and engaging with
                                                         death of common carriage178 were not published
uptake165 without data from operators166. So, this                                                                financial interest in the content of what they’re        Institutional Review Boards198 to navigate ethical
                                                         fourteen years ago. Despite the lack of any pro-
year they are finally re-discussing a backup plan:                                                                carrying. But that idea — although it is the same        issues in Internet measurement research. Worse
                                                         posed operationally enforceable definition179 of
privatizing IPv4 address markets167, in case they                                                                 type of structural regulation that made the In-          yet, conservative interpretations of the current
                                                         network neutrality, the conversation thrives180—
run out of IPv4 addresses168 before IPv6 gains                                                                    ternet possible — offends any capitalist sense of        relevant statutes conclude that most network
                                                         an understandable post-traumatic reaction to
                                                                                                                  profit margins.                                          measurement research is currently approximately
traction. There is little background169 research         the recent jettison of at least eight centuries of
                                                                                                                  E. Academic Internet researchers also operate            illegal199, but there is no consensus on what kind
on the implications170 of private ownership of           legal doctrine181 from our primary communica-
                                                                                                                  in a funding environment188 that does not pro-           of legislative changes are needed200, if any. The
addresses171, but what exists is not auspicious172.      tions fabric. Even the FCC is looking for ideas182
                                                                                                                  mote tackling 10-year problems189, nor are they          stunted legal process prevents sharing of data sets
Furthermore, the possibility that a legitimate           (strangely, they’re explicitly not interested in da-
                                                                                                                  equipped to navigate the conflict of interests           that could help solve immediate problems, but
market for IPv4 address may emerge will itself           ta183, despite clear indications184 that the free ma-
                                                                                                                  between the university and the providers of net-         the collateral damage is that it prevents informed
                                                                                                                  work data. Providers either legally cannot or are        discussion of what even needs to be known on
160         “Into the Future with the Internet Vendor Task Force - A very Curmudgeonly View or Testing                                                                     the net, and who needs to know it. Do we want
Spaghetti - a Wall’s Point of View”,
                                                                                                                  reluctant to share data without restrictions on
161         IETF (Internet Vendor Task Force) summary, Sep 2002,       what can be published about their network, and           to know how much peer-to-peer traffic is transit-
presentations/ripe43-plenary-ietf-icann/sld022.html                                                               universities have rules limiting such restrictions.      ing backbone links? How much encrypted traf-
162         “An ICANN reform plan”, Jun 2002,                                                                                fic? How much copyrighted traffic? Right now
163         ARIN IPv6 Wiki,
                                                                                                                  And so federal agencies funding research con-
164         See footnote 132                                                                                      tinue to spend millions of R&D dollars per year          there is insufficient access to data to any of
165         IPv6 AS-level Topology 2008,
166         See footnote 146
167         IPv4 Transfer Policy Proposal (2008-2),
168         “The End of the (IPv4) World is Nigher! “, Jul 2007,
html                                                                                                              185        “The Evolution of U.S. Internet Peering Ecosystem”, Nov 2003,
169         “RFC1744 - Observations on the Management of the Internet Address”, Dec 1994, http://www.faqs.        pdf/norton.pdf
org/rfcs/rfc1744.html                                                                                             186        “Where’s the Money? - Internet Interconnection and Financial Settlements”, Jan 2005, http://isp-
170         “RFC 1744 and its discontents”, 
171         “Financial Incentives for Route Aggregation and Efficient Address Utilization in the Internet”,       187        “My Interview in WorldChanging”, Feb 2008,
1996,                                                      worldchanging.html
172         “‘Address Ownership’ Considered Fatal”, Mar 1995,   188        NITRD,
area.and.wg.reports/ops/cidrd/                                                             189        See footnote 148
173         Media Access Project,                                                     190        DatCat, indexed Internet measurement data,
174         Free Press,                                                                 191        “The Devil and Packet Trace Anonymization”,
175                                                                       pdf
176         “Why You Should Care About Network Neutrality”, May 2006,            192        See footnote 48
177         Eli M. Noam bio,                                                193        The Datapository: A collaborative network data analysis and storage facility, http://www.dataposi-
178         “Beyond Liberalization II: The Impending Doom of Common Carriage”, Mar 1994, http://www.                                                                              194        Deep Internet Performance Zoom (DipZoom),
179         “OPINION: Would the real Network Neutrality please stand up?”, Nov 2005, http://www.telepoca-         195        “Issues and Etiquette Concerning Use of Shared Measurement Data”,                                                                                    2007/papers/imc80.pdf
180         “Net Neutrality’s Quiet Crusader: Free Press’s Ben Scott Faces Down Titans, Regulators in             196        Data Sharing and Anonymization,
Battle Over Internet Control”, Mar 2008,         197        The 1st ACM Workshop on Network Data Anonymization, 2008,
AR2008032703618.html?hpid=sec-business                                                                            nda08.html
181                           198        “IRBs and Security Research: Myths, Facts and Mission Creep”, Mar 2008,
182         “FCC seeks public comment on network neutrality”, Jan 2008,           events/upsec08/tech/full_papers/garfinkel/garfinkel_html/
seeks-public-comment-on-network-neutrality.aspx                                                                   199        “Legal Issues Surrounding Monitoring During Network Research (Invited Paper)”, http://www.
