idnits 2.17.1 draft-arkko-iab-internet-consolidation-02.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- ** The document seems to lack a Security Considerations section. ** The document seems to lack an IANA Considerations section. (See Section 2.2 of https://www.ietf.org/id-info/checklist for how to handle the case when there are no actions for IANA.) == There are 1 instance of lines with non-RFC6890-compliant IPv4 addresses in the document. If these are example addresses, they should be changed. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (July 09, 2019) is 1752 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- No issues found here. Summary: 2 errors (**), 0 flaws (~~), 2 warnings (==), 1 comment (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group J. Arkko 3 Internet-Draft Ericsson 4 Intended status: Informational B. Trammell 5 Expires: January 10, 2020 ETH Zurich 6 M. Nottingham 7 Fastly 8 C. Huitema 9 Private Octopus Inc. 10 M. Thomson 11 Mozilla 12 J. Tantsura 13 Apstra, Inc. 14 N. ten Oever 15 University of Amsterdam 16 July 09, 2019 18 Considerations on Internet Consolidation and the Internet Architecture 19 draft-arkko-iab-internet-consolidation-02 21 Abstract 23 Many of us have held a vision of the Internet as the ultimate 24 distributed platform that allows communication, the provision of 25 services, and competition from any corner of the world. But as the 26 Internet has matured, it seems to also feed the creation of large, 27 centralised entities in many areas. This phenomenon could be looked 28 at from many different angles, but this memo considers the topic from 29 the perspective of how available technology and Internet architecture 30 drives different market directions. 32 Status of This Memo 34 This Internet-Draft is submitted in full conformance with the 35 provisions of BCP 78 and BCP 79. 37 Internet-Drafts are working documents of the Internet Engineering 38 Task Force (IETF). Note that other groups may also distribute 39 working documents as Internet-Drafts. The list of current Internet- 40 Drafts is at http://datatracker.ietf.org/drafts/current/. 42 Internet-Drafts are draft documents valid for a maximum of six months 43 and may be updated, replaced, or obsoleted by other documents at any 44 time. It is inappropriate to use Internet-Drafts as reference 45 material or to cite them other than as "work in progress." 47 This Internet-Draft will expire on January 10, 2020. 49 Copyright Notice 51 Copyright (c) 2019 IETF Trust and the persons identified as the 52 document authors. All rights reserved. 54 This document is subject to BCP 78 and the IETF Trust's Legal 55 Provisions Relating to IETF Documents 56 (http://trustee.ietf.org/license-info) in effect on the date of 57 publication of this document. Please review these documents 58 carefully, as they describe your rights and restrictions with respect 59 to this document. Code Components extracted from this document must 60 include Simplified BSD License text as described in Section 4.e of 61 the Trust Legal Provisions and are provided without warranty as 62 described in the Simplified BSD License. 64 1. Introduction 66 Many of us have held a vision of the Internet as the ultimate 67 distributed platform that allows communication, the provision of 68 services, and competition from any corner of the world. But as the 69 Internet has matured, it seems to also feed the creation of large, 70 centralised entities in many areas. 72 We use the term Internet consolidation to refer to the process of the 73 increasing control over Internet infrastructure and services by a 74 small set of organizations. Such concentration has an obvious effect 75 on traffic flows or on services and systems that are daily used by a 76 large population of Internet users. However, it can also create 77 secondary effects, where the ability to collect information or to 78 affect something is concentrated in that small set of organizations. 80 Consolidation may also affect technology choices and the evolution of 81 the Internet architecture. For example, large organizations or 82 organizations providing important technology components may have a 83 significant impact on what technology is deployed for large numbers 84 of users or by other organizations. 86 Our first question is whether the Internet is indeed consolidating. 87 It certainly appears so, but more quantitative research on this topic 88 would be welcome. It is also possible that there is only a 89 perception of consolidation, as market forces have caused business 90 changes in new areas of business. Arguably, today's consolidation 91 areas seem to be more in the application layer than further down in 92 the stack or in operating systems, as was the case some years ago. 93 The second question is if we are seeing consolidation simply moving 94 to new areas. 