183         “FCC chief backs AT&T bid to waive cost data”, Apr 2008,
mentFilingsNews/idUSN2433101820080424?feedType=RSS&feedName=governmentFilingsNews&rpc=408                         200        See footnote 8
184         “Telecom carriers: ‘Phantom’ voice traffic costing billions”, Apr 2008,
12                                                                                                                                                                                                                            13
these questions. And answering them will come           study them. Recently, a group of well-respected         the government doing any better anytime soon, current legal frameworks229. Although it is well-
at a cost to the social contract of privacy. The        academics have become sufficiently desperate            neither are we in a position to claim the current established that supporting and enforcing these
conversation over how to make these trade-offs          at their inability to study, modify, and share208       lack of governance is working.                        legal frameworks (a tax-funded activity whose
has barely begun201. For one, the academic com-         aspects of the Internet, that they’ve proposed                                                                costs are unknown) does great economic dam-
                                                                                                                G. For the U.S. regulatory agency220 still reeling221
munity is too busy fighting lawsuits202, the great-     building their own sandbox209 to develop and                                                                  age230 while sacrificing privacy and freedom231(not
                                                                                                                from the damage wrought by the 1996 (U.S. Tele-
                                                        test innovative network technologies. It’s like                                                               the best trade citizens have made), Hollywood
                                                                                                                com) act222 and its lifetime employment for law-
                                                        network neutrality at the research layer, an ap-                                                              insists (based on no verified data, natch), that on
                                                                                                                yers223, the opaqueness of the U.S. infrastructure,
                                                        parently irresistible attempt to recover some ob-                                                             the contrary, it’s the sharing of zero marginal cost
                                                                                                                even to them, keeps them in the difficult position
                                                        jectivity210 in the field, but in both cases symptom-                                                         goods that is causing the economic damage232.
                                                                                                                of trying to set policy in the dark. (Ironically the
                                                        atic211 of the need for deeper inquiry212                                                                     While some governments admit they have no in-
                                                                                                                FCC us the agency who should lead solutions
                                                                                                                                                                      terest in tracking kids sharing music233, for-profit
                                                                                                                to this problem, but as mentioned, their behav-
                                                        F. The (predominantly libertarian) engineers                                                                  entities now forced to partner with content pro-
                                                                                                                ior suggests they want as little data as possible224,
                                                        in the router trenches have self-organized into                                                               viders for economic reasons (since as we know
                                                                                                                since they have already made up their mind225
                                                        squadrons of individual engineers213 and ana-                                                                 by now, you can’t maximize profit just moving
                                                                                                                about how to (not) regulate the Internet.)
                                                        lysts214: skilled, bright, principled people who un-                                                          bits around234) will find the temptation235 irresist-
                                                        til recently mostly believed that if they worked        H. Innovative software developers move away ible236.
                                                        hard enough, they could clean up the gutters of         from more oppressive legal frameworks226, the net
                                                                                                                                                                      All these communities have tremendous insights
est incentive yet for universities to not retain data   cyberspace without government intervention.             effect of which is to deprive the country of asso-
                                                                                                                                                                      into pieces of the problem, all are filled with ear-
on network usage. So, while academic research-          Even these groups are now finally acknowledg-           ciated tax revenue and innovative climate.
                                                                                                                                                                      nest people trying to do their job, constrained by
ers do generate203 quite a bit204 of intellectually     ing215 that without better support for protected
                                                                                                                I. Last but most important, the users, the young- their institutional237 context238. But no one has
meritorious205 work206, they are forced to choose       data-sharing216, partnerships with government217,
                                                                                                                est and most progressive of which are embracing oversight for coordination or even articulation of
scientific problems based on what data they can         and more educated law construction and enforce-
                                                                                                                activity227 that is arguably criminal228 under        the global picture239. While the best available da-
manage to scrape together (bottom- up) rather           ment218, even their best efforts plus the market
than picking the most important problems207 to          cannot fix219 the security problems. And although
study and getting the data needed to rigorously         no one currently has positive expectations about
                                                                                                                223        “Untangling the Next Telecom Act”, Mar 2006,
202         “Universities Baffled By Massive Surge In RIAA Copyright Notices”, Apr 2008, http://blog.wired.     224        “FCC chief backs AT&T bid to waive cost data”, Apr 2008,
com/27bstroke6/2008/04/riaa-sends-spik.html                                                                     mentFilingsNews/idUSN2433101820080424?feedType=RSS&feedName=governmentFilingsNews&rpc=408
203         ACM SIGCOMM 2008,                                      225        Written statement of Kevin J. Martin before the U.S. Senate Committee on Commerce, Science and
204         Passive and Active Measurement Conference 2008,                          Transportation, Apr 2008,
205         Internet Measurement Conference (IMC),                                       226        Testimony of Sam Yagan, Sep 2005,
206         IEEE/ACM Transactions on Networking,                                   227        “Japanese File-Sharing Population Explodes”, Dec 2007,
207         “Richard Hamming ‘You and Your Research’”, Mar 1986,         ing-population-explodes-071221/
loads/hamming.pdf                                                                                               228        “95% of Kids Aged 18-24 Are Pirating Music”, 2008,
208         Free Software Foundation,                                                       18+24-are-pirating-music
209         Global Environment for Network Innovations,                                    229
210         Objectivity, Oct 2007,             230        “What’s Wrong With Copy Protection”, Feb 2001,
211         “Misuing Network Neutrality to Eliminate Common Carriage Threatens Free Speech and the Post-        231        Digital Freedom,
al System”, 2006,                              232        “The True Cost of Motion Picture Piracy to the U.S. Economy”, Sep 2006,
212                                                  ipi%5CIPIPublications.nsf/PublicationLookupExecutiveSummary/A6EB1EAC4310AF6F862571F7007CB6AF
213         Castlecops,                                                              233 “Anti file-sharing laws considered”, Oct 2007,
214         MAAW (Messaging Anti-abuse Working Group),                              234        “Broadweave to heal iProvo by shedding wholesale fiber model”, May 2008, http://telephonyonline.
215         “Law Enforcement Engagement & Incident Response Handling: NANOG Engagement”, Feb 2008,              com/fttp/news/broadweave-iprovo-wholesale-0507/                                            235        “AT&T and Other I.S.P.’s May Be Getting Ready to Filter”, Jan 2008, http://bits.blogs.nytimes.
216         Multi-State Information Sharing and Analysis Center,                          com/2008/01/08/att-and-other-isps-may-be-getting-ready-to-filter/index.html
217         “Stopping Spam at its Source: Microsoft Partnered with Government and Law Enforcement in            236        “AT&T Pushes Internet Piracy Filters At CES”, Jan 2008,
003”,                              ATT-Pushes-Internet-Piracy-Filters-At-CES-90827
218         NANOG mailing list: “Re: handling ddos attacks”, May 2004,            237        See footnote 101
nanog/0405/0521.html                                                                                            238        Supercapitalism: The Transformation of Business, Democracy, and Everyday Life, Sep 2008,
219         “Information Security Economics - and Beyond”, 2008,
czech.pdf                                                                                                       239        “Why States Need an International Law for Information Operations”, Jan 2008, http://papers.ssrn.
14                                                                                                                                                                                                                           15
ta makes it obvious that legal repair and renewal        funded network research community’s “real ac-           C. As with most infrastructure issues, the U.S.        infrastructure via public-private partnerships.
is crucial to democracy — communications                 complishment was not in computing but in con-           federal government is slow to respond256 regard-       Local experimentation is critical, and eye-open-
technology being no exception — we are cur-              necting people241” captures a century of thought.       ing a national broadband strategy257. But the          ing: after a decade of pay-per-minute hotspots,
rently pursuing enlightened policy in the dark.          Although the openness of the architecture is the        USG is investing resources and regulatory at-          airports are realizing that free (as in beer) wifi
Which begs the question: what is the most im-            root of its many vulnerabilities, it was also the       tention in the to help foster global Internet          access269 appeals to visitors and residents.