96 This phenomenon could be looked at from many different angles, but 97 this memo considers the topic from the perspective of how available 98 technology and Internet architecture drives different market 99 directions. Our third question is if the Internet technology has 100 influenced the consolidation trends in some manner. And conversely, 101 the fourth question asks how Internet consolidation is influencing 102 the development of the Internet infrastructure and architecture. 104 The engineering remit at the IETF is to focus on technology, but of 105 course we also want to understand the implications and externalities 106 of the technical arrangements we design. Technology affects 107 economics and vice versa. The Internet technology community 108 continues to make decisions that have ramifications on Internet 109 systems, just as we are subject to forces that affect them. 111 As technologists, our fourth question is whether there are changes in 112 technology that would help reduce those large-player advantages that 113 are technically-driven. 115 This memo reviews areas where consolidation may be occurring in the 116 Internet, and discusses the potential reasons for this. The memo 117 starts by reviewing other work in this area in Section 2. Section 3 118 discusses consolidation and the reasons behind the creation of larger 119 entities, and Section 4 looks at some actions that might alleviate 120 the situation. 122 If you are interested on this or other architecture-related topics, 123 please subscribe to the IAB architecture-discuss mailing list as one 124 forum for discussion. Similarly, the Internet Society has chosen 125 consolidation as a focus topic for their year 2019 activities. Their 126 report is in [ISOC]. 128 2. Other Work 130 One of the causes for the current consolidation of the Internet 131 infrastructure can be traced back to some of the assumptions that 132 were made during the commercialization of the Internet in the early 133 1990s [Abbate], despite [RFC1192] describing some potential issues 134 that could arise. Overall it was expected the combination of 135 commercialization, together with the technical and architectural 136 characteristics of the Internet, such as its modularity and layering 137 principles, would lead to perfect markets, free competition and 138 decentralized structures [LitanRivlin]. 140 But as we know now, this did not happen entirely as expected. Some 141 even argue that 'market concentrations, control and power struggles 142 are categories to adequately describe the fundamental dynamics of the 143 commercial Internet' [DolataSchrape]. While the privatization was 144 supposed to lead to competition and innovation [Cowheyetal] 145 [VanSchewick], some argue that it actually led to the emergence of 146 Internet oligopolies [Mansell] [Smyrnaios]. 148 Current scientific economic thinking harbors two different schools of 149 thought vis-a-vis efficient markets and monopolies. The school of 150 thought based on Adam Smith argues that unfettered markets tend to 151 concentration of wealth and income, whereas liberal economists 152 believe in efficient markets that stimulate competition. 154 On the other hand, according to Joseph Stiglitz, 'many sectors - 155 telecommunications, cable TV, digital branches from social media to 156 Internet search, health insurance, pharmaceuticals, agro-business, 157 and many more - cannot be understood through the lens of competition' 158 [Stiglitz]. The considerations of technologists and policy makers at 159 the time of the commercialization and privatization of the Internet 160 infrastructure might have been based on a belief in efficient 161 markets, whereas we are now finding out this might not always be how 162 markets function. 164 Recently there is a growing body of literature that the currently 165 observed consolidation into oligopolies and monopolies can be 166 described as a failure of economic policy, which could be addressed 167 with revamped, or improved anti-trust policies [Wu] [Khan]. On the 168 other hand there are those who criticize these proposals for their 169 economic determinism; merger reviews, company break-ups and 170 'trustbusting' do not necessarily change the structure of a market. 171 Technology might actually have a role to play in this as well. The 172 IETF in specific, and the Internet governance regime complex [Nye] in 173 general, has been designed as a distributed arrangement to prevent 174 capture of the infrastructure by a single interest group or actor. 175 Where power or control was centralized, specific governance 176 arrangements were put into place to counter the centralization of 177 power. 