portant ingredient to enlightened policy?                aspect that allowed enough self-organizing mo-          growth, including: encouraging IPv6 deploy-
                                                         mentum to grow the network as fast as it did242.        ment258 to mitigate the coming address space           E. The OECD now considers the Internet rele-
“Such is the irresistible nature of truth that all       The results are noisy243, the journey messy244, the     crunch; improving259 the security of the naming        vant to its mission270 , and is issuing balanced rec-
it asks, and all it wants, is the liberty of ap-         future uncertain245, the most pessimistic scenar-       system260 with community-developed standards261        ommendations based on its best available data,
pearing.” Thomas Paine (1737 - 1809)
                                                         ios246 ominous. But the positive effects are also       for authenticated DNS262 responses; and, in            which they forcefully admit is problematic. In
                                 point nine              incalculable, particularly the potential for an         partnership with industry and academia, de-            their recent ministerial meeting271 on the future
                                                         unprecedented increase in individual freedom247,        veloping a roadmap263 for federal research and         of the Internet, they committed to “improv-
  IX. The news is not all bad: there is a                the often de-emphasized, yet primary, social ob-        development in cybersecurity and information           ing statistical indicators to measure access and
  reason everyone wants to be connected                  jective of both democracy and markets.                  assurance. (Yes, the emphasis is on security and       use of the order to provide more
  to all the world’s knowledge — as well
  as each other — besides its status as                                                                          sustainability issues, but that’s where federal in-    reliable data and analysis.” Only in the U.S. do
                                                         B. The p2p file-sharing248 phenomenon, and
  the most powerful complex system ever                                                                          vestment is today.)                                    policy makers believe that OECD rankings are
                                                         more recently the user-generated video shar-
  created by man. The Internet’s practical                                                                                                                              lying272.
                                                         ing249 phenomenon, are finally shedding some            D. With infrastructure, progressive movement
  promise240 for individual freedom, dem-                light on the inconvenient truth: we have not                                                                   F. There are many educated273 people speaking
  ocratic engagement, and economic em-                                                                           tends to begin at the state and local264 levels as
  powerment, is also unparalleled. This                  yet demonstrated a sustainable competitive              governments265 experiment with266 alternative267       out on the topic of informing policy based on
  promise is sufficient inspiration for an               model for moving raw bits around. Not that we           ownership models268 for provisioning Internet          what we know274, and reserving judgment275 else-
  open, technically literate conversation                excel at competitive models for moving things
  about how to invest in technologies                    around over large distances to almost anywhere.
  and policies to support articulated so-                Witness railroads, water, electricity, highways,        256        “Whatever happened to the Bush broadband policy?”,
  cial objectives.                                                                                               happened-to-bush-broadband.html
                                                         postal service, telephony. Soon, airlines. The
                                                                                                                 257        See footnote 131
                                                         economics clearly need some sunlight250. And            258        See foonote 20
A. David Clark’s conclusion that the federally           the p2p251 debate252 will253 require254 some255.        259        DNSSEC (Domain Name System Security Extensions) Deployment Initiative, http://en.wikipedia.
                                                                                                                 269        DNS security,
240         The Wealth of Networks: How Social Production Transforms Markets and Freedom, Oct 2007,              261        DNS Security Introduction and Requirements,                               262
241         Proceedings of the Twenty-Fourth Internet Engineering Task Force, Jul 1992, http://www3.ietf.        263        Cyber Security and Information Assurance (CSIA),
org/proceedings/prior29/IETF24.pdf                                                                               Supp-Web/TOC%20Pages/08supp-CSIA.pdf
242         Internet growth,            264        “ConnectKentucky’s incumbent-centric model cannot meet US broadband needs”, http://www.
243          “A Look At The Unidentified Half of Netflow”, Jan 2008,
half/missing-half.pdf                                                                                            265        Telecommunications network for Burlington,
244         Histories of the Internet,                                     266        “Philadelphia revives citywide Wi-Fi project” , Jun 2008,
245         The Future of the Internet: Q&A with K. Claffy,   ogyNews/idUSN1737601520080617
246         See footnote 117                                                                                     267        More about the report and analysis of the Philadelphia Wireless project, http://www.saschamein-
247         ”Property, Commons, and the First Amendment: Towards a Core Common Infrastructure”, Mar    
2001,                                                                      268        “City Takes Back Wi-Fi Net It Sold to EarthLink”, Apr 2008,
248                                                        2573C4006938800025742E006429B9.html?ref=t
249                                          269        Guide to WiFi in airports world-wide,
250         “The Internet as emerging critical infrastructure: what needs to be measured?”, Sep 2007, http://    270        OECD proposes roadmap for the future of the Internet economy, Jun 2008,                                                  ocument/18/0,3343,en_2649_37441_40862162_1_1_1_1,00.html
251         “Cisco: P2P Flat in North America? Some Experiencing Major Growth”, Jul 2008, http://www.            271        OECD Ministerial Meeting on the Future of Internet Economy, Jun 2008,                                                        /0,3407,en_21571361_38415463_1_1_1_1_1,00.html
252         “Sandvine: close to half of all bandwidth sucked up by P2P”, Jun 2008,       272        “The UK Broadband Infrastructure And the Debate We Should Be Having”, Jun 2008, http://
253         See footnote 233                                                                                     273        “The Key Questions About the New FISA Bill”, Jun 2008,
254         “The Effect of File Sharing on Record Sales: An Empirical Analysis”, Dec 2006, http://www.unc.       questions-about-new-fisa-bill.html
edu/~cigar/papers/JPE%2031618%20FileSharing%202006-12-12.pdf                                                     274        “Follow Up On Medical Devices: Smarter Devices And Smarter Policy, Not More Bandwidth”,
255         “Challenges and Directions for Monitoring P2P File Sharing Networks or Why My Printer Re-            Apr 2008,
ceived a DMCA Takedown Notice”, Aug 2001,                        275        “Reserving Judgment on Sprint/Clearwire/Google/Intel/ForcesofDarkness Deal”, May 2008,
16                                                                                                                                                                                                                        17
where. (Recommended thinkers276.) There are              ers, led by288 U.S. federal agency ARPA, sup-           other policies298.                                   these signals: abandon this, it is beyond your ca-
evolutionary lessons and insights277 to glean from       ported by strong regulatory protection for inno-                                                             pabilities; redesign this, regardless of short-run
other networked fields facing similar problems,          vation289built the Internet in an amazingly short        “But, important as these problems are, they
                                                                                                                                                                      costs; regulate this, regardless of the imperfec-
e.g., semantic web in big pharma278 and efficient        time relative to the history of communications.          were not the main point. The main point of
                                                                                                                                                                      tions of regulation. But like the operators of
routing279 as well as lessons to draw from ideas         Within twenty years the new ecosystem fatally            the book is to see these human constructions
                                                                                                                                                                      TMI (three-mile island) who could not con-
we have tried that have not worked yet, such as          threatened290 the old. The obvious response by           as systems, not as collections of individuals or
                                                                                                                                                                      ceive of the worst — and thus could not see the
public catalogs280 or open commerce in network           the incumbent carriers was to manipulate the             representatives of ideologies. From our open-
                                                                                                                                                                      disasters facing them — we have misread these
data281. There’s plenty of work to do, but there’s       regulatory architecture away from the line-shar-         ing accident with the coffeepot and job inter-
                                                                                                                                                                      signals too often, reinterpreting them to fit our
no shortage of qualified people.                         ing that made innovations such as the Internet           view through the exotics of space, weapons,
                                                                                                                                                                      preconceptions. Better training alone will not
                                                         possible. No surprise there, these same carriers 291     and microbiology, the theme has been that it is
                                                                                                                                                                      solve the problem, or promise that it won’t hap-
                                                         fought292 innovation293 last294 century295 too,          the way the parts fit together, interact, that is
G. Authors and journalists have captured282 and                                                                                                                       pen again. Worse yet, we may accept the precon-
                                                         including the Internet296. Regulating protection         important. The dangerous accidents liein the
interpreted history283, and academic researchers                                                                                                                      ception that military superiority and private
                                                         of innovation at the edge is neither new nor             system, not in the components. The nature of
have done their share of capturing284 and inter-                                                                                                                      profits are worth the risks. This book’s decoding
                                                         somehow obviated by the technological devel-             the transformation process eludes the capaci-
preting285 the history of communications286 and its                                                                                                                   asserts that the problems are not with individu-
                                                         opments of the Internet. On the contrary, the            ties of any human system we can tolerate in
implications for the Internet. There is detailed                                                                                                                      al motives, individual errors, or even political
                                                         technological ability to innovate at the edge of         the case of nuclear power and weapons; the
understanding of the history of many aspects                                                                                                                          ideologies. The signals come from systems, tech-
                                                         the Internet is easy to remove in the middle by          air transport system works well — diverse in-
of the Internet, including how pieces of the co-                                                                                                                      nological, and economic. They are systems that
                                                         a network owner. So as with the rest of history          terests and technological changes support one
evolving complex systems287 of technology, eco-                                                                                                                       elites have constructed, and thus can be changed
                                                         of telecom, and as with other social goals such          another; we may worry much about the DNA
nomics, and regulation fit together.                                                                                                                                  or abandoned.”