179 It cannot be denied that 'market actors have contributed immensely to 180 the evolution of the Internet in terms of investment, products, 181 services, and infrastructure, and the government's light-handed 182 approach to regulation has given producers and consumers substantial 183 freedom to innovate and to self-regulate with respect to many issues 184 affecting the Internet community in ways that have produced 185 substantial social benefits' [Frischmann]. But the current 186 consolidation in ownership of and control over the Internet 187 infrastructure was not foreseen [Clark], and arguably the loss of 188 decentralized control goes against its design objectives. For 189 instance, [RFC1958] says: 191 This allows for uniform and relatively seamless operations in a 192 competitive, multi-vendor, multi-provider public network. 194 and 196 Heterogeneity is inevitable and must be supported by design. 198 And [RFC3935] says: 200 We embrace technical concepts such as decentralized control, edge- 201 user empowerment and sharing of resources, because those concepts 202 resonate with the core values of the IETF community. 204 3. Factors Driving Consolidation 206 Consolidation is driven by economic factors relating to scale and 207 ability to easily reach a large market of users over the Internet. 208 This kind of setting tends to enable winners to take large market 209 shares, whether those winners came about through the model that 210 liberal economists believe in or the model that Adam Smith believes 211 in. 213 The most visible aspects of this involve well-recognized Internet 214 services. The Internet Society's report summarised the market 215 position of popular Internet service brands as follows [ISOC]: 217 o Facebook and Google have been estimated to account for 84% of 218 global digital advertising investment (excluding China). 220 o Amazon is expected to account for 49.1% of all online retail 221 spending in the US. Similarly, Alibaba is estimated to have close 222 to 60% of the e-commerce market in China. 224 o Google alone holds 90% of the global search market, over 60% of 225 web browsers, and has the number 1 (by far) mobile operating 226 system (Android). 228 o Facebook - incorporating Facebook Messenger, WhatsApp, and 229 Instagram - holds 4 of the world's top 6 social media platforms. 231 But it is important to recognize that the Internet is a complex 232 ecosystem. There are many underlying services whose diversity, or 233 lack thereof, are as important as that of, say, consumer-visible 234 social networks. For instance, the diversity of cloud services, 235 operating systems, and browser engines is as important as that as of 236 application stores or the browsers themselves. 238 Of course, the Internet allows plenty of choice in these and other 239 areas. Too many or too few choices create different kinds of 240 problems. 242 It would be useful to break these general factors and observations 243 down a bit further. In particular, it is useful to distinguish 244 market or economic factors from technical factors. 246 3.1. Economics 248 Scaling benefits are natural for many types of businesses. And many 249 Internet-based businesses can potentially serve a very large customer 250 base, as the cost of replicating and delivering their service to new 251 customers or areas is small. 253 However, typically the network effect has an even more pronounced 254 impact. Each additional user adds to the value of the network for 255 all users in a network. In some applications, such as the open web, 256 this value grows for everyone, as the web is a globally connected, 257 interoperable service for anyone with a browser can use. 259 There is an important distinction between different applications of 260 the network effect, however. Consider email as another example; 261 anyone with an account at any email server can use it globally. 262 However, here we have seen much more consolidation into few large 263 email providers, both due to innovative, high-quality services but 264 also because running email services by small entities is becoming 265 difficult; among other things due to spam prevention practices that 266 tend to recognize well only the largest entities. 268 In some other applications, such as social media, the services have a 269 more closed nature. The value of being a customer of one social 270 media service depends highly on how many other customers that 271 particular service has. Hence, the larger the service, the more 272 valuable it is. And the bigger the value difference to the 273 customers, the less practical choice they have in selecting a 274 service. 276 In some cases, these developments also allow asymmetric relationships 277 to form, with the customers having less ability to affect the service 278 than they would perhaps wish. 280 3.2. Interdependencies 282 Entities with a large role in a market place tend to have inertia, of 283 course through having many customers, but also due to their 284 interconnectedness with the rest of the ecosystem. These 285 interconnections can range from business relationships to peering 286 connections, linking, and the use of various infrastructure services 287 from an entity as building blocks in applications. 289 These interconnections make it difficult for a market to move away 290 from a large entity. Indeed, even for an individual it can be a 291 challenge not to use the most commonly used Internet services 292 [Gizmodo]. 294 Optimistic technologists ("digital libertarians") tend to believe 295 that states have limited ability to regulate the Internet: "The Net 296 interprets censorship as damage and routes around it" [Gilmore]. 297 However, as argued by [Boyle] states may have multiple ways to 298 influence and monitor the Internet. One of the issues related to 299 consolidation is that it tends to be easier to exert control of few 300 large entities, than a large set of small, distributed players. This 301 concern is particularly acute around intellectual property rights or 302 surveillance capabilities, particularly when extra-territorial 303 requirements are placed on the large entities. These entities cannot 304 avoid comforming to regulation and laws in any of the locations they 305 have presence in. 307 As a result, there's an added angle of interconnectedness with 308 governments. At the same time, this of course also provides an 309 avenue for control of market forces, e.g., in the form of competition 310 law. 312 3.3. Data- and Capital-intensive Services 314 The scaling advantages are only getting larger with the advent of AI- 315 and machine learning-based technologies. 317 The more users a service has, the more data is available for training 318 machine learning models, and the better the service becomes, bringing 319 again more users. This feedback loop and the general capital- 320 intensive nature of the technology (data and processing at scale) 321 makes it likely that the largest companies are ahead in the use of 322 these technologies. 324 One could also take the pessimistic view that many of the ongoing 325 disputes in standards organizations relate to which market actors 326 will ultimately be able to collect the more data from private persons 327 and how. The question isn't as much about the protection of these 328 persons' privacy but rather whether some industry (or country) will 329 be able to benefit from access to data. 331 3.4. Permissionless Innovation 333 The email vs. social media example also highlights the interesting 334 roles of interoperability and the "permissionless innovation" 335 principle - the idea that a network can be simple but still powerful 336 enough that essentially any application could be built on top of it 337 without needing any special support from anyone else. Permissionless 338 innovation has brought us all the innovative applications that we 339 enjoy today, on top of a highly interoperable underlying network, 340 along with advances in video coding and other techniques used by 341 applications. 343 Paradoxically, if the underlying network is sufficiently powerful, 344 the applications on top can evolve without similar pressures for 345 interoperability, leading to the closed but highly valuable services 346 discussed above. We call this the Permissionless Completeness 347 Problem. 349 3.5. Fundamentals of Communication 351 There are also fundamental issues, such as the speed of light. Low- 352 latency services can fundamentally only be provided through globally 353 distributed data centers. These are often provided and/or built by 354 large organizations, although collaborative data center or cloud 355 computing service approaches also exist. 357 A similar issue has arisen in recent years around large-scale denial- 358 of-service attacks, and how various entities can deal with them. 359 While the largest attacks affect all players (see, for instance, the 360 Dyn attacks in October 2016), it is also true that large cloud- and 361 content-delivery providers can better deal with such attacks due to 362 their scale. This is one reason that attracts many network services 363 to such providers. 365 3.6. Technology Factors 367 One of the key questions is whether we are seeing developments that 368 are driven by economic factors or whether fundamental reasons or lack 369 of available technology drives particular models. For instance, 370 centralised solutions might be desirable due to business incentives, 371 or they might be necessary because there is no distributed, 372 collaborative solution. 374 Some technical issues have historically not been easy to solve, such 375 as e-mail spam, which has led to reliance on non-technical solutions. 376 Today, it is becoming increasingly difficult to run your own mail 377 services, essentially forcing many organizations and individuals to 378 employ larger providers. The issues relate directly to size of 379 entities; no one can afford to disconnect from the largest providers. 