                                                         as universal access, it will largely be a matter of      system with its unregulated reward structure,
                                                                                                                                                                      –Normal Accidents, Charles Perrow, 1999
H. Relatively few government-funded research-            pointing legislatures to results achieved297 from        less about chemical plants; and though the pro-
                                                                                                                  cesses are less difficult and dangerous in min-
                                                                                                                  ing and marine transport, we find the system
                                                                                                                                                                      point ten
                                                                                                                  of each is an unfortunate concatenation of di-          X. Moreover, even in the dim light
277        Evolution of Networks: From Biological Nets to the Internet and WWW, Mar 2003, http://www.             verse interests at cross-purposes. These systems        of the under-attended interdisci-                                          are human constructions, whether designed by            plinary research into the network,
278        “Pharma Researchers Adopt An Orphan Internet Standard”, Oct 2007,           engineers and corporate presidents, or the re-          the available data implies clear di-
cen/html/100807150541.html                                                                                                                                                rections for solutions, all of which
279        “Navigability of Complex Networks”, Sep 2007,                           sult of unplanned, unwitting, crescive, slowly          cross policy-technology boundaries.
280        See footnote 190                                                                                       evolving human attempts to cope. Either way
281        See footnote 194                                                                                       they are very resistant to change. Private
282        Where Wizards Stay Up Late: The Origins Of The Internet, Jan 1998,
Where-Wizards-Stay-Up-Late/dp/0684832674/                                                                         privileges and profits make the planned con-
283        What the Dormouse Said: How the Sixties Counterculture Shaped the Personal Computer Indus-             structions resistant to change; layers upon lay-    A. We can learn from our mistakes. The false as-
try, Feb 2006,                    ers of accommodations and bargains that go
284        Internet history archive,                                                                                               sumption that competing members of a profit-
285        “Licensing in the Web 2.0 Era”, Jul 2007,    by the name of tradition make the unplanned         maximizing ecosystem will cooperate toward ar-
286        “The history of communications and its implications for the Internet”, Jun 2000, http://www.dtc.       ones unyielding. But they are human construc-       chitectural innovations not in their short-term                                                                  tions, and humans can destruct them or recon-
287        “The Telecommunications Economy and Regulation As Coevolving Complex Adaptive Systems:                                                                     interest is remarkably consistent across failed
Implications for Federalism”,                                   struct them. The catastrophes send us warning       attempts to solve299 major problems of the Inter-
288        J.C.R. Licklider,                                               signals. This book has attempted to decode          net300 (e.g., ATM301, multicast302, routing secu-
289        Computer Inquiries,
290        “Economist: How The Internet Killed the Phone Business”, Sep 2005, http://techliberation.
291        History of AT&T,                                            298       “Explaining International Broadband Leadership”, May 2008,
292        Customer Premises Equipment Part 68,                           php?id=142
293        Hush-a-phone case,                                 299       IAB Concerns and Recommendations Regarding Internet Research and Evolution, Aug 2004,
294        Carterfone case,                         
295        AT&T AntiTrust Suit III: Bell Operating Companies,        300       See footnote 62
trust.htm#div                                                                                                    301
296        See footnote 289                                                                                      nology
297        OECD Broadband Portal,,3343,en_2649_34225_38690102_1_1              302

18                                                                                                                                                                                                                   19
rity303, IPv6304, DNSSEC305, QOS306). Engineers         that solving323, or even studying324, the first three   social networks. Ubiquitous connectivity is                nature is what makes transparent335 and account-
have made valiant efforts to provide architectur-       problems325 is no one’s responsibility. Expecting       transforming economic conditions, supporting               able336 experimentation337 so necessary.
al solutions to security and scalability problems,      he private sector to navigate these dimensions          collaborations among individuals that achieve
providing vivid illustrations of how the compu-         (security, scalability, sustainability, and stew-       more efficient328 means of production and con-             C. What we believe about the infrastructure
tational thinking307 approach, embracing modu-          ardship) while subject to relentless pressure to        sumption than either government programs or                influences our technology and policy decisions.