380 But as a small entity, there is little leverage to convince peer 381 entities or various supporting white/blacklist entities to deal with 382 you properly. 384 Many Internet services are based on gathering data about users, and 385 using that data for targeted advertisements. More data from more 386 users makes it possible to run a service more accurately or with 387 better results; here again scale brings advantages. 389 Another trend is that more and more content is becoming available 390 locally, from a content delivery or provider function directly on 391 one's own ISP network. We predict that eventually most content will 392 be delivered this way, reducing the role that global IP connectivity 393 across the Internet plays. By some metrics this has already 394 happened; what practical - positive or negative - impacts might this 395 have on the Internet technology? 397 There are also security tradeoffs. Large entities are generally 398 better equipped to move to more recent and more secure technology. 399 For instance, the Domain Name System (DNS) shows signs of ageing but 400 due to the legacy of deployed systems, has changed very slowly. 401 Newer technology developed in accordance with IETF standards enables 402 DNS queries to be performed confidentially, but its deployment is 403 happening mostly in browsers that use global DNS resolver services, 404 such as Cloudflare's 1.1.1.1 or Google's 8.8.8.8. This results in 405 faster evolution and better security for end users. 407 However, if one steps back and considers the overall security effects 408 of these developments, the resulting effects can be different. While 409 the security of the actual protocol exchanges improves with the 410 introduction of this new technology, at the same time this implies a 411 move from using a worldwide distributed set of DNS resolvers into, 412 again, more centralised global resolvers. While these resolvers are 413 very well maintained (and a great service), they are potentially 414 high-value targets for pervasive monitoring and Denial-of-Service 415 (DoS) attacks. In 2016, for example, DoS attacks were launched 416 against Dyn, one of the largest DNS providers, leading to some 417 outages. 419 4. Call to Action 421 Are there assumptions about the Internet architecture that no longer 422 hold in a world where larger, more centralised entities provide big 423 parts of the Internet service? If the world changes, the Internet 424 and its technology/architecture may have to match those changes. 426 It appears that leveling the playing field for new entrants or small 427 players brings potential benefits. Are there technical solutions 428 that are missing today? 430 Of course, it may well be that technology improvements are hard to 431 come by. Nevertheless, recognizing the risks of consolidation in 432 both current and proposed future technologies is the first step in 433 proactively avoiding those risks where possible. 435 Assuming that one does not wish for regulation, technologies that 436 support distributed architectures, open source implementations of 437 currently centralised network functions, or helping increase users' 438 control can be beneficial. Federation, for example, would help 439 enable distributed services in situations where smaller entities 440 would like to collaborate. 442 Similarly, in an asymmetric power balance between users and services, 443 tools that enable the user to control what information is provided to 444 a particular service can be very helpful. Some such tools exist, for 445 instance, in the privacy and prevention-tracking modes of popular 446 browsers, but why are these modes not the default, and could we 447 develop them further? 449 It is also surprising that in the age of software-defined everything, 450 we can program almost anything else except the globally provided, 451 packaged services. Opening up interfaces would allow the building of 452 additional, innovative services, and better match users' needs. 454 Silver bullets are rare, of course. Internet service markets 455 sometimes fragment rather than cooperate through federation. And the 456 asymmetric power balances are easiest changed with data that is user- 457 controlled, but it is much harder to change when someone else holds 458 it. Nevertheless, the exploration of solutions to ensure the 459 Internet is kept open for new innovations and in the control of users 460 is very important. 462 o What IETF topics that should be pursued to address some of the 463 issues around consolidation? Some of the topics for consideration 464 are discussed in Section 4.1 and Section 4.2. 466 o Are there open source efforts that should be pursued or continue 467 to be pursued to ensure that a diversity of operators and networks 468 can use a particular technology? This is further discussed in 469 {{opensource]]. 471 o What measurements relating to the development of centralization or 472 consolidation should be pursued? And what other research, such as 473 distributed Internet architectures, should be driven forward? 474 Some potential topics are discussed in Section 4.4. 476 4.1. Open Interfaces 478 Standards and open source efforts continue to build many open 479 interfaces and APIs that allow systems interoperability and tailoring 480 of services. In some cases, however, the availability of open 481 interfaces definitions and software has not led to the realization of 482 actual interfaces in this open manner. For instance, different 483 instant messaging systems have had a technical ability to be 484 interoperable with other systems, just like e-mail is interoperable 485 across systems, but have chosen to be disconnected. 487 Work in determining what open interfaces can provide benefits to 488 users as well be successfully deployed in the Internet ecosystem 489 would of course be useful. 491 4.2. Specific Standardization Choices 493 Sometimes the issue is not the availability of interfaces as such, 494 but rather fundamental architectural choices with regards to how 495 Internet systems should be built. 497 Often this relates to how centralized or distributed deployments are 498 targeted. And even if a distributed, broad deployment model is 499 targeted, expectations may not match reality when economies of scale, 500 DDoS resilience, market consolidation, or other factors have come 501 into play. These factors can result in the deployed reality being 502 highly concentrated. 504 This is a serious issue for the Internet, as concentrated, 505 centralized deployment models present risks to user choice, privacy, 506 and future protocol evolution. On occasion, the differences to 507 expectations can be immediate, but can also occur much later. 509 Some examples of these issues include current work in DNS where we 510 may see concentration through the use of globally available common 511 resolver services, which evolve rapidly and can offer better 512 security. But the concentration of these queries into a few services 513 creates new security and privacy concerns. 515 Another example is email, which started out as many providers running 516 in a largely uncoordinated fashion, but which has since then seen 517 significant market consolidation and a need for coordination to 518 defend against attacks such as spam. The coordination and 519 centralized defense mechanisms scale better for large entities, which 520 has fueled additional consolidation. 522 Awareness of these issues while working on standards would be useful, 523 so that the issues can be taken into account and appropriately 524 mitigated. To begin with, those of us interested in the broader 525 questions about Internet research and development are needed to 526 identify some of the effects that new technology developments may 527 pose. 529 4.3. Open Source 531 Many key Internet infrastructure services (e.g., DNS servers), end- 532 user applications (e.g., browsers) and technology components (e.g., 533 operating systems or protocol implementations) are commonly 534 implemented using open source solutions. 536 This is often true even if there is a large entity that is in charge 537 of a large fraction of development and deployment for a particular 538 technology. This is a good thing, as it provides a means for others 539 to have a fair chance of changing the technology in question, should 540 the large entity drive their use of the technology in some direction 541 that does not benefit the users. For instance, users and other 542 organizations have traditionally been able to either run their own 543 browser versions or provided extensions that suit their needs better 544 than the default system. 546 As a result, continuing to have this ecosystem is an important safety 547 valve and competition opportunity. 549 4.4. Research Challenges 551 There are a number of different research directions for which further 552 work would be useful. 554 The first is about measurements; is the Internet indeed 555 consolidating, and if so, by how much, and in what aspects? Also, 556 where are the Internet's traffic flows concentrating, and how is this 557 changing over time? 559 There are also questions about the trends and their relationships to 560 technology: Has Internet technology influenced the consolidation 561 trends in some manner? And conversely, how does Internet 562 consolidation influence the development of Internet infrastructure 563 and architecture? 565 Finally, research on topics that would likely yield results that 566 increase the diversity and de-centralization in the Internet would 567 obviously be welcome. This can include any de-centralized 568 technologies, but in particular distributed Internet architecture is 569 an interesting topic. 571 5. Contributors 573 Much of the text in this memo is from a blog article written by Jari 574 Arkko, Mark Nottingham, Christian Huitema, Martin Thomson, and Brian 575 Trammell for the Internet Architecture Board (IAB), and from a blog 576 article written by Jari Arkko and Brian Trammell APNIC and RIPE. 577 Some parts of the text have also come from a future workshop 578 description developed in the IAB, primarily by Christian Huitema and 579 Ted Hardie. 581 6. Acknowledgements 583 The authors would like to thank IAB members, Geoff Huston, Amelia 584 Andersdotter, Gonzalo Camarillo, Mirjam Kuehne, Robert Mitchell, Olaf 585 Kolkman, Greg Skinner and many others for interesting discussions in 586 this problem space. The authors would also like to thank all 587 participants in the 2019 Design Expectations vs. Deployment Reality 588 (DEDR) IAB workshop held in Kirkkonummi, Finland, as well as the 589 participants in the 2019 EuroDIG workshop on "Internet consolidation 590 - opportunities and challenges". 592 7. Informative References 594 [Abbate] Abbate, J., "Inventing the Internet", MIT Press , 2000, 595 . 597 [Boyle] "Foucault in Cyberspace: Surveillance, Sovereignty, and 598 Hardwired Censors", Duke Law School, 599 https://scholarship.law.duke.edu/ 600 faculty_scholarship/619/ , 1997. 602 [Clark] Clark, D., "The Design Philosophy of the DARPA Internet 603 Protocols", In Symposium Proceedings on Communications 604 Architectures and Protocols, 106-114. SIGCOMM '88. New 605 York, NY, USA, ACM https://doi.org/10.1145/52324.52336 , 606 1988. 608 [Cowheyetal] 609 Cowhey, P., Aronson, J., and J. Richards, "Shaping the 610 Architecture of the US Information and Communication 611 Technology Architecture: A Political Economic Analysis", 612 Review of Policy Research 26 (1-2), pp. 105-125. , 2009. 614 [DolataSchrape] 615 Dolata, U. and J. Schrape, "Collectivity and Power on the 616 Internet: A Sociological Perspective", Springer 617 International Publishing. Page 85. , 2018, 618 . 620 [Frischmann] 621 Frischmann, B., "Privatization and Commercialization of 622 the Internet Infrastructure", Colum. Sci. & Tech. L. Rev. 623 2, pp. 1-25. , 2001. 625 [Gilmore] "First Nation in Cyberspace", TIME International (see 626 http://kirste.userpage.fu-berlin.de/outerspace/ 627 internet-article.html) , December 1993. 629 [Gizmodo] "I Cut The 'Big Five' Tech Giants From My Life. It Was 630 Hell", Gizmodo, https://www.gizmodo.com.au/2019/02/i-cut- 631 the-big-five-tech-giants-from-my-life-it-was-hell/ , 632 February 2019. 634 [ISOC] "Consolidation in the Internet economy", Internet Society, 635 https://future.internetsociety.org/2019/ , 2019. 637 [Khan] Khan, L., "Amazon's Antritrust Paradox", The Yale Law 638 Journal 126:710 , 2017. 640 [LitanRivlin] 641 Litan, R. and A. Rivlin, "Projecting the Economic Impact 642 of the Internet", American Economic Review 91 (2), pp. 643 313-317 , 2001. 645 [Mansell] Mansell, R. and M. Javary, "Emerging Internet Oligopolies: 646 A Political Economy Analysis", In Miller, Edythe S.and 647 Samuels, Warren J., (eds.) An Institutionalist Approach to 648 Public Utilities Regulation. Michigan State University 649 Press, East Lansing, Michigan, pp. 162-201. ISBN 650 9780870136245 , 2002, . 653 [Nye] Nye, J., "The Regime Complex for Managing Global Cyber 654 Activities. Global Commission on Internet Governance", 655 CIGI, Global Commission on Internet Governance , 2014. 657 [RFC1192] Kahin, B., "Commercialization of the Internet summary 658 report", RFC 1192, DOI 10.17487/RFC1192, November 1990, 659 . 661 [RFC1958] Carpenter, B., Ed., "Architectural Principles of the 662 Internet", RFC 1958, DOI 10.17487/RFC1958, June 1996, 663 . 665 [RFC3935] Alvestrand, H., "A Mission Statement for the IETF", 666 BCP 95, RFC 3935, DOI 10.17487/RFC3935, October 2004, 667 . 669 [Smyrnaios] 670 Smyrnaios, N., "Internet Oligopoly: The Corporate Takeover 671 of Our Digital World", Emerald Publishing Ltd. , 2018. 673 [Stiglitz] 674 Stiglitz, J., "Joseph Stiglitz: Are markets efficient, or 675 do they tend towards monopoly? The verdict is in", 2016, 676 . 680 [VanSchewick] 681 van Schewick, B., "Internet Architecture and Innovation", 682 MIT Press , 2012. 684 [Wu] Wu, T., "The Curse of Bigness", Columbia Global Reports, 685 New York, New York , 2018. 687 Authors' Addresses 689 Jari Arkko 690 Ericsson 692 Email: jari.arkko@piuha.net 694 Brian Trammell 695 ETH Zurich 697 Email: ietf@trammell.ch 699 Mark Nottingham 700 Fastly 702 Email: mnot@mnot.net 704 Christian Huitema 705 Private Octopus Inc. 707 Email: huitema@huitema.net 709 Martin Thomson 710 Mozilla 712 Email: martin.thomson@gmail.com 713 Jeff Tantsura 714 Apstra, Inc. 716 Email: jefftant.ietf@gmail.com 718 Niels ten Oever 719 University of Amsterdam 721 Email: mail@nielstenoever.net