larization and separation of issues, can fail to        minimize costs is a recipe for failure; even pub-       competitive markets have achieved. This trans-             The current barriers to data access leave us with-
account for how tightly linked the technology,          lic-private partnerships are not free of326 these       formation leaves some incongruity about current            out any mechanism to verify claims or weed
economic, and social dimensions of the prob-            pressures327. Furthermore, since all four dimen-                                                                          out false beliefs about the infrastructure,
lems are. As the Internet becomes the substrate         sions transcend the jurisdiction of any sovereign                                                                         including the increasing suspicion that
underlying our professional, personal and politi-       government, we also cannot expect any solution                                                                            the majority of Internet traffic represents
cal lives, we must recognize the links within and       that emphasizes national boundaries.                                                                                      illegal activity. Copyright infringement,
across its four biggest problems:                                                                                                                                                 only one example, may be so rampant as
(1) the fundamentally insecure308 software309           B. While competing in the middle prohibits ar-                                                                            to be economically unviable to pre vent,
ecosystem310, (2) the311 fundamentally312 unscal-       chitectural innovation, cooperating at the edge                                                                           but without an objective338 look at how
able313 routing314 and addressing architecture315,      seems to be a common ingredient of the most                                                                               the network is used, we are subject to vain
(3) the fundamentally316 unsustainable317 eco-          successful innovations on the Internet, includ-                                                                           attempts to criminalize typical network339
nomic318 architecture319, and (4) a stewardship320      ing the web and search engines, VOIP, Linux,                                                                              usage rather than updating the laws to
model broken321 along so many dimensions322             Wikipedia, Ebay, the blogosphere and other                                                                                accomplish their intended purpose340 in
                                                                                                                economic architecture329 for the Internet, which           light of technological developments. Ironically,
303         A Survey of BGP Security Issues and Solutions,     has a deeply embedded preference for markets330            traffic measurement undertaken by law en-
304         Technical and Economic Assessment of Internet Protocol Version 6 (IPv6),   and private sector control331 of communications            forcement for national seurity purposes341 and
ntiahome/ntiageneral/ipv6/draft/draftchap4.htm                                                                  infrastructure332 as well as information333. The           attempted342 by scientific researchers343 is also
305         See footnote 31                                                                                     extremely dynamic and unpredictable structure,             arguably illegal344 under current anachronistic
306         Considerations of Provider-to-Provider Agreements for Internet-Scale Quality of Service (QOS),
Mar 2008,                                                                   usage, and growth of the Internet does not re-             legislation. Again, our choice is to cripple so-
307         Computational Thinking, Mar 2004,           duce the necessity of regulation334 to well-func-          cially important goals — law enforcement and
Wing06.pdf                                                                                                      tioning markets; on the contrary, its elusive              scientific Internet research345 — or update346 the
308         Online Identity Theft, Jun 2008,,3343,en_2649_34223_40830139
309         Economics Malware: Security Decisions, Incentives and Externalities, May 2008, http://www.oecd.
org/dataoecd/53/17/40722462.pdf                                                                                 328        “Sharing Nicely”, 2004,
310         Daily Submission Volume, Aug 2008,                             329
311                                                                     have-learned-so-far/
312         “Scalability of Routing: Compactness and Dynamics”,   330        “Let There Be Markets: The Evangelical Roots of Economics”, May 2005,
RRG-3.pdf                                                                                                       Industry/2005/Evangelical-Economics1may05.htm
313         See footnote 89                                                                                     331        Administration NII Accomplishments,
314         “Damping BGP”, Jun 2007,                         332        “The 700 MHz Auction as the Next Front In the Cable/Telco War”, May 2007, http://www.wetma-
315         “Report from the IAB Workshop on Routing and Addressing”, Sep 2007,
rfc4984.txt                                                                                                     333        Intellectual Property and the National Information Infrastructure,
316         “Broadband pricing: solutions that are orthogonal to any real problem”, Jun 2008, http://www.       fices/com/doc/ipnii/                                           334        Understanding Regulation: Theory, Strategy, and Practice, Jul 1999,
317         See footnote 181                                                                                    derstanding-Regulation-Theory-Strategy-Practice/dp/0198774389/
318         “The (un)Economic Internet”,                 335        Sunlight Foundation,
319         “$200 Billion Broadband Scandal”,                  336        “Harnessing Conflict and Competitiveness for Society’s Benefit”, Aug 2000, http://www.davidbrin.
320         See footnote 39                                                                                     com/disputationarticle1.html
321         ICANN - The Case for Reform, Feb 2002,     337        GNU General Public License, Jun 2007,
feb02.htm                                                                                                       338        See footnote 210
322         The Broadband Problem: Anatomy of a Market Failure and a Policy Dilemma, Jun 2004, http://          339                                         340        See footnote 229
323         See footnote 160                                                                                    341        U.S. Code Collection: Chapter 36-Foreign Intelligence Surveillance, Aug 2008,
324         See footnote 62                                                                           
325         IPv4 Census Map,                           342
326         “Connected Nation’s Private Interests Hit In FCC Comments”, Jul 2008, http://www.publicknowl-       343, 344 See footnote 199                                                                                              345        “Why We Don’t Know How To Simulate The Internet”, Dec 1997,
327    ps
rection/                                                                                                        346        See footnote 8
20                                                                                                                                                                                                                              21
relevant347 communications348 privacy349 (EC- E. Scientific researchers are in a difficult posi-               organization can share more data from its re- and supporting scientific research balanced
PA350) legislation.                                tion, trying to do science without data355, but-            search backbone with scientific researchers.       carefully against individual privacy. Such a trade
                                                   they are in a position to make progress, with the                                                              seems less extreme an idea in light of the failure
D. Public investment in knowledge produc- help of a few good legal experts. They (we) could                    F. The FCC is not exempt from the facts either of the D-block auction, and the FCC admission
tion, including science and medical research, propose a list of the most important Internet re-                — the agency should be pursuing empirically that economic conditions make it a bad time to
gains enormously from universal connectivity, search questions/problems to investigate, such                   grounded validation362 of the claimed efficiency try to auction it now365. Reforming our policy
offering distribution of resulting products to as the ongoing discrepancies356 between suppos-                 of its own policies, even if it requires trading for this spectrum could achieve efficiency366, ac-
all taxpayers at zero marginal cost. The same edly scientific studies357, and suggest what data is             temporary spectrum unlicensing364 as an experi- cess367, public safety, and network science objec-
reasoning reduces the justification for strong needed to investigate them. The academic com-                   ment to gather realistic baseline data on wireless tives at least cost to taxpayers.
intellectual property systems, since they require munity could even use existing assets358 such as             network behavior to policy makers. The aca-
expensive technology to prevent networks from their own under-utilized backbone359 to mitigate                 demic community could even help design such a “We can have facts without thinking but we can-
doing what networks do naturally: share data. It the data dearth, by incenting measurement data                network, geared toward public safety objectives not have thinking without facts.” — John Dewey
is thus in interest of taxpayers for governments out of cooperating networks in exchange for
to promote and sometimes directly fund351 uni- network bandwidth. In the process they could
versal deployment of network infrastructure. help local communities experiment with and
More generally, government needs to prevent measure performance, cost, and efficiency of al-
monopoly control over essential resources352, ternative network ownership models. Internet2
mandate collection of traffic reports from ISPs353 should also work with researchers and their in-
to validate their claims, be a better role model stitutional review boards360 (IRBs) at member
for operational security, and coordinate the de- universities to assist researchers in developing
velopment of a road map for Internet security privacy-respecting361 network analysis technolo-
similar to that of the energy sector354 (DHS is gies and data handling policies, so that the
working on this last one).

347          U.S. Code Collection: 2511. Interception and disclosure of wire, oral, or electronic communica-
tions prohibited, Aug 2008,
348          U.S. Code Collection: Unlawful access to stored communications, Aug 2008,
349          “Privacy Act II”, Jul 2008,
350          Electronic Communications Privacy Act,
351          Broadband Opportunities for Rural America,
352          The Essential Facilities Doctrine In The Deregulated Telecommunications Industry, 1998, http://
353          Internet 3.0: Identifying Problems and Solutions to the Network Neutrality Debate, Jun 2007,      362 1st ACM Workshop on Network Data Anonymization, 2008,                                                     html
354          Roadmap to Secure Control Systems in the Energy Sector, January 2006, http://www.controlsys-      363        How To Give America Wireless Broadband For Christmas 2009 - the Lesson from 3.65 GHz De-                                                                                               ployment, Jan 2008,
355          See footnote 342                                                                                  364        “FCC chief says economy could hurt wireless sale”, Jan 2008,
356          Internet2 and Quality of Service: Research, Experience, and Conclusions, May 2006, https://www.   nologyNews/idUSN1555479420080115                                                                        365        “Gaping Hole in Models for Using Spectrum Efficiently”, Mar 2008,
357          See point five                                                                                    posts/832812_models_wireless_spectrum/
358          High-speed academic networks and the future of the Internet, May 2008,    366        Report: US must have universal gigabit broadband by 2015, Jun 2008,
articles/culture/future-internet.ars                                                                           news.ars/post/20080624-report-us-must-have-universal-gigabit-broadband-by-2015.html
360          “IRBs and Security Research: Myths, Facts and Mission Creep”, Mar 2008, http://www.usenix.
361          1st ACM Workshop on Network Data Anonymization, 2008,

22                                                                                                                                                                                                                      23
                                  BY OUR

     This pamphlet was produced as part of the COMMONS, a CAIDA and Cisco collaborative
     project, to simultaneously solve three acute and growing problems facing the Internet: a self-
     reported financial crisis in the Internet infrastructure provider industry; a data acquisition cri-
     sis which has severely stunted the field of network science; and a struggle for survival within
     emerging community and municipal networks.

                               500 GILMAN DRIVE, MC0505. LA JOLLA, CA 92093-0505.

24                                                                                                         24

Shared By: