idnits 2.17.1 draft-irtf-hrpc-research-11.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 abstract seems to contain references ([RFC6973]), which it shouldn't. Please replace those with straight textual mentions of the documents in question. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year == Line 252 has weird spacing: '...locking the p...' == Line 258 has weird spacing: '...sorship techn...' == Line 315 has weird spacing: '...ltering the p...' -- The document date (February 25, 2017) is 2616 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- -- Looks like a reference, but probably isn't: '1' on line 3446 -- Looks like a reference, but probably isn't: '2' on line 3448 -- Looks like a reference, but probably isn't: '3' on line 3450 == Unused Reference: 'RFC4033' is defined on line 3174, but no explicit reference was found in the text -- Obsolete informational reference (is this intentional?): RFC 226 (Obsoleted by RFC 247) -- Obsolete informational reference (is this intentional?): RFC 760 (Obsoleted by RFC 791) -- Obsolete informational reference (is this intentional?): RFC 793 (Obsoleted by RFC 9293) -- Obsolete informational reference (is this intentional?): RFC 2460 (Obsoleted by RFC 8200) -- Obsolete informational reference (is this intentional?): RFC 3536 (Obsoleted by RFC 6365) -- Obsolete informational reference (is this intentional?): RFC 3979 (Obsoleted by RFC 8179) -- Obsolete informational reference (is this intentional?): RFC 4646 (Obsoleted by RFC 5646) -- Obsolete informational reference (is this intentional?): RFC 4941 (Obsoleted by RFC 8981) -- Obsolete informational reference (is this intentional?): RFC 5246 (Obsoleted by RFC 8446) -- Obsolete informational reference (is this intentional?): RFC 7540 (Obsoleted by RFC 9113) -- Obsolete informational reference (is this intentional?): RFC 7626 (Obsoleted by RFC 9076) Summary: 1 error (**), 0 flaws (~~), 5 warnings (==), 15 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Human Rights Protocol Considerations Research Group N. ten Oever 3 Internet-Draft ARTICLE 19 4 Intended status: Informational C. Cath 5 Expires: August 29, 2017 Oxford Internet Institute 6 February 25, 2017 8 Research into Human Rights Protocol Considerations 9 draft-irtf-hrpc-research-11 11 Abstract 13 This document aims to propose guidelines for human rights 14 considerations, similar to the work done on the guidelines for 15 privacy considerations [RFC6973]. If you want to apply this work to 16 your own, you can directly go to Section 6. The rest of the document 17 explains the background of the guidelines and how they were 18 developed. 20 This document is not an Internet Standards Track specification; it is 21 published for informational purposes. 23 This document is a product of the Internet Research Task Force 24 (IRTF). The IRTF publishes the results of Internet-related research 25 and development activities. This documents is a consensus document 26 of the Human Rights Protocol Consideration Research Group of the 27 Internet Research Task Force (IRTF). 29 Discussion of this draft at: hrpc@irtf.org // 30 https://www.irtf.org/mailman/listinfo/hrpc 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 August 29, 2017. 49 Copyright Notice 51 Copyright (c) 2017 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 Table of Contents 66 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 67 2. Vocabulary used . . . . . . . . . . . . . . . . . . . . . . . 5 68 3. Research Questions . . . . . . . . . . . . . . . . . . . . . 11 69 4. Literature and Discussion Review . . . . . . . . . . . . . . 11 70 5. Methodology . . . . . . . . . . . . . . . . . . . . . . . . . 14 71 5.1. Data Sources . . . . . . . . . . . . . . . . . . . . . . 15 72 5.1.1. Discourse analysis of RFCs . . . . . . . . . . . . . 16 73 5.1.2. Interviews with members of the IETF community . . . . 16 74 5.1.3. Participant observation in Working Groups . . . . . . 16 75 5.2. Data analysis strategies . . . . . . . . . . . . . . . . 16 76 5.2.1. Identifying qualities of technical concepts that 77 relate to human rights . . . . . . . . . . . . . . . 16 78 5.2.2. Relating human rights to technical concepts . . . . . 18 79 5.2.3. Map cases of protocols, implementations and 80 networking paradigms that adversely impact human 81 rights or are enablers thereof . . . . . . . . . . . 21 82 6. Model for developing human rights protocol considerations . . 39 83 6.1. Human rights threats . . . . . . . . . . . . . . . . . . 40 84 6.2. Guidelines for human rights considerations . . . . . . . 41 85 6.2.1. Connectivity . . . . . . . . . . . . . . . . . . . . 42 86 6.2.2. Privacy . . . . . . . . . . . . . . . . . . . . . . . 42 87 6.2.3. Content agnosticism . . . . . . . . . . . . . . . . . 43 88 6.2.4. Security . . . . . . . . . . . . . . . . . . . . . . 43 89 6.2.5. Internationalization . . . . . . . . . . . . . . . . 44 90 6.2.6. Censorship resistance . . . . . . . . . . . . . . . . 45 91 6.2.7. Open Standards . . . . . . . . . . . . . . . . . . . 46 92 6.2.8. Heterogeneity Support . . . . . . . . . . . . . . . . 48 93 6.2.9. Anonymity . . . . . . . . . . . . . . . . . . . . . . 48 94 6.2.10. Pseudonymity . . . . . . . . . . . . . . . . . . . . 49 95 6.2.11. Accessibility . . . . . . . . . . . . . . . . . . . . 50 96 6.2.12. Localization . . . . . . . . . . . . . . . . . . . . 51 97 6.2.13. Decentralization . . . . . . . . . . . . . . . . . . 51 98 6.2.14. Reliability . . . . . . . . . . . . . . . . . . . . . 52 99 6.2.15. Confidentiality . . . . . . . . . . . . . . . . . . . 53 100 6.2.16. Integrity . . . . . . . . . . . . . . . . . . . . . . 54 101 6.2.17. Authenticity . . . . . . . . . . . . . . . . . . . . 55 102 6.2.18. Adaptability . . . . . . . . . . . . . . . . . . . . 56 103 6.2.19. Outcome Transparency . . . . . . . . . . . . . . . . 57 104 7. Document Status . . . . . . . . . . . . . . . . . . . . . . . 57 105 8. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 57 106 9. Security Considerations . . . . . . . . . . . . . . . . . . . 58 107 10. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 58 108 11. Research Group Information . . . . . . . . . . . . . . . . . 58 109 12. References . . . . . . . . . . . . . . . . . . . . . . . . . 58 110 12.1. Informative References . . . . . . . . . . . . . . . . . 58 111 12.2. URIs . . . . . . . . . . . . . . . . . . . . . . . . . . 74 112 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 74 114 1. Introduction 116 "There's a freedom about the Internet: As long as we accept the 117 rules of sending packets around, we can send packets containing 118 anything to anywhere." 120 [Berners-Lee] 122 "The Internet isn't value-neutral, and neither is the IETF." 124 [RFC3935] 126 The evergrowing interconnectedness of Internet and society increases 127 the impact of the Internet on the lives of individuals. Because of 128 this, the design and development of the Internet infrastructure also 129 has a growing impact on society. This has led to a broad recognition 130 that human rights [UDHR] [ICCPR] [ICESCR] have a role in the 131 development and management of the Internet [HRC2012] [UNGA2013] 132 [NETmundial]. It has also been argued that the Internet should be 133 strengthened as a human rights enabling environment [Brown]. 135 This document aims to expose the relation between protocols and human 136 rights, propose possible guidelines to protect the Internet as a 137 human-rights-enabling environment in future protocol development, in 138 a manner similar to the work done for Privacy Considerations in 139 [RFC6973], and to increase the awareness in both the human rights 140 community and the technical community on the importance of the 141 technical workings of the Internet and its impact on human rights. 143 Open, secure and reliable connectivity is necessary (although not 144 sufficient) to exercise human rights such as freedom of expression 145 and freedom of association [FOC], as defined in the Universal 146 Declaration of Human Rights [UDHR]. The purpose of the Internet to 147 be a global network of networks that provides unfettered connectivity 148 to all users and for any content [RFC1958]. This objective of 149 stimulating global connectivity contributes to the Internet's role as 150 an enabler of human rights. The Internet has given people a platform 151 to exchange opinions, gather information, and it has enabled people 152 of different backgrounds and genders to participate in the public 153 debate, it has also allowed people to congregate and organize. Next 154 to that, the strong commitment to security [RFC1984] [RFC3365] and 155 privacy [RFC6973] [RFC7258] in the Internet's architectural design 156 contribute to the strengthening of the Internet as a human rights 157 enabling environment. One could even argue that the Internet is not 158 only an enabler of human rights, but that human rights lie at the 159 basis of, and are ingrained in, the architecture of the networks that 160 make up the Internet. Internet connectivity increases the capacity 161 for individuals to exercise their rights, the core of the Internet, 162 its architectural design is therefore closely intertwined with the 163 human rights framework [CathFloridi]. The quintessential link 164 between the Internet's infrastructure and human rights has been 165 argued by many. [Bless] for instance argues that, 'to a certain 166 extent, the Internet and its protocols have already facilitated the 167 realization of human rights, e.g., the freedom of assembly and 168 expression. In contrast, measures of censorship and pervasive 169 surveillance violate fundamental human rights.' [Denardis15] argues 170 that 'Since the first hints of Internet commercialization and 171 internationalization, the IETF has supported strong security in 172 protocol design and has sometimes served as a force resisting 173 protocol-enabled surveillance features.' By doing so, the IETF 174 enabled the manifestation of the right to privacy, through the 175 Internet's infrastructure. Additionally, access to freely available 176 information gives people access to knowledge that enables them to 177 help satisfy other human rights, as such the Internet increasingly 178 becomes a pre-condition for human rights rather than a supplement. 180 Human rights can be in conflict with each other, such as the right to 181 freedom of expression and the right to privacy. In such cases the 182 different affected rights need to be balanced. In order to do this 183 it is crucial that the rights impacts are clearly documented in order 184 to mitigate the potential harm. Making that process tangible and 185 practical for protocol developers is what this research aims to 186 ultimately contribute to. Technology can never be fully equated with 187 a human right. Whereas a specific technology might be strong enabler 188 of a specific human right, it might have an adverse impact on another 189 human right. In this case decisions on design and deployment need to 190 take this into account. 192 The open nature of the initial technical design and its open 193 standards, as well as developments like open source, fostered freedom 194 of communication. What emerged was a network of networks that could 195 enable everyone to connect and to exchange data, information and 196 code. For many, enabling such conections became a core value. 197 However as the scale and the commercialization of the Internet grew, 198 topics like access, rights and connectivity are forced to compete 199 with other values. Therefore, important human rights enabling 200 characteristics of the Internet might be degraded if they're not 201 properly defined, described and protected as such. And, the other 202 way around, not protecting human right enabling characteristics could 203 also result in (partial) loss of functionality and connectivity, and 204 other inherent parts of the Internet's architecture of networks. New 205 protocols, particularly those that upgrade the core infrastructure of 206 the network, should be designed to continue to enable fundamental 207 human rights. 209 The IETF has produced guidelines and procedures to ensure and 210 galvanize the privacy of indiduals and security of the network in 211 protocol development. This document aims to explore the possibility 212 of the development of similar procedures for guidelines for human 213 rights considerations to ensure that protocols developed in the IETF 214 do not have an adverse impact on the realization of human rights on 215 the Internet. By carefully considering the answers to the questions 216 posed in the Section 6 part of this document, document authors should 217 be able to produce a comprehensive analysis that can serve as the 218 basis for discussion on whether the protocol adequately protects 219 against human rights threats, and potentially stimulate authors to 220 think about alternative design choices. 222 2. Vocabulary used 224 In the discussion of human rights and Internet architecture concepts 225 developed in computer science, networking, law, policy-making and 226 advocacy are coming together [Dutton],[Kaye],[Franklin], [RFC1958]. 227 The same concepts might have a very different meaning and 228 implications in other areas of expertise. In order to foster a 229 constructive interdisciplinary debate, and minimize differences in 230 interpretation, the following glossary is provided, building as much 231 as possible on existing definitions, and where these were not 232 available definitions have been developed. 234 Accessibility Full Internet Connectivity as described in [RFC4084] 235 to provide unfettered access to the Internet 237 The design of protocols, services or implementation that provide 238 an enabling environment for people with disabilities. 240 The ability to receive information available on the Internet 242 Anonymity The condition of an identity being unknown or concealed. 243 [RFC4949] 245 Anonymous A state of an individual in which an observer or attacker 246 cannot identify the individual within a set of other individuals 247 (the anonymity set). [RFC6973] 249 Authenticity The property of being genuine and able to be verified 250 and be trusted. [RFC4949] 252 Blocking the practice of preventing access to resources in the 253 aggregate [RFC7754]. Both blocking and filtering can be 254 implemented at the level of "services" (web hosting or video 255 streaming, for example) or at the level of particular "content." 256 [RFC7754] 258 Censorship technical mechanisms, that include both blocking and 259 filtering, that certain political or private actors around the 260 world use to block or degrade Internet traffic. For further 261 details on the various elements of Internet censorship see [hall] 263 Censorship resistance Methods and measures to mitigate Internet 264 censorship. 266 Confidentiality The property that data is not disclosed to system 267 entities unless they have been authorized to know the data. 268 [RFC4949]. 270 Connectivity The extent to which a device or network is able to 271 reach other devices or networks to exchange data. The Internet is 272 the tool for providing global connectivity [RFC1958]. Different 273 types of connectivity are further specified in [RFC4084]. 275 The combination of the end-to-end principle, interoperability, 276 distributed architecture, resilience, reliability and robustness 277 are the enabling factors that result in connectivity to and on the 278 Internet. 280 Content agnosticism Treating network traffic identically regardless 281 of content. 283 Decentralized Implementation or deployment of standards, protocols 284 or systems without one single point of control. 286 End-to-End The principle that application-specific functions should 287 not be embedded into the network and thus stay at the end-points: 289 in many cases, especially when dealing with failures, the right 290 decisions can only be made with the corresponding application- 291 specific knowledge, which is available at the end-points not in 292 the network. 294 The end-to-end principle is one of the key architectural 295 guidelines of the Internet. The argument in favor of the end-to- 296 end approach to system design is laid out in the fundamental paper 297 by Saltzer, Reed, and Clark [Saltzer] [Clark]. In it, the authors 298 argue in favor of radical simplification: systems designers should 299 only build the essential and shared functions into the network, as 300 most functions can only be implemented at network end points. 301 Building features into the network for the benefit of certain 302 applications, will come at the expense of others. As such, as a 303 general system designers should attempt to steer clear of building 304 anything into the network that is not a bare necessity for its 305 functioning. Following the end-to-end principle is crucial for 306 innovation, as it makes innovation at the edges possible without 307 having to make changes to the network, and the robustness of the 308 network. Various aspects of end-to-end connectivity are further 309 elaborated on in [RFC2775]. 311 Federation The possibility of connecting autonomous and possibly 312 centralized systems into single system without a central 313 authority. 315 Filtering the practice of preventing access to specific resources 316 within an aggregate [RFC7754]. 318 Heterogeneity The Internet is characterized by heterogeneity on many 319 levels: devices and nodes, router scheduling algorithms and queue 320 management mechanisms, routing protocols, levels of multiplexing, 321 protocol versions and implementations, underlying link layers 322 (e.g., point-to-point, multi-access links, wireless, FDDI, etc.), 323 in the traffic mix and in the levels of congestion at different 324 times and places. Moreover, as the Internet is composed of 325 independent organizations and Internet service providers, each 326 with their own separate policy concerns,there is a large 327 heterogeneity of administrative domains and pricing structures. 328 As a result, the heterogeneity principle proposed in [RFC1958] 329 needs to be supported by design. [FIArch] 331 Human rights Human rights are principles and norms that are 332 indivisible, interrelated, unalienable, universal, and mutually 333 reinforcing that have been codified in national and international 334 bodies of law. The Universal Declaration of Human Rights [UDHR] 335 is the most well-known document in the history of human rights. 336 The apirations from this documents were later codified into 337 treaties such as the [ICCPR] and the [ICESCR], after which 338 signatory countries were obliged to reflect them in their national 339 bodies of law. There is also a broad recognition that not only 340 states have an obligations vis a vis human rights, but non-state 341 actors do so as well. 343 Integrity The property that data has not been changed, destroyed, or 344 lost in an unauthorized or accidental manner. [RFC4949]. 346 Interoperable A property of a documented standard or protocol which 347 allows different independent implementations to work with each 348 other without any restriction on functionality. 350 Internationalization (i18n) The practice of making protocols, 351 standards, and implementations usable in different languages and 352 scripts (see Localization). 354 "In the IETF, "internationalization" means to add or improve the 355 handling of non-ASCII text in a protocol" [RFC6365]. A different 356 perspective, more appropriate to protocols that are designed for 357 global use from the beginning, is the definition used by W3C: 359 "Internationalization is the design and development of a product, 360 application or document content that enables easy localization for 361 target audiences that vary in culture, region, or language." 362 [W3Ci18nDef] 364 Many protocols that handle text only handle one charset (US- 365 ASCII), or leave the question of encoding up to local guesswork 366 (which leads, of course, to interoperability problems) [RFC3536]. 367 If multiple charsets are permitted, they must be explicitly 368 identified [RFC2277]. Adding non-ASCII text to a protocol allows 369 the protocol to handle more scripts, hopefully all of the ones 370 useful in the world. In today's world, that is normally best 371 accomplished by allowing Unicode encoded in UTF-8 only, thereby 372 shifting conversion issues away from ad hoc choices. 374 Localization (l10n) The practice of translating an implementation to 375 make it functional in a specific language or for users in a 376 specific locale (see Internationalization). 378 (cf [RFC6365]): The process of adapting an internationalized 379 application platform or application to a specific cultural 380 environment. In localization, the same semantics are preserved 381 while the syntax may be changed. [FRAMEWORK] 383 Localization is the act of tailoring an application for a 384 different language or script or culture. Some internationalized 385 applications can handle a wide variety of languages. Typical 386 users only understand a small number of languages, so the program 387 must be tailored to interact with users in just the languages they 388 know. The major work of localization is translating the user 389 interface and documentation. Localization involves not only 390 changing the language interaction, but also other relevant changes 391 such as display of numbers, dates, currency, and so on. The 392 better internationalized an application is, the easier it is to 393 localize it for a particular language and character encoding 394 scheme. 396 Open standards Conform with [RFC2026]: Various national and 397 international standards bodies, such as ANSI, ISO, IEEE, and ITU- 398 T, develop a variety of protocol and service specifications that 399 are similar to Technical Specifications defined here. National 400 and international groups also publish "implementors' agreements" 401 that are analogous to Applicability Statements, capturing a body 402 of implementation-specific detail concerned with the practical 403 application of their standards. All of these are considered to be 404 "open external standards" for the purposes of the Internet 405 Standards Process. 407 Openness Absence of centralized points of control - a feature that 408 is assumed to make it easy for new users to join and new uses to 409 unfold [Brown]. 411 Permissionless innovation The freedom and ability to freely create 412 and deploy new protocols on top of the communications constructs 413 that currently exist. 415 Privacy The right of an entity (normally a person), acting in its 416 own behalf, to determine the degree to which it will interact with 417 its environment, including the degree to which the entity is 418 willing to share its personal information with others. [RFC4949] 420 The right of individuals to control or influence what information 421 related to them may be collected and stored and by whom and to 422 whom that information may be disclosed. 424 Privacy is a broad concept relating to the protection of 425 individual or group autonomy and the relationship between an 426 individual or group and society, including government, companies 427 and private individuals. It is often summarized as "the right to 428 be left alone" but it encompasses a wide range of rights including 429 protections from intrusions into family and home life, control of 430 sexual and reproductive rights, and communications secrecy. It is 431 commonly recognized as a core right that underpins human dignity 432 and other values such as freedom of association and freedom of 433 speech. 435 The right to privacy is also recognized in nearly every national 436 constitution and in most international human rights treaties. It 437 has been adjudicated upon both by international and regional 438 bodies. The right to privacy is also legally protected at the 439 national level through provisions in civil and/or criminal codes. 441 Reliability Reliability ensures that a protocol will execute its 442 function consistently as described and function without unexpected 443 result. A system that is reliable degenerates gracefully and will 444 have a documented way to announce degradation. It also has 445 mechanisms to recover from failure gracefully, and if applicable, 446 allow for partial healing [dict]. 448 Resilience The maintaining of dependability and performance in the 449 face of unanticipated changes and circumstances [Meyer]. 451 Robustness The resistance of protocols and their implementations to 452 errors, and to involuntary, legal or malicious attempts to disrupt 453 its mode of operations. [RFC0760] [RFC0791] [RFC0793] [RFC1122]. 454 Or framed more positively, a system can provide functionality 455 consistently and without errors despite involuntary, legal or 456 malicious attempts to disrupt its mode of operations. 458 Scalability The ability to handle increased or decreased system 459 parameters (e.g., number of end-systems, users, data flows, 460 routing entries. etc.) predictably within defined expectations. 461 There should be a clear definition of its scope and applicability. 462 The limits of a systems scalability should be defined. Growth or 463 shrinkage of these parameters is typically considered by orders of 464 magnitude. 466 Strong encryption / cryptography Used to describe a cryptographic 467 algorithm that would require a large amount of computational power to 468 defeat it. [RFC4949]. In the modern usage of the definition 'strong 469 encryption' this refers to an amount of computing power current not 470 available, not even to major state-level actors. 472 Transparency In this context transparency is linked to the 473 comprehensibility of a protocol in relation to the choices it 474 makes for both user and protocol developers and implementers and 475 to its outcome. 477 outcome transparency, is linked to the comprehensibility of the 478 effects of a protocol in relation to the choices it makes for both 479 user and protocol developers and implementers, including the 480 comprehensibility of possible unintended consequences of protocol 481 choices (e.g. lack of authenticity may lead to lack of integrity 482 and negative externalities) 484 3. Research Questions 486 The Human Rights Protocol Considerations Research Group (hrpc) in the 487 Internet Research Taskforce (IRTF) embarked on its mission to answer 488 the following two questions which are also the main two questions 489 which this documents seeks to answer: 491 1. How can Internet protocols and standards impact human rights, 492 either by enabling them or by creating a restrictive environment? 494 2. Can guidelines be developed to improve informed and transparent 495 decision making about potential human rights impact of protocols? 497 4. Literature and Discussion Review 499 Protocols and standards are regularly seen as merely performing 500 technical functions. However, these protocols and standards do not 501 exist outside of their technical context nor outside of their 502 political, historical, economic, legal or cultural context. This is 503 best exemplified by the way in which some Internet processes and 504 protocols have become part and parcel of political processes and 505 public policies: one only has to look at the IANA transition, the RFC 506 on pervasive monitoring or global innovation policy for concrete 507 examples [Denardis15]. According to [Abbate]: "protocols are 508 politics by other means". This statement would probably not garner 509 IETF consensus, but it nonetheless confers that protocols are based 510 on decision making, most often by humans. In this process the values 511 and ideas about the role that a particular technology should perform 512 in society is embedded into the design. Often these design decisions 513 are part pure-technical, and part inspired by certain world view of 514 how technology should function that is inspired by personal, 515 corporate and political views. Within the community of IETF 516 participants there is a strong desire to solve technical problems and 517 minimize engagement with political processes and non-protocol related 518 political issues. 520 Since the late 1990's a burgeoning group of academics and 521 practitioners researched questions surrounding the societal impact of 522 protocols, and the politics of protocols. These studies vary in 523 focus and scope: some focus on specific standards [Davidsonetal] 524 [Musiani], others look into the political, legal, commercial or 525 social impact of protocols [BrownMarsden] [Lessig], [Mueller] and yet 526 others look at how the engineers' personal set of values get 527 translated into technology [Abbate] [CathFloridi] [Denardis15] 528 [WynsbergheMoura]. 530 Commercial and political influences on the management of the 531 Internet's infrastructure are well-documented in the academic 532 literature and will thus not be discussed here [Benkler] [Brownetal] 533 [Denardis15] [Lessig] [Mueller] [Zittrain]. It is sufficient to 534 say that the IETF community consistently tries to push back against 535 the standardization of surveillance and certain other issues that 536 negatively influence end-users' experience of and trust in the 537 Internet [Denardis14]. The role human rights play in engineering, 538 infrastructure maintenance and protocol design is much less clear. 540 It is very important to understand how protocols and standards impact 541 human rights. In particular because Standard Developing 542 Organizations (SDOs) are increasingly becoming venues where social 543 values (like human rights) are discussed, although often from a 544 technological point of view. These SDOs are becoming a new focal 545 point for discussions about values-by-design, and the role of 546 technical engineers in protecting or enabling human rights 547 [Brownetal] [Clarketal] [Denardis14] [CathFloridi] [Lessig] 548 [Rachovitsa]. 550 In the academic literature five clear positions can be discerned, in 551 relation to the role of human rights in protocol design and how to 552 account for these human rights in protocol development: Clark et al. 553 argue that there is a need to 'design for variation in outcome, so 554 that the outcome can be different in different places, and the tussle 555 takes place within the design (...) [as] Rigid designs will be 556 broken; designs that permit variation will flex under pressure and 557 survive [Clarketal].' They hold that human rights should not be 558 hard-coded into protocols because of three reasons: first, the rights 559 in the UDHR are not absolute. Second, technology is not the only 560 tool in the tussle over human rights. And last but not least, it is 561 dangerous to make promises that can't be kept. The open nature of 562 the Internet will never, they argue, be enough to fully protect 563 individuals' human rights. 565 Conversely, Brown et al. [Brownetal] state that 'some key, universal 566 values - of which the UDHR is the most legitimate expression - should 567 be baked into the architecture at design time.' They argue that 568 design choices have offline consequences, and are able to shape the 569 power positions of groups or individuals in society. As such, the 570 individuals making these technical decisions have a moral obligation 571 to take into account the impact of their decisions on society, and by 572 extension human rights. Brown et al recognise that values and the 573 implementation of human rights vary across the globe. Yet they argue 574 that all members of the United Nations have found 'common agreement 575 on the values proclaimed in the Universal Declaration of Human 576 Rights. In looking for the most legitimate set of global values to 577 embed in the future Internet architectures, the UDHR has the 578 democratic assent of a significant fraction of the planet's 579 population, through their elected representatives." 581 The main disagreement between these two academic positions lies 582 mostly in the question on whether a particular value system should be 583 embedded into the Internet's architectures or whether the 584 architectures need to account for a varying set of values. 586 A third position that is similar to that of Brown et al., is taken by 587 [Broeders] who argues that 'we must find ways to continue 588 guaranteeing the overall integrity and functionality of the public 589 core of the Internet.' He argues that the best way to do this is by 590 declaring the backbone of the Internet - which includes the TCP/IP 591 protocol suite, numerous standards, the Domain Name System (DNS), and 592 routing protocols - a common public good. This is a different 593 approach than that of [Clarketal] and [Brownetal] because Broeders 594 does not suggest that social values should (or should not) be 595 explicitly coded into the Internet, but rather that the existing 596 infrastructure should be seen as an entity of public value. 598 Bless and Orwat [Bless] represent a fourth position. They argue that 599 it is too early to make any definitive claims, but that there is a 600 need for more careful analysis of the impact of protocol design 601 choices on human rights. They also argue that it is important to 602 search for solutions that 'create awareness in the technical 603 community about impact of design choices on social values. And work 604 towards a methodology for co-design of technical and institutional 605 systems.' 607 Berners-Lee and Halpin argue that the Internet could lead to even new 608 capacities, and these capacities may over time be viewed as new kinds 609 of rights. For example, Internet access may be viewed as a human 610 right in of itself if it is taken to be a pre-condition for other 611 rights, even if it could not have been predicted at the declaration 612 of the UNHDR after the end of World War 2.[BernersLeeHalpin]. 614 It is important to contextualize the technical discussion with the 615 academic discussions on this issue. The academic discussions also 616 are important to document as they inform the position of the authors 617 of this document. The Research Groups position is that hard-coding 618 human rights into protocols is complicated and changes with the 619 context. At this point is difficult to say whether hard-coding human 620 rights into protocols is wise or feasible. Additionally, there are 621 many human rights, but that not all are relevant for ICTs. A partial 622 catalog, with references to sources, of human rights related to ICTs 623 can be found here [Hill2014]. It is however important to make 624 conscious and explicit design decisions that take into account the 625 human rights protocol considerations guidelines developed below. 626 This will contribute to the understanding of the impact protocols can 627 have on human rights, both for developers and for users. In 628 addition, it contributes to the careful consideration of the impact 629 that a specific protocol might have on human rights and that concrete 630 design decisions are documented in the protocol. 632 Pursuant to the principle of constant change, since the function and 633 scope of the Internet evolves, so does the role of the IETF in 634 developing standards. Internet standards are adopted on the basis of 635 a series of criteria, including high technical quality, support by 636 community consensus, and their overall benefit to the Internet. The 637 latter calls for an assessment of the interests of all affected 638 parties and the specifications' impact on the Internet's users. In 639 this respect, the effective exercise of the human rights of the 640 Internet users is a relevant consideration that needs to be 641 appreciated in the standardization process insofar as it is directly 642 linked to the reliability and core values of the Internet. [RFC1958] 643 [RFC0226] [RFC3724] 645 This document details the steps taken in the research into human 646 rights protocol considerations by the hrpc research group to clarify 647 the relation between technical concepts used in the IETF and human 648 rights. This document sets out some preliminary steps and 649 considerations for engineers to take into account when developing 650 standards and protocols. 652 5. Methodology 654 Mapping the relation between human rights, protocols and 655 architectures is a new research challenge, which requires a good 656 amount of interdisciplinary and cross organizational cooperation to 657 develop a consistent methodology. 659 The methodological choices made in this document are based on the 660 political science-based method of discourse analysis and ethnographic 661 research methods [Cath]. This work departs from the assumption that 662 language reflects the understanding of concepts. Or as [Jabri] 663 holds, policy documents are 'social relations represented in texts 664 where language is used to construct meaning and representation'. 665 This process happens in 'the social space of society' [Schroeder] and 666 manifests itself in institutions and organizations [King], exposed 667 using the ethnographic methods of semi-structured interviews and 668 participant observation. Or in non-academic language, the way the 669 language in IETF/IRTF documents describes and approaches the issues 670 they are trying to address is an indicator for the underlying social 671 assumptions and relations of the engineers to their engineering. By 672 reading and analyzing these documents, as well as interviewing 673 engineers and participating in the IETF/IRTF working groups, it is 674 possible to distill the relation between human rights, protocols and 675 the Internet's infrastructure as it pertains to the work of the IETF. 677 The discourse analysis was operationalized using qualitative and 678 quantitative means. The first step taken by the authors and 679 contributors was reading RFCs and other official IETF documents. The 680 second step was the use of a python-based analyzer, using the tool 681 Big Bang, adapted by Nick Doty [Doty] to scan for the concepts that 682 were identified as important architectural principles (distilled on 683 the initial reading and supplemented by the interviews and 684 participant observation). Such a quantitative method is very precise 685 and speeds up the research process [Richie]. But this tool is unable 686 to understand 'latent meaning' [Denzin]. In order to mitigate these 687 issues of automated word-frequency based approaches, and to get a 688 sense of the 'thick meaning' [Geertz] of the data, a second 689 qualitative analysis of the data set was performed. These various 690 rounds of discourse analysis were used to inform the interviews and 691 further data analysis. As such the initial rounds of quantitative 692 discourse analysis were used to inform the second rounds of 693 qualitative analysis. The results from the qualitative interviews 694 were again used to feed new concepts into the quantitative discourse 695 analysis. As such the two methods continued to support and enrich 696 each other. 698 The ethnographic methods of the data collection and processing 699 allowed the research group to acquire the data necessary to 'provide 700 a holistic understanding of research participants' views and actions' 701 [Denzin] that highlighted ongoing issues and case studies where 702 protocols impact human rights. The interview participants were 703 selected through purposive sampling [Babbie], as the research group 704 was interested in getting a wide variety of opinions on the role of 705 human rights in guiding protocol development. This sampling method 706 also ensured that individuals with extensive experience working at 707 the IETF in various roles were targeted. The interviewees included 708 individuals in leadership positions (Working Group (WG) chairs, Area 709 Directors (ADs)), 'regular participants', individuals working for 710 specific entities (corporate, civil society, political, academic) and 711 represented various backgrounds, nationalities and genders. 713 5.1. Data Sources 715 In order to map the potential relation between human rights and 716 protocols, the HRPC research group gathered data from three specific 717 sources: 719 5.1.1. Discourse analysis of RFCs 721 To start addressing the issue, a mapping exercise analyzing Internet 722 infrastructure and protocols features, vis-a-vis their possible 723 impact on human rights was undertaken. Therefore, research on the 724 language used in current and historic RFCs and mailing list 725 discussions was undertaken to expose core architectural principles, 726 language and deliberations on human rights of those affected by the 727 network. 729 5.1.2. Interviews with members of the IETF community 731 Over 30 interviews with the current and past members of the Internet 732 Architecture Board (IAB), current and past members of the Internet 733 Engineering Steering Group (IESG) and chairs of selected working 734 groups and RFC authors were done at the IETF92 Dallas meeting in 735 March 2015. To get an insider understanding of how they view the 736 relationship (if any) between human rights and protocols to play out 737 in their work. Several of the participants opted to remain 738 anonymous, if you are interested in this data set please contact the 739 authors. 741 5.1.3. Participant observation in Working Groups 743 By participating in various working groups, in person at IETF 744 meetings and on mailinglists, information was gathered about the 745 IETFs day-to-day workings. From which general themes, technical 746 concepts, and use-cases about human rights and protocols were 747 extracted. This process started at the IETF91 meeting and continues 748 today. 750 5.2. Data analysis strategies 752 The data above was processed using three consecutive strategies: 753 mapping protocols related to human rights, extracting concepts from 754 these protocols, and creation of a common glossary (detailed under 755 Section 2). Before going over these strategies some elaboration on 756 the process of identifying technical concepts as they relate to human 757 rights needs to be given: 759 5.2.1. Identifying qualities of technical concepts that relate to human 760 rights 762 5.2.1.1. Mapping protocols and standards to human rights 764 By combining data from the three data sources named above, an 765 extensive list of protocols and standards that potentially enable the 766 Internet as a tool for freedom of expression and association was 767 created. In order to determine the enabling (or inhibiting) features 768 we relied on direct references of such impact in the RFCs, as well as 769 input from the community. On the basis of this analysis a list of 770 RFCs that describe standards and protocols that are potentially 771 closely related to human rights was compiled. 773 5.2.1.2. Extracting concepts from selected RFCs 775 Identifying the protocols and standards that are related to human 776 rights and create a human rights enabeling environment was the first 777 step. For that we needed to focus on specific technical concepts 778 that underlie these protocols and standards. On the basis of this 779 list a number of technical concepts that appeared frequently was 780 extracted, and used to create a second list of technical terms that, 781 when combined and applied in different circumstances, create an 782 enabling environment for excercising human rights on the Internet. 784 5.2.1.3. Building a common vocabulary of technical concepts that impact 785 human rights 787 While interviewing experts, investigating RFCs and compiling 788 technical definitions several concepts of convergence and divergence 789 were identified. To ensure that the discussion was based on a common 790 understanding of terms and vocabulary, a list of definitions was 791 created. The definitions are based on the wording found in various 792 IETF documents, and if these were unavailable definitions were taken 793 from definitions from other Standards Developing Organizations or 794 academic literature, as indicated in the vocabulary section. 796 5.2.1.4. Translating Human Rights Concepts into Technical Definitions 798 The previous steps allowed for the clarification of relations between 799 human rights and technical concepts. The steps taken show how the 800 research process zoomed in, from compiling a broad lists of protocols 801 and standards that relate to human rights to extracting the precise 802 technical concepts that make up these protocols and standards, in 803 order to understand the relationship between the two. This sub- 804 section presents the next step: translating human rights to technical 805 concepts by matching the individuals components of the rights to the 806 accompanying technical concepts, allowing for the creation of a list 807 of technical concepts that when partially combined can create an 808 enabling environment for human rights. 810 5.2.1.5. List technical terms that when partially combined can create 811 an enabling environment for human rights 813 On the basis of the prior steps the following list of technical 814 terms, that when partially combined can create an enabling 815 environment for human rights, such a freedom of expression and 816 freedom of association, was drafted. 818 Architectural principles Enabling features 819 and system properties for user rights 821 /------------------------------------------------\ 822 | | 823 +=================|=============================+ | 824 = | = | 825 = | End to end = | 826 = | Reliability = | 827 = | Resilience = Access as | 828 = | Interoperability = Human Right | 829 = Good enough | Transparency = | 830 = principle | Data minimization = | 831 = | Permissionless innovation = | 832 = Simplicity | Graceful degradation = | 833 = | Connectivity = | 834 = | Heterogeneity support = | 835 = | = | 836 = | = | 837 = \------------------------------------------------/ 838 = = 839 +===============================================+ 841 figure 1 - relation between architectural principles and enabling 842 features for user rights. 844 5.2.2. Relating human rights to technical concepts 846 The combination of the technical concepts that have been gathered the 847 steps above have been grouped according to their impact on specific 848 rights as they have been mentioned in the interviews done at IETF92 849 as well as study of literature (see literature and discussion review 850 above). 852 This analysis aims to assist protocol developers in better 853 understanding the roles specific technical concepts have with regards 854 to their contribution to an enabeling environment for people to 855 excise their human rights. 857 This analysis does not claim to be a complete or exhaustive mapping 858 of all possible ways in which a protocols could potentially impact 859 human rights, but it presents an initial concept mapping based on 860 interviews and literature and discussion review. 862 +-----------------------+-----------------------------------------+ 863 | Technical Concepts | Rights potentially impacted | 864 +-----------------------+-----------------------------------------+ 865 | Connectivity | | 866 | Privacy | | 867 | Security | | 868 | Content agnosticism | Right to freedom of expression | 869 | Internationalization | | 870 | Censorship resistance | | 871 | Open Standards | | 872 | Heterogeneity support | | 873 +-----------------------+-----------------------------------------+ 874 | Anonymity | | 875 | Privacy | | 876 | Pseudonymity | Right to non-discrimination | 877 | Accessibility | | 878 +-----------------------+-----------------------------------------+ 879 | Content agnosticism | | 880 | Security | Right to equal protection | 881 +-----------------------+-----------------------------------------+ 882 | Accessibility | | 883 | Internationalization | Right to political participation | 884 | Censorship resistance | | 885 | Connectivity | | 886 +-----------------------+-----------------------------------------+ 887 | Open standards | | 888 | Localization | Right to participate in cultural life, | 889 | Internationalization | arts and science & | 890 | Censorship resistance | Right to education | 891 | Accessibility | | 892 +-----------------------+-----------------------------------------+ 893 | Connectivity | | 894 | Decentralization | | 895 | Censorship resistance | Right to freedom of assembly | 896 | Pseudonymity | and association | 897 | Anonymity | | 898 | Security | | 899 +-----------------------+-----------------------------------------+ 900 | Reliability | | 901 | Confidentiality | | 902 | Integrity | Right to security | 903 | Authenticity | | 904 | Anonymity | | 905 | | | 906 +-----------------------+-----------------------------------------+ 907 figure 2 - relation between specific technical concepts with regards 908 to their contribution to an enabeling environment for people to 909 exercise their human rights 911 5.2.3. Map cases of protocols, implementations and networking paradigms 912 that adversely impact human rights or are enablers thereof 914 Given the information above, the following list of cases of 915 protocols, implenentations and networking paradigms that adversely 916 impact or enable human rights was formed. 918 It is important to note that the assessment here is not a general 919 judgment on these protocols, nor an exhaustive listing of all the 920 potential negative or positive impacts on human rights they might 921 have. When they were conceived, there were many criteria to take 922 into account. For instance, relying on an centralized service can be 923 bad for freedom of speech (it creates one more control point, where 924 censorship could be applied) but it may be a necessity if the 925 endpoints are not connected and reachable permanently. So, when we 926 say "protocol X has feature Y, which may endanger the freedom of 927 speech", it does not mean that protocol X is bad and even less that 928 its authors were evil. The goal here is to show, with actual 929 examples, that the design of protocols have practical consequences 930 for some human rights and these consequences have to be considered in 931 the design phase. 933 5.2.3.1. IPv4 935 The Internet Protocol version 4 (IPv4), also known as 'layer 3' of 936 the Internet, and specified as a common encapsulation and protocol 937 header, is defined in [RFC0791]. The evolution of Internet 938 communications led to continued development in this area, 939 encapsulated in the development of version 6 (IPv6) of the protocol 940 in [RFC2460]. In spite of this updated protocol, we find that 25 941 years after the specification of version 6 of the protocol, the older 942 v4 standard continues to account for a sizeable majority of Internet 943 traffic, and most of the issues discussed here (with the big 944 exception of NAT, see Address Translation) are valid for IPv4 as well 945 as IPv6. 947 The Internet was designed as a platform for free and open 948 communication, most notably encoded in the end-to-end principle, and 949 that philosophy is also present in the technical implementation of 950 the Internet Protocol. [RFC3724] While the protocol was designed to 951 exist in an environment where intelligence is at the end hosts, it 952 has proven to provide sufficient information that a more intelligent 953 network core can make policy decisions and enforce policy-based 954 traffic shaping and restricting the communications of end hosts. 956 These capabilities for network control and limitations of the freedom 957 of expression by end hosts can be traced back to the IPv4 design, 958 helping us to understand which technical protocol decisions have led 959 to harm of this human rights. A feature that can harm freedom of 960 expression as well as the right to privacy through misuse of the 961 Internet Protocol is the exploitation of the public visibility of the 962 host pairs for all communications, and the corresponding ability to 963 discriminate and block traffic as a result of that metadata. 965 5.2.3.1.1. Network visibility of Source and Destination 967 The IPv4 protocol header contains fixed location fields for both the 968 source and destination IP addresses [RFC0791]. These addresses 969 identify both the host sending and receiving each message, and allow 970 the core network to understand who is talking to whom, and to 971 practically limit communication selectively between pairs of hosts. 972 Blocking of communication based on the pair of source and destination 973 is one of the most common limitations on the ability for people to 974 communicate today, [caida] and can be seen as a restriction of the 975 ability for people to assemble or to consensually express themselves. 977 Inclusion of an Internet-wide identified source in the IP header is 978 not the only possible design, especially since the protocol is most 979 commonly implemented over Ethernet networks exposing only link-local 980 identifiers [RFC0894]. 982 A variety of alternative designs do exist, such as the Accountable 983 and Private Internet Protocol [APIP] and Hornet [Hornet] as well as 984 source routing. The latter would allow for the sender to choose a 985 pre-defined (safe) route and spoofing of the source IP address, which 986 are technically supported by the IPv4 protocol, but neither are 987 considered good practice on the Internet [Farrow]. While projects 988 like [torproject] provide an alternative implementation of anonymity 989 in connections, they have been developed in spite of the IPv4 990 protocol design. 992 5.2.3.1.2. Address Translation and Mobility 994 A major structural shift in the Internet which undermined the 995 protocol design of IPv4, and significantly reduced the freedom of end 996 users to communicate and assemble is the introduction of network 997 address translation. [RFC3022] Network address translation is a 998 process whereby organizations and autonomous systems connect two 999 networks by translating the IPv4 source and destination addresses 1000 between the two. This process puts the router performing the 1001 translation into a privileged position, where it can decide which 1002 subset of communications are worthy of translation, and whether an 1003 unknown request for communication will be correctly forwarded to a 1004 host on the other network. 1006 This process of translation has widespread adoption despite promoting 1007 a process that goes against the stated end-to-end process of the 1008 underlying protocol [natusage]. In contrast, the proposed mechanism 1009 to provide support for mobility and forwarding to clients which may 1010 move, encoded instead as an option in the IP protocol in [RFC5944], 1011 has failed to gain traction. In this situation the compromise made 1012 in the design of the protocol resulted in a technology that is not 1013 coherent with the end-to-end principles and thus creates an extra 1014 possible hurdle for freedom of expression in its design, even though 1015 a viable alternative exists. There is a particular problem 1016 surrounding NATs and VPN (as well as other connections used for 1017 privacy purposes) as NATs sometimes cause VPNs not to work. 1019 5.2.3.2. DNS 1021 The Domain Name System (DNS) [RFC1035], provides service discovery 1022 capabilities, and provides a mechanism to associate human readable 1023 names with services. The DNS system is organized around a set of 1024 independently operated 'Root Servers' run by organizations which 1025 function in line with ICANN's policy by answering queries for which 1026 organizations have been delegated to manage registration under each 1027 Top Level Domain (TLD). The DNS is organized as a rooted tree, and 1028 this brings up political and social concerns over control. Top Level 1029 domains are maintained and determined by ICANN. These namespaces 1030 encompass several classes of services. The initial name spaces 1031 including '.Com' and '.Net', provide common spaces for expression of 1032 ideas, though their policies are enacted through US based companies. 1033 Other name spaces are delegated to specific nationalities, and may 1034 impose limits designed to focus speech in those forums both to 1035 promote speech from that nationality, and to comply with local limits 1036 on expression and social norms. Finally, the system has recently 1037 been expanded with additional generic and sponsored name spaces, for 1038 instance '.travel' and '.ninja', which are operated by a range of 1039 organizations which may independently determine their registration 1040 policies. This new development has both positive and negative 1041 implications in terms of enabling human rights. Some individuals 1042 argue that it undermines the right to freedom of expression because 1043 some of these new gtlds have restricted policies on registration and 1044 particular rules on hate speech content. Others argue that precisely 1045 these properties are positive because they enable certain (mostly 1046 minority) communities to build safer spaces for association, thereby 1047 enabling their right to freedom of association. An often mentioned 1048 example is an application like .gay [CoE]. 1050 DNS has significant privacy issues per [RFC7626]. Most notable the 1051 lack of encryption to limit the visibility of requests for domain 1052 resolution from intermediary parties, and a limited deployment of 1053 DNSSEC to provide authentication, allowing the client to know that 1054 they received a correct, "authoritative", answer to a query. In 1055 response to the privacy issues, the IETF DNS PRIVate Exchange 1056 (DPRIVE) Working Group is developing mechanisms to provide 1057 confidentiality to DNS transactions, to address concerns surrounding 1058 pervasive monitoring [RFC7258]. 1060 Authentication through DNSSEC creates a validation path for records. 1061 This authentication protects against forged or manipulated DNS data. 1062 As such DNSSEC protects the directory look-up and makes hijacking of 1063 a session harder. This is important because currently interference 1064 with the operation of the DNS is becoming one of the central 1065 mechanisms used to block access to websites. This interference 1066 limits both the freedom of expression of the publisher to offer their 1067 content, and the freedom of assembly for clients to congregate in a 1068 shared virtual space. Even though DNSSEC doesn't prevent censorship, 1069 it makes it clear that the returned information is not the 1070 information that was requested, which contributes to the right to 1071 security and increases trust in the network. It is however important 1072 to note that DNSSEC is currently not widely supported or deployed by 1073 domain name registrars, making it difficult to authenticate and use 1074 correctly. 1076 5.2.3.2.1. Removal of records 1078 There have been a number of cases where the records for a domain are 1079 removed from the name system due to political events. Examples of 1080 this removal includes the 'seizure' of wikileaks [bbc-wikileaks] and 1081 the names of illegally operating gambling operations by the United 1082 States Immigrations and Customs Enforcement unit (ICE). In the first 1083 case, a US court ordered the registrar to take down the domain. In 1084 the second, ICE compelled the US-based registry in charge of the .com 1085 TLD to hand ownership of those domains over to the US government. 1086 The same technique has been used in Libya to remove sites in 1087 violation of "our Country's Law and Morality (which) do not allow any 1088 kind of pornography or its promotion." [techyum] 1090 At a protocol level, there is no technical auditing for name 1091 ownership, as in alternate systems like [namecoin]. As a result, 1092 there is no ability for users to differentiate seizure from the 1093 legitimate transfer of name ownership, which is purely a policy 1094 decision of registrars. While DNSSEC addresses network distortion 1095 events described below, it does not tackle this problem. 1097 (While mentioning alternative techniques, this is not a comparison of 1098 DNS with Namecoin: the latter has its own problems and limitations. 1099 The idea here is to show that there are several possible choices, and 1100 they have consequences for human rights.) 1102 5.2.3.2.2. Distortion of records 1104 The most common mechanism by which the DNS system is abused to limit 1105 freedom of expression is through manipulation of protocol messages by 1106 the network. One form occurs at an organizational level, where 1107 client computers are instructed to use a local DNS resolver 1108 controlled by the organization. The DNS resolver will then 1109 selectively distort responses rather than request the authoritative 1110 lookup from the upstream system. The second form occurs through the 1111 use of deep packet inspection, where all DNS protocol messages are 1112 inspected by the network, and objectionable content is distorted, as 1113 can be observed in Chinese network. 1115 A notable instance of distortion occurred in Greece [ververis], where 1116 a study found evidence of both of deep packet inspection to distort 1117 DNS replies, and more excessive blocking of content than was legally 1118 required or requested (also known as overblocking). ISPs prevented 1119 clients from resolving the names of domains which they were 1120 instructed to do through a governmental order, prompting this 1121 particular blocking systems there. 1123 At a protocol level, the effectiveness of these attacks is made 1124 possible by a lack of authentication in the DNS protocol. DNSSEC 1125 provides the ability to determine authenticity of responses when 1126 used, but it is not regularly checked by resolvers. DNSSEC is not 1127 effective when the local resolver for a network is complicit in the 1128 distortion, for instance when the resolver assigned for use by an ISP 1129 is the source of injection. Selective distortion of records is also 1130 been made possible by the predictable structure of DNS messages, 1131 which make it computationally easy for a network device to watch all 1132 passing messages even at high speeds, and the lack of encryption, 1133 which allows the network to distort only an objectionable subset of 1134 protocol messages. Specific distortion mechanisms are discussed 1135 further in [hall]. 1137 Users can switch to another resolver, for instance a public one. The 1138 distorter can then try to block or hijack the connection to this 1139 resolver. This may start an arm's race, the user switching to 1140 secured connections to this alternative resolver ([RFC7858]), the 1141 disruptor then trying to find more sophisticated ways to block or 1142 hijack. In some cases, this search for an alternative, non- 1143 disrupting resolver, may lead to more centralisation, many people 1144 going to a few big commercial public resolvers. 1146 5.2.3.2.3. Injection of records 1148 Responding incorrectly to requests for name lookups is the most 1149 common mechanism that in-network devices use to limit the ability of 1150 end users to discover services. A deviation, which accomplishes a 1151 similar objective may be seen as different from a freedom of 1152 expression perspective, is the injection of incorrect responses to 1153 queries. The most prominent example of this behavior occurs in 1154 China, where requests for lookups of sites deemed inappropriate will 1155 trigger the network to respond with a false response, causing the 1156 client to ignore the real response when it subsequently arrives. 1157 [greatfirewall] Unlike the other forms of discussion mentioned above, 1158 injection does not stifle the ability of a server to announce it's 1159 name, it instead provides another voice which answers sooner. This 1160 is effective because without DNSSEC, the protocol will respond to 1161 whichever answer is received first, without listening for subsequent 1162 answers. 1164 5.2.3.3. HTTP 1166 The Hypertext Transfer Protocol (HTTP), described in its version 1.1 1167 in RFC 7230 to 7237, is a request-response application protocol 1168 developed throughout the 1990s, and factually contributed to the 1169 exponential growth of the Internet and the inter-connection of 1170 populations around the world. Its simple design strongly contributed 1171 to the fact that HTTP has become the foundation of most modern 1172 Internet platforms and communication systems, from websites, to chat 1173 systems, and computer-to-computer applications. In its manifestation 1174 with the World Wide Web, HTTP radically revolutionized the course of 1175 technological development and the ways people interact with online 1176 content and with each other. 1178 However, HTTP is also a fundamentally insecure protocol, that doesn't 1179 natively provide encryption properties. While the definition of the 1180 Secure Sockets Layer (SSL) [RFC6101], and later of Transport Layer 1181 Security (TLS)[RFC5246], also happened during the 1990s, the fact 1182 that HTTP doesn't mandate the use of such encryption layers to 1183 developers and service providers, was one of the reasons for a very 1184 late adoption of encryption. Only in the middle of the 2000s did we 1185 observe big Internet service providers, such as Google, starting to 1186 provide encrypted access to their web services. 1188 The lack of sensitivity and understanding of the critical importance 1189 of securing web traffic incentivized certain (offensive) actors to 1190 develop, deploy and utilize at large interception systems and later 1191 active injection attacks, in order to swipe large amounts of data, 1192 compromise Internet-enabled devices. The commercial availability of 1193 systems and tools to perform these types of attacks also led to a 1194 number of human rights abuses that have been discovered and reported 1195 over the years. 1197 Generally we can identify in Traffic Interception and Traffic 1198 Manipulation the two most problematic attacks that can be performed 1199 against applications employing a clear-text HTTP transport layer. 1200 That being said, the IETF is taking steady steps to move to the 1201 encrypted version of HTTP, HTTPSecure (HTTPS). 1203 While this is commendable, we must not lose track of the fact that 1204 different protocols, implementations, configurations and networking 1205 paradigms can intersect such that they (can be used to) adversely 1206 impact human rights. For instance, certain countries will throttle 1207 HTTPS connections forcing users to switch to the (unthrottled) HTTP 1208 to facilitate surveillance [Aryanetall]. 1210 5.2.3.3.1. Traffic Interception 1212 While we are seeing an increasing trend in the last couple of years 1213 to employ SSL/TLS as a secure traffic layer for HTTP-based 1214 applications, we are still far from seeing an ubiquitous use of 1215 encryption on the World Wide Web. It is important to consider that 1216 the adoption of SSL/TLS is also a relatively recent phenomena. 1217 E-mail providers such as riseup.net were the first ones to enable SSL 1218 by default. Google introduced an option for its GMail users to 1219 navigate with SSL only in 2008 [Rideout], and turned TLS on by 1220 default later in 2010 [Schillace]. It took an increasing amount of 1221 security breaches and revelations on global surveillance from Edward 1222 Snowden to have other mail service providers to follow suit. For 1223 example, Yahoo enabled SSL/TLS by default on its webmail services 1224 only towards the end of 2013 [Peterson]. 1226 TLS itself has been subject to many attacks and bugs which can be 1227 attributed to some fundamental design weaknesses such as lack of a 1228 state machine, which opens a vulnerability for a Triple Handshake 1229 Attack, and flaws caused by early U.S. government restrictions on 1230 cryptography, leading to cipher-suite downgrade attacks (Logjam 1231 attack). These vulnerabilities are being corrected in TLS1.3. 1232 [Bhargavan] [Adrian] 1234 HTTP upgrading to HTTPS is also vulnerable to having an attacker 1235 remove the "S" in any links to HTTPS URIs from a web-page transferred 1236 in cleartext over HTTP, an attack called "SSL Stripping" [sslstrip]. 1237 Thus, for high security use of HTTPS IETF standards such as HSTS 1238 [RFC6797], certificate pinning [RFC7469] and/or DANE [RFC6698] should 1239 be used. 1241 As we learned through the Snowden's revelations, intelligence 1242 agencies have been intercepting and collecting unencrypted traffic at 1243 large for many years. There are documented examples of such mass 1244 surveillance programs with GCHQ's TEMPORA [WP-Tempora] and NSA's 1245 XKEYSCORE [Greenwald]. Through these programs NSA/GCHQ have been 1246 able to swipe large amounts of data including email and instant 1247 messaging communications which have been transported by the 1248 respective providers in clear for years, unsuspecting of the 1249 pervasiveness and scale of governments' efforts and investment into 1250 global mass surveillance capabilities. 1252 However, similar mass interception of unencrypted HTTP communications 1253 is also often employed at a nation-level by some democratic countries 1254 by exercising control over state-owned Internet Service Providers 1255 (ISP) and through the use of commercially available monitoring, 1256 collection, and censorship equipment. Over the last few years a lot 1257 of information has come to public attention on the role and scale of 1258 a surveillance industry dedicated to develop interception gear of 1259 different types, making use of known and unknown weaknesses in 1260 existing protocols [RFC7258]. We have several records of such 1261 equipment being sold and utilized by some regimes in order to monitor 1262 entire segments of population especially at times of social and 1263 political distress, uncovering massive human rights abuses. For 1264 example, in 2013 the group Telecomix revealed that the Syrian regime 1265 was making use of BlueCoat products in order to intercept clear-text 1266 traffic as well as to enforce censorship of unwanted content [RSF]. 1267 Similarly in 2012 it was found that the French Amesys provided the 1268 Gaddafi's government with equipment able to intercept emails, 1269 Facebook traffic, and chat messages at a country level [WSJ]. The 1270 use of such systems, especially in the context of the Arab Spring and 1271 of civil uprisings against the dictatorships, has caused serious 1272 concerns of significant human rights abuses in Libya. 1274 5.2.3.3.2. Traffic Manipulation 1276 The lack of a secure transport layer under HTTP connections not only 1277 exposes the users to interception of the content of their 1278 communications, but is more and more commonly abused as a vehicle for 1279 actively compromising computers and mobile devices. If an HTTP 1280 session travels in the clear over the network, any node positioned at 1281 any point in the network is able to perform man-in-the-middle attacks 1282 and observe, manipulate, and hijack the session and modify the 1283 content of the communication in order to trigger unexpected behavior 1284 by the application generating the traffic. For example, in the case 1285 of a browser the attacker would be able to inject malicious code in 1286 order to exploit vulnerabilities in the browser or any of its 1287 plugins. Similarly, the attacker would be able to intercept, add 1288 malware, and repackage binary software updates that are very commonly 1289 downloaded in clear by applications such as word processors and media 1290 players. If the HTTP session would be encrypted, the tampering of 1291 the content would not be possible, and these network injection 1292 attacks would not be successful. 1294 While traffic manipulation attacks have been long known, documented, 1295 and prototyped especially in the context of WiFi and LAN networks, in 1296 the last few years we observed an increasing investment into the 1297 production and sale of network injection equipment both available 1298 commercially as well as deployed at scale by intelligence agencies. 1300 For example, we learned from some of the documents provided by Edward 1301 Snowden to the press, that the NSA has constructed a global network 1302 injection infrastructure, called QUANTUM, able to leverage mass 1303 surveillance in order to identify targets of interests and 1304 subsequently task man-on-the-side attacks to ultimately compromise a 1305 selected device. Among other attacks, NSA makes use of an attack 1306 called QUANTUMINSERT [Haagsma] which intercepts and hijacks an 1307 unencrypted HTTP communication and forces the requesting browser to 1308 redirect to a host controlled by NSA instead of the intended website. 1309 Normally, the new destination would be an exploitation service, 1310 referred in Snowden documents as FOXACID, which would attempt at 1311 executing malicious code in the context of the target's browser. The 1312 Guardian reported in 2013 that NSA has for example been using these 1313 techniques to target users of the popular anonymity service Tor 1314 [Schneier]. The German NDR reported in 2014 that NSA has also been 1315 using its mass surveillance capabilities to identify Tor users at 1316 large [Appelbaum]. 1318 Recently similar capabilities of Chinese authorities have been 1319 reported as well in what has been informally called the "Great 1320 Cannon" [Marcak], which raised numerous concerns on the potential 1321 curb on human rights and freedom of speech due to the increasing 1322 tighter control of Chinese Internet communications and access to 1323 information. 1325 Network injection attacks are also made widely available to state 1326 actors around the world through the commercialization of similar, 1327 smaller scale equipment that can be easily acquired and deployed at a 1328 country-wide level. Certain companies are known to have network 1329 injection gear within their products portfolio [Marquis-Boire]. The 1330 technology devised and produced by some of them to perform network 1331 traffic manipulation attacks on HTTP communications is even the 1332 subject of a patent application in the United States [Googlepatent]. 1333 Access to offensive technologies available on the commercial lawful 1334 interception market has led to human rights abuses and illegitimate 1335 surveillance of journalists, human rights defenders, and political 1336 activists in many countries around the world [Collins]. While 1337 network injection attacks haven't been the subject of much attention, 1338 they do enable even unskilled attackers to perform silent and very 1339 resilient compromises, and unencrypted HTTP remains one of the main 1340 vehicles. 1342 There is a new version of HTTP, called HTTP/2, which was published as 1343 [RFC7540] and which aimed to be largely backwards compatible but also 1344 offer new option such as data compression of HTTP headers and 1345 pipelining of request and multiplexing multiple requests over a 1346 single TCP connection. In addition to decreasing latency to improve 1347 page loading speeds it also facilitates more efficient use of 1348 connectivity in low-bandwith environments, which is an enabler for 1349 freedom of expression, the right to assembly, right to political 1350 participation and the right to participate in cultural life, art and 1351 science. [RFC7540] does not mandate Transport Layer Security or any 1352 other form of encryption, is also does not support opportunistic 1353 encryption, so the vulnerabilities listed above for HTTP/1 are also 1354 valid for HTTP/2 as defined in [RFC7540]. 1356 5.2.3.4. XMPP 1358 The Extensible Messaging and Presence Protocol (XMPP), specified in 1359 [RFC6120], provides a standard for interactive chat messaging, and 1360 has evolved to encompass interoperable text, voice, and video chat. 1361 The protocol is structured as a federated network of servers, similar 1362 to email, where users register with a local server which acts one 1363 their behalf to cache and relay messages. This protocol design has 1364 many advantages, allowing servers to shield clients from denial of 1365 service and other forms of retribution for their expression, and 1366 designed to avoid central entities which could control the ability to 1367 communicate or assemble using the protocol. 1369 None-the-less, there are plenty of aspects of the protocol design of 1370 XMPP which shape the ability for users to communicate freely, and to 1371 assembly through the protocol. 1373 5.2.3.4.1. User Identification 1375 The XMPP specification dictates that clients are identified with a 1376 resource (node@domain/home [1] / node@domain/work [2]) to distinguish 1377 the conversations to specific devices. While the protocol does not 1378 specify that the resource must be exposed by the client's server to 1379 remote users, in practice this has become the default behavior. In 1380 doing so, users can be tracked by remote friends and their servers, 1381 who are able to monitor presence not just of the user, but of each 1382 individual device the user logs in with. This has proven to be 1383 misleading to many users [pidgin], since many clients only expose 1384 user level rather than device level presence. Likewise, user 1385 invisibility so that communication can occur while users don't notify 1386 all buddies and other servers of their availability is not part of 1387 the formal protocol, and has only been added as an extension within 1388 the XML stream rather than enforced by the protocol. 1390 5.2.3.4.2. Surveillance of Communication 1392 The XMPP protocol specifies the standard by which communication of 1393 channels may be encrypted, but it does not provide visibility to 1394 clients of whether their communications are encrypted on each link. 1395 In particular, even when both clients ensure that they have an 1396 encrypted connection to their XMPP server to ensure that their local 1397 network is unable to read or disrupt the messages they send, the 1398 protocol does not provide visibility into the encryption status 1399 between the two servers. As such, clients may be subject to 1400 selective disruption of communications by an intermediate network 1401 which disrupts communications based on keywords found through Deep 1402 Packet Inspection. While many operators have commited to only 1403 establishing encrypted links from their servers in recognition of 1404 this vulnerability, it remains impossible for users to audit this 1405 behavior and encrypted connections are not required by the protocol 1406 itself [xmppmanifesto]. 1408 In particular, section 13.14 of the protocol specification [RFC6120] 1409 explicitly acknowledges the existence of a downgrade attack where an 1410 adversary controlling an intermediate network can force the inter 1411 domain federation between servers to revert to a non-encrypted 1412 protocol were selective messages can then be disrupted. 1414 5.2.3.4.3. Group Chat Limitations 1416 Group chat in the XMPP protocol is defined as an extension within the 1417 XML specification of the XMPP protocol (https://xmpp.org/extensions/ 1418 xep-0045.html). However, it is not encoded or required at a protocol 1419 level, and not uniformly implemented by clients. 1421 The design of multi-user chat in the XMPP protocol suffers from 1422 extending a protocol that was not designed with assembly of many 1423 users in mind. In particular, in the federated protocol provided by 1424 XMPP, multi-user communities are implemented with a distinguished 1425 'owner', who is granted control over the participants and structure 1426 of the conversation. 1428 Multi-user chat rooms are identified by a name specified on a 1429 specific server, so that while the overall protocol may be federated, 1430 the ability for users to assemble in a given community is moderated 1431 by a single server. That server may block the room and prevent 1432 assembly unilaterally, even between two users neither of whom trust 1433 or use that server directly. 1435 5.2.3.5. Peer to Peer 1437 Peer-to-Peer (P2P) is a distributed network architecture [RFC5694] in 1438 which all the participant nodes can be responsible for the storage 1439 and dissemination of information from any other node (defined in 1440 [RFC7574], an IETF standard that used a P2P architecture). A P2P 1441 network is a logical overlay that lives on top of the physical 1442 network, and allows nodes (or "peers") participating to it to 1443 establish contact and exchange information directly from one to each 1444 other. The implementation of a P2P network may very widely: it may 1445 be structured or unstructured, and it may implement stronger or 1446 weaker cryptographic and anonymity properties. While its most common 1447 application has traditionally been file-sharing (and other types of 1448 content delivery systems), P2P is a popular architecture for networks 1449 and applications that require (or encourage) decentralization. A 1450 prime example is Bitcoin (and similar cryptocurrencies), as well as 1451 Bitcoin and proprietary multimedia applications. 1453 In a time of heavily centralized online services, peer-to-peer is 1454 regularly described as an alternative, more democratic, and resistant 1455 option that displaces structures of control over data and 1456 communications and delegates all peers equally to be responsible for 1457 the functioning, integrity, and security of the data. While in 1458 principle peer-to-peer remains imporant to the design and development 1459 of future content distribution, messaging, and publishing systems, it 1460 poses numerous security and privacy challenges which are mostly 1461 delegated to individual developers to recognize, analyze, and solve 1462 in each implementation of a given P2P network. 1464 5.2.3.5.1. Network Poisoning 1466 Since content, and in some occasions peer lists, are safeguarded and 1467 distributed by its members, P2P networks are prone to what are 1468 generally defined as "poisoning attacks". Poisoning attacks might be 1469 aimed directly at the data that is being distributed, for example by 1470 intentionally corrupting it, or at the index tables used to instruct 1471 the peers where to fetch the data, or at routing tables, with the 1472 attempt of providing connecting peers with lists of rogue or non- 1473 existing peers, with the intention to effectively cause a Denial of 1474 Service on the network. 1476 5.2.3.5.2. Throttling 1478 Peer-to-Peer traffic (and BitTorrent in particular) represents a 1479 significant percentage of global Internet traffic [Sandvine] and it 1480 has become increasingly popular for Internet Service Providers to 1481 perform throttling of customers lines in order to limit bandwidth 1482 usage [torrentfreak1] and sometimes probably as an effect of the 1483 ongoing conflict between copyright holders and file-sharing 1484 communities [wikileaks]. Such throttling undermines the end-to-end 1485 principle. 1487 Throttling the peer-to-peer traffic makes some uses of P2P networks 1488 ineffective and it might be coupled with stricter inspection of 1489 users' Internet traffic through Deep Packet Inspection techniques 1490 which might pose additional security and privacy risks. 1492 5.2.3.5.3. Tracking and Identification 1494 One of the fundamental and most problematic issues with traditional 1495 peer-to-peer networks is a complete lack of anonymization of its 1496 users. For example, in the case of BitTorrent, all peers' IP 1497 addresses are openly available to the other peers. This has lead to 1498 an ever-increasing tracking of peer-to-peer and file-sharing users 1499 [ars]. As the geographical location of the user is directly exposed, 1500 and so could be his identity, the user might become target of 1501 additional harassment and attacks, being of physical or legal nature. 1502 For example, it is known that in Germany law firms have made 1503 extensive use of peer-to-peer and file-sharing tracking systems in 1504 order to identify downloaders and initiate legal actions looking for 1505 compensations [torrentfreak2]. 1507 It is worth noting that there are varieties of P2P networks that 1508 implement cryptographic practices and that introduce anonymization of 1509 its users. Such implementations may be proved to be successful in 1510 resisting censorship of content, and tracking of the network peers. 1511 A primary example is FreeNet [freenet1], a free software application 1512 designed to significantly increase the difficulty of users and 1513 content identification, and dedicated to foster freedom of speech 1514 online [freenet2]. 1516 5.2.3.5.4. Sybil Attacks 1518 In open-membership P2P networks, a single attacker can pretend to be 1519 many participants, typically by creating multiple fake identities of 1520 whatever kind the P2P network uses [Douceur]. Attackers can use 1521 Sybil attacks to bias choices the P2P network makes collectively 1522 toward the attacker's advantage, e.g., by making it more likely that 1523 a particular data item (or some threshold of the replicas or shares 1524 of a data item) are assigned to attacker-controlled participants. If 1525 the P2P network implements any voting, moderation, or peer review- 1526 like functionality, Sybil attacks may be used to "stuff the ballots" 1527 toward the attacker's benefit. Companies and governments can use 1528 Sybil attacks on discussion-oriented P2P systems for "astroturfing" 1529 or creating the appearance of mass grassroots support for some 1530 position where there is none in reality. It is important to know 1531 that there are no known complete, environmentally sustainable, and 1532 fully distributed solutions to Sybil attacks, and routing via 1533 'friends' allows users to be de-anonymized via their social graph. 1534 It is important to note that Sybil attacks in this context (e.f. 1535 astroturfing) are relevant to more than P2P protocols. And are also 1536 common on web based systems, and exploited by governments and 1537 commercial entitities. 1539 Encrypted P2P and Anonymous P2P networks already emerged and provided 1540 viable platforms for sharing material [tribler], publish content 1541 anonymously, and communicate securely [bitmessage]. These platforms 1542 are not perfect, and more research needs to be done. If adopted at 1543 large, well-designed and resistant P2P networks might represent a 1544 critical component of a future secure and distributed Internet, 1545 enabling freedom of speech and freedom of information at scale. 1547 5.2.3.6. Virtual Private Network 1549 The Virtual Private Networks (VPN) that are being discussed here are 1550 point-to-point connections that enables two computers to communicate 1551 over an encrypted tunnel. There are multiple implementations and 1552 protocols used in the deployment of VPNs, and they generally 1553 diversify by encryption protocol or particular requirements, most 1554 commonly in proprietary and enterprise solutions. VPNs are used 1555 commonly either to enable some devices to communicate through 1556 peculiar network configurations, or in order to use some privacy and 1557 security properties in order to protect the traffic generated by the 1558 end user; or both. VPNs have also become a very popular technology 1559 among human rights defenders, dissidents, and journalists worldwide 1560 to avoid local monitoring and eventually also to circumvent 1561 censorship. Among human rights defenders VPNs are often debated as a 1562 potential alternative to Tor or other anonymous networks. Such 1563 comparison is misleading, as some of the privacy and security 1564 properties of VPNs are often misunderstood by less tech-savvy users, 1565 which could ultimately lead to unintended problems. 1567 As VPNs increased in popularity, commercial VPN providers have 1568 started growing in business and are very commonly picked by human 1569 rights defenders and people at risk, as they are normally provided 1570 with an easy-to-use service and sometimes even custom applications to 1571 establish the VPN tunnel. Not being able to control the 1572 configuration of the network, and even less so the security of the 1573 application, assessing the general privacy and security state of 1574 common VPNs is very hard. Often such services have been discovered 1575 leaking information, and their custom applications have been found 1576 flawed. While Tor and similar networks receive a lot of scrutiny 1577 from the public and the academic community, commercial or non- 1578 commercial VPN networks are way less analyzed and understood 1579 [Insinuator] [Alshalanetal] , and it might be valuable to establish 1580 some standards to guarantee a minimal level of privacy and security 1581 to those who need them the most. 1583 5.2.3.6.1. No anonymity against VPN provider 1585 One of the common misconceptions among users of VPNs is the level of 1586 anonymity VPN can provide. This sense of anonymity can be betrayed 1587 by a number of attacks or misconfigurations of the VPN provider. It 1588 is important to remember that, in contrast to Tor and similar 1589 systems, VPN was not designed to provide anonymity properties. From 1590 a technical point of view, the VPN might leak identifiable 1591 information, or might be subject of correlation attacks that could 1592 expose the originating address of the connecting user. Most 1593 importantly, it is vital to understand that commercial and non- 1594 commercial VPN providers are bound by the law of the jurisdiction 1595 they reside in or in which their infrastructure is located, and they 1596 might be legally forced to turn over data of specific users if legal 1597 investigations or intelligence requirements dictate so. In such 1598 cases, if the VPN providers retain logs, it is possible that the 1599 information of the user is provided to the user's adversary and leads 1600 to his or her identification. 1602 5.2.3.6.2. Logging 1604 With VPN being point-to-point connections, the service providers are 1605 in fact able to observe the original location of the connecting users 1606 and they are able to track at what time they started their session 1607 and eventually also to which destinations they're trying to connect 1608 to. If the VPN providers retain logs for long enough, they might be 1609 forced to turn over the relevant data or they might be otherwise 1610 compromised, leading to the same data getting exposed. A clear log 1611 retaining policy could be enforced, but considerig that countries 1612 enforce different levels of data retention policies, VPN providers 1613 should at least be transparent on what information do they store and 1614 for how long is being kept. 1616 5.2.3.6.3. 3rd Party Hosting 1618 VPN providers very commonly rely on 3rd parties to provision the 1619 infrastructure that is later going to be used to run VPN endpoints. 1620 For example, they might rely on external dedicated server hosting 1621 providers, or on uplink providers. In those cases, even if the VPN 1622 provider itself isn't retaining any significant logs, the information 1623 on the connecting users might be retained by those 3rd parties 1624 instead, introducing an additional collection point for the 1625 adversary. 1627 5.2.3.6.4. IPv6 Leakage 1629 Some studies proved that several commercial VPN providers and 1630 applications suffer of critical leakage of information through IPv6 1631 due to improper support and configuration [PETS2015VPN]. This is 1632 generally caused by a lack of proper configuration of the client's 1633 IPv6 routing tables. Considering that most popular browsers and 1634 similar applications have been supporting IPv6 by default, if the 1635 host is provided with a functional IPv6 configuration, the traffic 1636 that is generated might be leaked if the VPN application isn't 1637 designed to manipulate such traffic properly. 1639 5.2.3.6.5. DNS Leakage 1641 Similarly, VPN services that aren't handling DNS requests and are not 1642 running DNS servers of their own, might be prone to DNS leaking which 1643 might not only expose sensitive information on the activity of the 1644 user, but could also potentially lead to DNS hijacking attacks and 1645 following compromises. 1647 5.2.3.6.6. Traffic Correlation 1649 Some implementations of VPN appear to be particularly vulnerable to 1650 identification and collection of key exchanges which, some Snowden 1651 documents revealed, are systematically collected and stored for 1652 future reference. The ability of an adversary to monitor network 1653 connections at many different points over the Internet, can allow 1654 them to perform traffic correlation attacks and identify the origin 1655 of certain VPN traffic by cross referencing the connection time of 1656 the user to the endpoint and the connection time of the endpoint to 1657 the final destination. These types of attacks, although very 1658 expensive and normally only performed by very resourceful 1659 adversaries, have been documented [spiegel] to be already in practice 1660 and could completely vanify the use of a VPN and ultimately expose 1661 the activity and the identity of a user at risk. 1663 5.2.3.7. HTTP Status Code 451 1665 Every Internet user has run into the '404 Not Found' Hypertext 1666 Transfer Protocol (HTTP) status code when trying, and failing, to 1667 access a particular website [Cath]. It is a response status that the 1668 server sends to the browser, when the server cannot locate the URL. 1669 '403 Forbidden' is another example of this class of code signals that 1670 gives users information about what is going on. In the '403' case 1671 the server can be reached, but is blocking the request because the 1672 user is trying to access content forbidden to them. This can be 1673 because the specific user is not allowed access to the content (like 1674 a government employee trying to access pornography on a work- 1675 computer) or because access is restricted to all users (like social 1676 network sites in certain countries). As surveillance and censorship 1677 of the Internet is becoming more commonplace, voices were raised at 1678 the IETF to introduce a new status code that indicates when something 1679 is not available for 'legal reasons' (like censorship): 1681 The 451 status code would allow server operators to operate with 1682 greater transparency in circumstances where issues of law or public 1683 policy affect their operation. This transparency may be beneficial 1684 both to these operators and to end-users [RFC7725]. 1686 The status code is named '451', a reference to Bradbury's famous 1687 novel on censorship, and the temperature (in Fahrenheit) at which 1688 bookpaper autoignites. 1690 During the IETF92 meeting in Dallas, there was discussion about the 1691 usefulness of '451'. The main tension revolved around the lack of an 1692 apparent machine-readable technical use of the information. The 1693 extent to which '451' is just 'political theatre' or whether it has a 1694 concrete technical use was heatedly debated. Some argued that 'the 1695 451 status code is just a status code with a response body' others 1696 said it was problematic because 'it brings law into the picture'. 1697 Again others argued that it would be useful for individuals, or 1698 organizations like the 'Chilling Effects' project, crawling the web 1699 to get an indication of censorship (IETF discussion on '451' - 1700 author's field notes March 2015). There was no outright objection 1701 during the Dallas meeting against moving forward on status code 1702 '451', and on December 18, 2015 the Internet Engineering Steering 1703 Group approved publication of 'An HTTP Status Code to Report Legal 1704 Obstacles'. It is now an IETF approved HTTP status code to signal 1705 when resource access is denied as a consequence of legal demands 1706 [RFC7725]. 1708 What is interesting about this particular case is that not only 1709 technical arguments but also the status code's outright potential 1710 political use for civil society played a substantial role in shaping 1711 the discussion, and the decision to move forward with this 1712 technology. 1714 It is nonetheless important to note that HTTP status code 451 is not 1715 a solution to detect all occasions of censorship. A large swath of 1716 Internet filtering occurs in the network rather than the server 1717 itself. For these forms of censorship 451 plays a limited role, as 1718 the servers will not be able to send the code, because they haven't 1719 received the requests (as is the case with servers with resources 1720 blocked by the Chinese Golden shield). Such filtering regimes are 1721 unlikely to voluntarily inject a 451 status code. The use of 451 is 1722 most likely to apply in the case of cooperative, legal versions of 1723 content removal resulting from requests to providers. One can think 1724 of content that is removed or blocked for legal reasons, like 1725 copyright infringement, gambling laws, child abuse, et cetera. Large 1726 Internet companies and search engines are constantly asked to censor 1727 content in various jurisdictions. 451 allows this to be easily 1728 discovered, for instance by initiatives like the Lumen Database. 1730 Overall, the strength of 451 lies in its ability to provide 1731 transparency by giving the reason for blocking, and giving the end- 1732 user the ability to file a complaint. It allows organizations to 1733 easily measure censorship in an automated way, and prompts the user 1734 to access the content via another path (e.g. TOR, VPNs) when (s)he 1735 encounters the 451 status code. 1737 Status code 451 impact human rights by making censorship more 1738 transparent and measurable. The status code increases transparency 1739 both by signaling the existence of censorship (instead of a much more 1740 broad HTTP error message like HTTP status code 404) as well as 1741 providing details of the legal restriction, which legal authority is 1742 imposing it, and what class of resources it applies to. This 1743 empowers the user to seek redress. 1745 5.2.3.8. DDoS attacks 1747 Many individuals, not excluding IETF engineers, have argued that DDoS 1748 attacks are fundamentally against freedom of expression. Technically 1749 DDoS attacks are when one or multiple host overload the bandwidth or 1750 resources of another host by flooding it with traffic or making 1751 resource intensive requests, causing it to temporarily stop being 1752 available to users. One can roughly differentiate three types of 1753 DDoS attacks: Volume Based Attacked (This attack aims to make the 1754 host unreachable by using up all it's bandwith, often used techniques 1755 are: UDP floods and ICMP floods), Protocol Attacks (This attacks aims 1756 to use up actual server resources, often used techniques are SYN 1757 floods, fragmented packet attacks, and Ping of Death [RFC4949]) and 1758 Application Layer Attacks (this attack aims to bring down a server, 1759 such as the webserver). 1761 DDoS attacks can thus stifle freedom of expression, complicate the 1762 ability of independent media and human rights organizations to 1763 exercise their right to (online) freedom of association, while 1764 facilitating the ability of governments to censor dissent. When it 1765 comes to comparing DDoS attacks to protests in offline life, it is 1766 important to remember that only a limited number of DDoS attacks 1767 involved solely willing participants. In the overwhelming majority 1768 of cases, the clients are hacked hosts of unrelated parties that have 1769 not consented to being part of a DDoS (for exceptions see Operation 1770 Abibil [Abibil] or the Iranian Green Movement DDoS [GreenMovement]). 1771 In addition, DDoS attacks are increasingly used as an extortion 1772 tactic. 1774 All of these issues seem to suggest that the IETF should try to 1775 ensure that their protocols cannot be used for DDoS attacks, which is 1776 consistent with the long-standing IETF consensus that DDoS is an 1777 attack that protocols should mitigate them to the extent they can 1778 [BCP72]. Decreasing the number of vulnerabilities in protocols and 1779 (outside of IETF) the number of bugs in the network stacks of routers 1780 or computers could address this issue. The IETF can clearly play a 1781 role in bringing about some of these changes but the IETF cannot be 1782 expected to take a positive stance on (specific) DDoS attacks, or 1783 create protocols to enable some attacks and inhibit others. What the 1784 IETF can do is critically reflect on its role in the development of 1785 the Internet, and how this impacts the ability of people to excercise 1786 their human rights, such as freedom of expression. 1788 6. Model for developing human rights protocol considerations 1790 This section outlines a set of human rights protocol considerations 1791 for protocol developers. It provides questions engineers should ask 1792 themselves when developing or improving protocols if they want to 1793 understand their human rights impact. It should however be noted 1794 that the impact of a protocol cannot solely be deduced from its 1795 design, but its usage and implementation should also be studied to 1796 form a full protocol human rights impact assessment. 1798 The questions are based on the research performed by the hrpc 1799 research group which has been documented before these considerations. 1800 The research establishes that human rights relate to standards and 1801 protocols and offers a common vocabulary of technical concepts that 1802 impact human rights and how these technical concept can be combined 1803 to ensure that the Internet remains an enabling environment for human 1804 rights. With this the contours of a model for developing human 1805 rights protocol considerations has taken shape. 1807 6.1. Human rights threats 1809 Human rights threats on the Internet come in a myriad of forms. 1810 Protocols and standards can harm or enable the right to freedom of 1811 expression, right to non-discrimination, right to equal protection, 1812 right to participate in cultural life, arts and science, right to 1813 freedom of assembly and association, and the right to security. An 1814 end-user who is denied access to certain services, data or websites 1815 may be unable to disclose vital information about the malpractices of 1816 a government or other authority. A person whose communications are 1817 monitored may be prevented from exercising their right to freedom of 1818 association or participate in political processes [Penney]. In a 1819 worst-case scenario, protocols that leak information can lead to 1820 physical danger. A realistic example to consider is when individuals 1821 perceived as threats to the state are subjected to torture or 1822 extrajudicial killing or detention on the basis of information 1823 gathered by state agencies through information leakage in protocols. 1825 This section details several 'common' threats to human rights, 1826 indicating how each of these can lead to human rights violations/ 1827 harms and present several examples of how these threats to human 1828 rights materialize on the Internet. This threat modeling is inspired 1829 by [RFC6973] Privacy Considerations for Internet Protocols, which is 1830 based on the security threat analysis. This method is by no means a 1831 perfect solution for assessing human rights risks in Internet 1832 protocols and systems; it is however the best approach currently 1833 available. Certain specific human rights threats are indirectly 1834 considered in Internet protocols as part of the security 1835 considerations [BCP72], but privacy guidelines [RFC6973] or reviews, 1836 let alone human rights impact assessments of protocols are not 1837 standardized or implemented. 1839 Many threats, enablers and risks are linked to different rights. 1840 This is not unsurprising if one takes into account that human rights 1841 are interrelated, interdependent and indivisible. Here however we're 1842 not discussing all human rights because not all human rights are 1843 relevant to ICTs in general and protocols and standards in particular 1844 [Bless]: "The main source of the values of human rights is the 1845 International Bill of Human Rights that is composed of the Universal 1846 Declaration of Human Rights [UDHR] along with the International 1847 Covenant on Civil and Political Rights [ICCPR] and the International 1848 Covenant on Economic, Social and Cultural Rights [ICESCR]. In the 1849 light of several cases of Internet censorship, the Human Rights 1850 Council Resolution 20/8 was adopted in 2012 [UNHRC2016], affirming ". 1851 . . that the same rights that people have offline must also be 1852 protected online. . . " . In 2015, the Charter of Human Rights and 1853 Principles for the Internet [IRP] was developed and released. 1854 According to these documents, some examples of human rights relevant 1855 for ICT systems are human dignity (Art. 1 UDHR), non-discrimination 1856 (Art. 2), rights to life, liberty and security (Art. 3), freedom of 1857 opinion and expression (Art. 19), freedom of assembly and association 1858 (Art. 20), rights to equal protection, legal remedy, fair trial, due 1859 process, presumed innocent (Art. 7-11), appropriate social and 1860 international order (Art. 28), participation in public affairs (Art. 1861 21), participation in cultural life, protection of intellectual 1862 property (Art. 27), and privacy (Art. 12)." A partial catalog of 1863 human rights related to ICTs, including economic rights, can be found 1864 in [Hill2014]. 1866 This is by no means an attempt to exclude specific rights or 1867 prioritize some rights over others. If other rights seem relevant, 1868 please contact the authors. 1870 6.2. Guidelines for human rights considerations 1872 This section provides guidance for document authors in the form of a 1873 questionnaire about protocols and their (potential) impact. The 1874 questionnaire may be useful at any point in the design process, 1875 particularly after document authors have developed a high-level 1876 protocol model as described in [RFC4101]. These guidelines do not 1877 seek to replace any existing referenced specifications, but rather 1878 contribute to them and look at the design process from a human rights 1879 perspective. 1881 Protocols and Internet Standard might benefit from a documented 1882 discussion of potential human rights risks arising from potential 1883 misapplications of the protocol or technology described in the RFC. 1884 This might be coupled with an Applicability Statement for that RFC. 1886 Note that the guidance provided in this section does not recommend 1887 specific practices. The range of protocols developed in the IETF is 1888 too broad to make recommendations about particular uses of data or 1889 how human rights might be balanced against other design goals. 1890 However, by carefully considering the answers to the following 1891 questions, document authors should be able to produce a comprehensive 1892 analysis that can serve as the basis for discussion on whether the 1893 protocol adequately takes specific human rights threats into account. 1894 This guidance is meant to help the thought process of a human rights 1895 analysis; it does not provide specific directions for how to write a 1896 human rights protocol considerations section (following the example 1897 set in [RFC6973]), and the addition of a human rights protocol 1898 considerations section has also not yet been proposed. In 1899 considering these questions, authors will need to be aware of the 1900 potential of technical advances or the passage of time to undermine 1901 protections. In general, considerations of rights are likely to be 1902 more effective if they are considered given a purpose and specific 1903 use cases, rather than as abstract absolute goals. 1905 6.2.1. Connectivity 1907 Question(s): Does your protocol add application-specific functions to 1908 intermediary nodes? Could this functionality be added to end nodes 1909 instead of intermediary nodes? Is your protocol optimized for low 1910 bandwidth and high latency connections? Could your protocol also be 1911 developed in a stateless manner? 1913 Explanation: The end-to-end principle [Saltzer] holds that 'the 1914 intelligence is end to end rather than hidden in the network' 1915 [RFC1958]. The end-to-end principle is important for the robustness 1916 of the network and innovation. Such robustness of the network is 1917 crucial to enabling human rights like freedom of expression. 1919 Example: Middleboxes (which can be Content Delivery Networks, 1920 Firewalls, NATs or other intermediary nodes that provide other 1921 'services' than routing) serve many legitimate purposes. But the 1922 protocols guiding them, can influence individuals' ability to 1923 communicate online freely and privately. The potential for abuse and 1924 intentional and unintentional censoring and limiting permissionless 1925 innovation, and thus ultimately the impact of middleboxes on the 1926 Internet as a place of unfiltered, unmonitored freedom of speech, is 1927 real. 1929 Impacts: 1931 - Right to freedom of expression 1933 - Right to freedom of assembly and association 1935 6.2.2. Privacy 1937 Question(s): Did you have a look at the Guidelines in the Privacy 1938 Considerations for Internet Protocols [RFC6973] section 7? Could 1939 your protocol in any way impact the confidentiality of protocol 1940 metadata? Could your protocol counter traffic analysis? Could you 1941 protocol improve data minimization? Does your document identify 1942 potentially sensitive logged data by your protocol and/or for how 1943 long that needs to be retained for technical reasons? 1945 Explanation: Privacy refers to the right of an entity (normally a 1946 person), acting in its own behalf, to determine the degree to which 1947 it will interact with its environment, including the degree to which 1948 the entity is willing to share its personal information with others. 1949 [RFC4949]. If a protocol provides insufficient privacy protection it 1950 may have a negative impact on freedom of expression as users self- 1951 censor for fear of surveillance, or find themselves unable to express 1952 themselves freely. 1954 Example: See [RFC6973] 1956 Impacts: 1958 - Right to freedom of expression 1960 - Right to non-discrimination 1962 6.2.3. Content agnosticism 1964 Question(s): If your protocol impacts packet handling, does it use 1965 user data (packet data that is not included in the header)? Is it 1966 making decisions based on the payload of the packet? Does your 1967 protocol prioritize certain content or services over others in the 1968 routing process ? Is the protocol transparent about the 1969 prioritization that is made (if any)? 1971 Explanation: Content agnosticism refers to the notion that network 1972 traffic is treated identically regardless of payload, with some 1973 exception where it comes to effective traffic handling, for instance 1974 where it comes to delay tolerant or delay sensitive packets, based on 1975 the header. 1977 Example: Content agnosticism prevents payload-based discrimination 1978 against packets. This is important because changes to this principle 1979 can lead to a two-tiered Internet, where certain packets are 1980 prioritized over others on the basis of their content. Effectively 1981 this would mean that although all users are entitled to receive their 1982 packets at a certain speed, some users become more equal than others. 1984 Impacts: 1986 - Right to freedom of expression 1988 - Right to non-discrimination 1990 - Right to equal protection 1992 6.2.4. Security 1994 Question(s): Did you have a look at Guidelines for Writing RFC Text 1995 on Security Considerations [BCP72]? Have you found any attacks that 1996 are out of scope for your protocol? Would these attacks be pertinent 1997 to the human rights enabling features of the Internet (as described 1998 throughout this document)? 2000 Explanation: Most people speak of security as if it were a single 2001 monolithic property of a protocol or system, however, upon 2002 reflection; one realizes that it is clearly not true. Rather, 2003 security is a series of related but somewhat independent properties. 2004 Not all of these properties are required for every application. 2005 Since communications are carried out by systems and access to systems 2006 is through communications channels, these goals obviously interlock, 2007 but they can also be independently provided [BCP72]. 2009 Example: See [BCP72]. 2011 Impacts: 2013 - Right to freedom of expression 2015 - Right to freedom of assembly and association 2017 - Right to non-discrimination 2019 - Right to security 2021 6.2.5. Internationalization 2023 Question(s): Does your protocol have text strings that have to be 2024 understood or entered by humans? Does your protocol allow Unicode? 2025 If so, do you have accept texts in one charset (which must be UTF-8), 2026 or several (which is dangerous for interoperability)? If character 2027 sets or encodings other than UTF-8 are allowed, does your protocol 2028 mandate a proper tagging of the charset? Did you have a look at 2029 [RFC6365]? Does your protocol allow Unicode encoded in UTF-8 only? 2030 If other character sets or encodings are allowed, does your protocol 2031 mandate a proper tagging of the charset? Did you have a look at 2032 [RFC6365]? 2034 Explanation: Internationalization refers to the practice of making 2035 protocols, standards, and implementations usable in different 2036 languages and scripts (see Localization). In the IETF, 2037 internationalization means to add or improve the handling of non- 2038 ASCII text in a protocol. [RFC6365] A different perspective, more 2039 appropriate to protocols that are designed for global use from the 2040 beginning, is the definition used by W3C: 2042 "Internationalization is the design and development of a 2043 product, application or document content that enables easy 2044 localization for target audiences that vary in culture, region, 2045 or language." {{W3Ci18nDef}} 2047 Many protocols that handle text only handle one charset (US-ASCII), 2048 or leave the question of what CCS and encoding are used up to local 2049 guesswork (which leads, of course, to interoperability problems). If 2050 multiple charsets are permitted, they must be explicitly identified 2051 [RFC2277]. Adding non-ASCII text to a protocol allows the protocol 2052 to handle more scripts, hopefully representing users across the 2053 world. In today's world, that is normally best accomplished by 2054 allowing Unicode encoded in UTF-8 only. 2056 In the current IETF policy [RFC2277], internationalization is aimed 2057 at user-facing strings, not protocol elements, such as the verbs used 2058 by some text-based protocols. (Do note that some strings are both 2059 content and protocol elements, such as the identifiers.) If the 2060 Internet wants to be a global network of networks, the protocols 2061 should work with other languages than English and other character 2062 sets than latin characters. It is therefore crucial that at least 2063 the content carried by the protocol can be in any script, and that 2064 all scripts are treated equally. 2066 Example: See localization 2068 Impacts: 2070 - Right to freedom of expression 2072 - Right to political participation 2074 - Right to participate in cultural life, arts and science 2076 6.2.6. Censorship resistance 2078 Question(s): Does this protocol introduce new identifiers or reuse 2079 existing identifiers (e.g. MAC addresses) that might be associated 2080 with persons or content? Does your protocol make it apparent or 2081 transparent when access to a resource it restricted? Can your 2082 protocol contribute to filtering in a way it could be implemented to 2083 censor data or services? Could this be designed to ensure this 2084 doesn't happen? 2086 Explanation: Censorship resistance refers to the methods and measures 2087 to prevent Internet censorship. 2089 Example: In the development of the IPv6 protocol it was discussed to 2090 embed a Media Access Control (MAC) address into unique IP addresses. 2091 This would make it possible for 'eavesdroppers and other information 2092 collectors to identify when different addresses used in different 2093 transactions actually correspond to the same node. [RFC4941] This is 2094 why Privacy Extensions for Stateless Address Autoconfiguration in 2095 IPv6 have been introduced. [RFC4941] 2097 Identifiers of content exposed within a protocol might be used to 2098 facilitate censorship, as in the case of Application Layer based 2099 censorship, which affects protocols like HTTP. Denial or restriction 2100 of access can be made apparent by the use of status code 451 - which 2101 allows server operators to operate with greater transparency in 2102 circumstances where issues of law or public policy affect their 2103 operation [RFC7725]. 2105 Impacts: 2107 - Right to freedom of expression 2109 - Right to political participation 2111 - Right to participate in cultural life, arts and science 2113 - Right to freedom of assembly and association 2115 6.2.7. Open Standards 2117 Question(s): Is your protocol fully documented in a way that it could 2118 be easily implemented, improved, built upon and/or further developed? 2119 Do you depend on proprietary code for the implementation, running or 2120 further development of your protocol? Does your protocol favor a 2121 particular proprietary specification over technically equivalent and 2122 competing specification(s), for instance by making any incorporated 2123 vendor specification "required" or "recommended" [RFC2026]? Do you 2124 normatively reference another standard that is not available without 2125 cost (and could it possible be done without)? Are you aware of any 2126 patents that would prevent your standard from being fully implemented 2127 [RFC3979] [RFC6701]? 2129 Explanation: The Internet was able to be developed into the global 2130 network of networks because of the existence of open, non-proprietary 2131 standards [Zittrain]. They are crucial for enabling 2132 interoperability. Yet, open standards are not explicitly defined 2133 within the IETF. On the subject, [RFC2026] states: Various national 2134 and international standards bodies, such as ANSI, ISO, IEEE, and ITU- 2135 T, develop a variety of protocol and service specifications that are 2136 similar to Technical Specifications defined at the IETF. National 2137 and international groups also publish "implementors' agreements" that 2138 are analogous to Applicability Statements, capturing a body of 2139 implementation-specific detail concerned with the practical 2140 application of their standards. All of these are considered to be 2141 "open external standards" for the purposes of the Internet Standards 2142 Process. Similarly, [RFC3935] does not define open standards but 2143 does emphasize the importance of 'open process': any interested 2144 person can participate in the work, know what is being decided, and 2145 make his or her voice heard on the issue. Part of this principle is 2146 the IETF's commitment to making its documents, WG mailing lists, 2147 attendance lists, and meeting minutes publicly available on the 2148 Internet. 2150 Open standards are important as they allow for permissionless 2151 innovation, which is important to maintain the freedom and ability to 2152 freely create and deploy new protocols on top of the communications 2153 constructs that currently exist. It is at the heart of the Internet 2154 as we know it, and to maintain its fundamentally open nature, we need 2155 to be mindful of the need for developing open standards. 2157 All standards that need to be normatively implemented should be 2158 freely available and with reasonable protection for patent 2159 infringement claims, so it can also be implemented in open source or 2160 free software. Patents have often held back open standardization or 2161 been used against those deploying open standards, particularly in the 2162 domain of cryptography [newegg]. An exemption of this is sometimes 2163 made when a protocol is standardized that normatively relies on 2164 speficiations produced by others SDOs that are not freely available. 2165 Patents in open standards or in normative references to other 2166 standards should have a patent disclosure [notewell], royalty-free 2167 licensing [patentpolicy], or some other form of reasonable 2168 protection. Reasonable patent protection should includes but is not 2169 limited to cryptographic primitives. 2171 Example: [RFC6108] describes a system for providing critical end-user 2172 notifications to web browsers, which has been deployed by Comcast, an 2173 Internet Service Provider (ISP). Such a notification system is being 2174 used to provide near-immediate notifications to customers, such as to 2175 warn them that their traffic exhibits patterns that are indicative of 2176 malware or virus infection. There are other proprietary systems that 2177 can perform such notifications, but those systems utilize Deep Packet 2178 Inspection (DPI) technology. In contrast to DPI, this document 2179 describes a system that does not rely upon DPI, and is instead based 2180 in open IETF standards and open source applications. 2182 Impacts: 2184 - Right to freedom of expression 2185 - Right to participate in cultural life, arts and science 2187 6.2.8. Heterogeneity Support 2189 Question(s): Does your protocol support heterogeneity by design? 2190 Does your protocol allow for multiple types of hardware? Does your 2191 protocol allow for multiple types of application protocols? Is your 2192 protocol liberal in what it receives and handles? Will it remain 2193 usable and open if the context changes? Does your protocol allow 2194 there to be well-defined extension points? Do these extension points 2195 allow for open innovation? 2197 Explanation: The Internet is characterized by heterogeneity on many 2198 levels: devices and nodes, router scheduling algorithms and queue 2199 management mechanisms, routing protocols, levels of multiplexing, 2200 protocol versions and implementations, underlying link layers (e.g., 2201 point-to-point, multi-access links, wireless, FDDI, etc.), in the 2202 traffic mix and in the levels of congestion at different times and 2203 places. Moreover, as the Internet is composed of autonomous 2204 organizations and Internet service providers, each with their own 2205 separate policy concerns, there is a large heterogeneity of 2206 administrative domains and pricing structures. As a result, the 2207 heterogeneity principle proposed in [RFC1958] needs to be supported 2208 by design [FIArch]. 2210 Example: Heterogeneity is inevitable and needs be supported by 2211 design. Multiple types of hardware must be allowed for, e.g. 2212 transmission speeds differing by at least 7 orders of magnitude, 2213 various computer word lengths, and hosts ranging from memory-starved 2214 microprocessors up to massively parallel supercomputers. Multiple 2215 types of application protocol must be allowed for, ranging from the 2216 simplest such as remote login up to the most complex such as 2217 distributed databases [RFC1958]. 2219 Impacts: 2221 - Right to freedom of expression 2223 - Right to freedom of expression 2225 - Right to political participtation 2227 6.2.9. Anonymity 2229 Question(s): Did you have a look at the Privacy Considerations for 2230 Internet Protocols [RFC6973], especially section 6.1.1 ? 2231 Explanation: Anonymity refers to the condition of an identity being 2232 unknown or concealed [RFC4949]. Even though full anonymity is hard 2233 to achieve, it is a non-binary concept. Making pervasive monitoring 2234 and tracking harder is important for many users as well as for the 2235 IETF [RFC7258]. Achieving a higher level of anonymity is an 2236 important feature for many end-users, as it allows them different 2237 degrees of privacy online. 2239 Example: Often standards expose private information, it is important 2240 to consider ways to mitigate the obvious privacy impacts. For 2241 instance, a feature which uses deep packet inspection or geolocation 2242 data could refuse to open this data to third parties, that might be 2243 able to connect the data to a physical person. 2245 Impacts: 2247 - Right to non-discrimination 2249 - Right to political participation 2251 - Right to freedom of assembly and association 2253 - Right to security 2255 6.2.10. Pseudonymity 2257 Question(s): Have you considered the Privacy Considerations for 2258 Internet Protocols [RFC6973], especially section 6.1.2 ? Does this 2259 specification collect personally derived data? Does the protocol 2260 generate or process anything that can be, or be tightly correlated 2261 with, personally identifiable information? Does the standard utilize 2262 data that is personally-derived, i.e. derived from the interaction of 2263 a single person, or their device or address? Does this specification 2264 generate personally derived data, and if so how will that data be 2265 handled? 2267 Explanation: Pseudonymity - the ability to use a persistent 2268 identifier not linked to one's offline identity" straight away - is 2269 an important feature for many end-users, as it allows them different 2270 degrees of disguised identity and privacy online. 2272 I would also include consideration that pseudonyms cannot be simply 2273 reverse engineered - some early approaches simply took approaches 2274 such as simple hashing of IP addreses. These could then be simply 2275 reversed by generating a hash for each potential IP address and 2276 comparing it to the pseudonym. 2278 Example: Designing a standard that exposes private information, it is 2279 important to consider ways to mitigate the obvious impacts. While 2280 pseudonyms cannot be simply reverse engineered - some early 2281 approaches simply took approaches such as simple hashing of IP 2282 addreses, these could then be simply reversed by generating a hash 2283 for each potential IP address and comparing it to the pseudonym - 2284 limiting the exposure of private information remains important. 2286 Pseudonymity means using a pseudonym instead of one's "real" name. 2287 There are many reasons for users to use pseudoyms, for instance to: 2288 hide their gender, protect themselves against harassment, protect 2289 their families' privacy, frankly discuss sexuality, or develop a 2290 artistic or journalistic persona without retribution from an 2291 employer, (potential) customers, or social surrounding. 2292 [geekfeminism] The difference between anonymity and pseudonymity is 2293 that a pseudonym often is persistent. "Pseudonymity is strengthened 2294 when less personal data can be linked to the pseudonym; when the same 2295 pseudonym is used less often and across fewer contexts; and when 2296 independently chosen pseudonyms are more frequently used for new 2297 actions (making them, from an observer's or attacker's perspective, 2298 unlinkable)." [RFC6973] 2300 Impacts: 2302 - Right to non-discrimination 2304 - Right to freedom of assembly and association 2306 6.2.11. Accessibility 2308 Question(s): Is your protocol designed to provide an enabling 2309 environment for people who are not able-bodied? Have you looked at 2310 the W3C Web Accessibility Initiative for examples and guidance? 2312 Explanation: The Internet is fundamentally designed to work for all 2313 people, whatever their hardware, software, language, culture, 2314 location, or physical or mental ability. When the Internet meets 2315 this goal, it is accessible to people with a diverse range of 2316 hearing, movement, sight, and cognitive ability [W3CAccessibility]. 2317 Sometimes in the design of protocols, websites, web technologies, or 2318 web tools, barriers are created that exclude people from using the 2319 Web. 2321 Example: The HTML protocol as defined in [HTML5] specifically 2322 requires that every image must have an alt attribute (with a few 2323 exceptions) to ensure images are accessible for people that cannot 2324 themselves decipher non-text content in web pages. 2326 Impacts: 2328 - Right to non-discrimination 2330 - Right to freedom of assembly and association 2332 - Right to education 2334 - Right to political participation 2336 6.2.12. Localization 2338 Question(s): Does your protocol uphold the standards of 2339 internationalization? Have made any concrete steps towards 2340 localizing your protocol for relevant audiences? 2342 Explanation: Localization refers to the adaptation of a product, 2343 application or document content to meet the language, cultural and 2344 other requirements of a specific target market (a locale) 2345 [W3Ci18nDef]. It is also described as the practice of translating an 2346 implementation to make it functional in a specific language or for 2347 users in a specific locale (see Internationalization). 2349 Example: The Internet is a global medium, but many of its protocols 2350 and products are developed with a certain audience in mind, that 2351 often share particular characteristics like knowing how to read and 2352 write in ASCII and knowing English. This limits the ability of a 2353 large part of the world's online population from using the Internet 2354 in a way that is culturally and linguistically accessible. An 2355 example of a protocol that has taken into account the view that 2356 individuals like to have access to data in their native language can 2357 be found in [RFC4646]. This protocol labels the information content 2358 with an identifier for the language in which it is written. And this 2359 allows information to be presented in more than one language. 2361 Impacts: 2363 - Right to non-discrimination 2365 - Right to participate in cultural life, arts and science 2367 - Right to freedom of expression 2369 6.2.13. Decentralization 2371 Question(s): Can your protocol be implemented without one single 2372 point of control? If applicable, can your protocol be deployed in a 2373 federated manner? What is the potential for discrimination against 2374 users of your protocol? How can the use of your protocol be used to 2375 implicate users? Does your protocol create additional centralized 2376 points of control? 2378 Explanation: Decentralization is one of the central technical 2379 concepts of the architecture of the networks, and embraced as such by 2380 the IETF [RFC3935]. It refers to the absence or minimization of 2381 centralized points of control; a feature that is assumed to make it 2382 easy for new users to join and new uses to unfold [Brown]. It also 2383 reduces issues surrounding single points of failure, and distributes 2384 the network such that it continues to function if one or several 2385 nodes are disabled. With the commercialization of the Internet in 2386 the early 1990's there has been a slow move to move away from 2387 decentralization, to the detriment of the technical benefits of 2388 having a decentralized Internet. 2390 Example: The bits traveling the Internet are increasingly susceptible 2391 to monitoring and censorship, from both governments and Internet 2392 service providers, as well as third (malicious) parties. The ability 2393 to monitor and censor is further enabled by the increased 2394 centralization of the network that creates central infrastructure 2395 points that can be tapped in to. The creation of peer-to-peer 2396 networks and the development of voice-over-IP protocols using peer- 2397 to-peer technology in combination with distributed hash table (DHT) 2398 for scalability are examples of how protocols can preserve 2399 decentralization [Pouwelse]. 2401 Impacts: 2403 - Right to freedom of expression 2405 - Right to freedom of assembly and association 2407 6.2.14. Reliability 2409 Question(s): Is your protocol fault tolerant? Does it degrade 2410 gracefully? Can your protocol resist malicious degradation attempts? 2411 Do you have a documented way to announce degradation? Do you have 2412 measures in place for recovery or partial healing from failure? Can 2413 your protocol maintain dependability and performance in the face of 2414 unanticipated changes or circumstances? 2416 Explanation: Reliability ensures that a protocol will execute its 2417 function consistently and error resistant as described, and function 2418 without unexpected result. A system that is reliable degenerates 2419 gracefully and will have a documented way to announce degradation. 2420 It also has mechanisms to recover from failure gracefully, and if 2421 applicable, allow for partial healing. It is important here to draw 2422 a distinction between random degradation and malicious degradation. 2423 Many current attacks against TLS, for example, exploit TLS's ability 2424 to gracefully degrade to older cipher suites - from a functional 2425 perspective, this is good. From a security perspective, this can be 2426 very bad. As with confidentiality, the growth of the Internet and 2427 fostering innovation in services depends on users having confidence 2428 and trust [RFC3724] in the network. For reliability it is necessary 2429 that services notify the users if a delivery fails. In the case of 2430 real-time systems in addition to the reliable delivery the protocol 2431 needs to safeguard timeliness. 2433 Example: In the modern IP stack structure, a reliable transport layer 2434 requires an indication that transport processing has successfully 2435 completed, such as given by TCP's ACK message [RFC0793], and not 2436 simply an indication from the IP layer that the packet arrived. 2437 Similarly, an application layer protocol may require an application- 2438 specific acknowledgement that contains, among other things, a status 2439 code indicating the disposition of the request (See [RFC3724]). 2441 Impacts: 2443 - Right to freedom of expression 2445 - Right to security 2447 6.2.15. Confidentiality 2449 Question(s): Does this protocol expose information related to 2450 identifiers or data? If so, does it do so to each other protocol 2451 entity (i.e., recipients, intermediaries, and enablers) [RFC6973]? 2452 What options exist for protocol implementers to choose to limit the 2453 information shared with each entity? What operational controls are 2454 available to limit the information shared with each entity? 2456 What controls or consent mechanisms does the protocol define or 2457 require before personal data or identifiers are shared or exposed via 2458 the protocol? If no such mechanisms or controls are specified, is it 2459 expected that control and consent will be handled outside of the 2460 protocol? 2462 Does the protocol provide ways for initiators to share different 2463 pieces of information with different recipients? If not, are there 2464 mechanisms that exist outside of the protocol to provide initiators 2465 with such control? 2467 Does the protocol provide ways for initiators to limit which 2468 information is shared with intermediaries? If not, are there 2469 mechanisms that exist outside of the protocol to provide users with 2470 such control? Is it expected that users will have relationships that 2471 govern the use of the information (contractual or otherwise) with 2472 those who operate these intermediaries? Does the protocol prefer 2473 encryption over clear text operation? 2475 Does the protocol provide ways for initiators to express individuals' 2476 preferences to recipients or intermediaries with regard to the 2477 collection, use, or disclosure of their personal data? 2479 Explanation: Confidentiality refers to keeping your data secret from 2480 unintended listeners [BCP72]. The growth of the Internet depends on 2481 users having confidence that the network protects their private 2482 information [RFC1984]. 2484 Example: Protocols that do not encrypt their payload make the entire 2485 content of the communication available to the idealized attacker 2486 along their path. Following the advice in [RFC3365], most such 2487 protocols have a secure variant that encrypts the payload for 2488 confidentiality, and these secure variants are seeing ever-wider 2489 deployment. A noteworthy exception is DNS [RFC1035], as DNSSEC 2490 [RFC4033]does not have confidentiality as a requirement. This 2491 implies that, in the absence of changes to the protocol as presently 2492 under development in the IETF's DNS Private Exchange (DPRIVE) working 2493 group, all DNS queries and answers generated by the activities of any 2494 protocol are available to the attacker. When store-and-forward 2495 protocols are used (e.g., SMTP [RFC5321]), intermediaries leave this 2496 data subject to observation by an attacker that has compromised these 2497 intermediaries, unless the data is encrypted end-to-end by the 2498 application-layer protocol or the implementation uses an encrypted 2499 store for this data [RFC7624]. 2501 Impacts: 2503 - Right to privacy 2505 - Right to security 2507 6.2.16. Integrity 2509 Question(s): Does your protocol maintain, assure and/or verify the 2510 accuracy of payload data? Does your protocol maintain and assure the 2511 consistency of data? Does your protocol in any way allow for the 2512 data to be (intentionally or unintentionally) altered? 2514 Explanation: Integrity refers to the maintenance and assurance of the 2515 accuracy and consistency of data to ensure it has not been 2516 (intentionally or unintentionally) altered. 2518 Example: Integrity verification of data is important to prevent 2519 vulnerabilities and attacks, like man-in-the-middle-attacks. These 2520 attacks happen when a third party (often for malicious reasons) 2521 intercepts a communication between two parties, inserting themselves 2522 in the middle changing the content of the data. In practice this 2523 looks as follows: 2525 Alice wants to communicate with Bob. 2526 Alice sends data to Bob. 2527 Corinne intercepts the data sent to Bob. 2528 Corinne reads and alters the message to Bob. 2529 Bob cannot see the data from Alice was altered by Corinne. 2530 Corinne intercepts and alters the communication as it is sent between 2531 Alice and Bob. 2532 Corinne is able to control the communication content. 2534 Impacts: 2536 - Right to freedom of expression 2538 - Right to security 2540 6.2.17. Authenticity 2542 Question(s): Do you have sufficient measures to confirm the truth of 2543 an attribute of a single piece of data or entity? Can the attributes 2544 get garbled along the way (see security)? If relevant have you 2545 implemented IPsec, DNSsec, HTTPS and other Standard Security Best 2546 Practices? 2548 Explanation: Authenticity ensures that data does indeed come from the 2549 source it claims to come from. This is important to prevent attacks 2550 or unauthorized access and use of data. 2552 Example: Authentication of data is important to prevent 2553 vulnerabilities and attacks, like man-in-the-middle-attacks. These 2554 attacks happen when a third party (often for malicious reasons) 2555 intercepts a communication between two parties, inserting themselves 2556 in the middle and posing as both parties. In practice this looks as 2557 follows: 2559 Alice wants to communicate with Bob. 2560 Alice sends data to Bob. 2561 Corinne intercepts the data sent to Bob. 2562 Corinne reads (and potentially alters) the message to Bob. 2563 Bob cannot see the data did not come from Alice but from Corinne. 2565 When there is proper authentication the scenario would be as follows: 2567 Alice wants to communicate with Bob. 2568 Alice sends data to Bob. 2569 Corinne intercepts the data sent to Bob. 2570 Corinne reads and alters the message to Bob. 2571 Bob can see the data did not come from Alice but from Corinne. 2573 Impacts: 2575 - Right to privacy 2577 - Right to freedom of expression 2579 - Right to security 2581 6.2.18. Adaptability 2583 Question(s): Is your protocol written in such a way that is would be 2584 easy for other protocols to be developed on top of it, or to interact 2585 with it? Does your protocol impact permissionless innovation? See 2586 'Connectivity' above. 2588 Explanation: Adaptability is closely interrelated with permissionless 2589 innovation, both maintain the freedom and ability to freely create 2590 and deploy new protocols on top of the communications constructs that 2591 currently exist. It is at the heart of the Internet as we know it, 2592 and to maintain its fundamentally open nature, we need to be mindful 2593 of the impact of protocols on maintaining or reducing permissionless 2594 innovation to ensure the Internet can continue to develop. 2596 Example: WebRTC generates audio and/or video data. In order to 2597 ensure that WebRTC can be used in different locations by different 2598 parties it is important that standard Javascript APIs are developed 2599 to support applications from different voice service providers. 2600 Multiple parties will have similar capabilities, in order to ensure 2601 that all parties can build upon existing standards these need to be 2602 adaptable, and allow for permissionless innovation. 2604 Impacts: 2606 - Right to education 2608 - Freedom of expression 2610 - Freedom of assembly and association 2612 6.2.19. Outcome Transparency 2614 Question(s): Are the effects of your protocol fully and easily 2615 comprehensible, including with respect to unintended consequences of 2616 protocol choices? 2618 Explanation: certain technical choice may have unintended 2619 consequences. 2621 Example: lack of authenticity may lead to lack of integrity and 2622 negative externalities, of which spam is an example. Lack of data 2623 that could be used for billing and accounting can lead to so-called 2624 "free" arrangements which obscure the actual costs and distribution 2625 of the costs, for example the barter arrangements that are commonly 2626 used for Internet interconnection; and the commercial exploitation of 2627 private data for targeted advertising which is the most common 2628 funding model for the so-called "free" services such as search 2629 engines and social networks. 2631 Impacts: - Freedom of expression - Privacy - Freedom of assembly and 2632 association - Access to information 2634 7. Document Status 2636 This document has been developed within the framework of the Human 2637 Rights Protocols Considerations Research Group, based on discussions 2638 on the hrpc mailinglist and during hrpc sessions, where this document 2639 also has been extensively discussed. The draft in its current form 2640 and iteration has received eleven in-depth reviews on list, and 2641 received many comments from inside and outside the IRTF and IETF 2642 community. The authors believe that the issues that have been raised 2643 by the reviewers have been addressed. 2645 8. Acknowledgements 2647 A special thanks to all members of the hrpc RG who contributed to 2648 this draft. The following deserve a special mention: 2650 - Joana Varon for helping draft the first iteration of the 2651 methodology, previous drafts and the direction of the film Net of 2652 Rights and working on the interviews at IETF92 in Dallas. 2654 - Daniel Kahn Gillmor (dkg) for helping with the first iteration of 2655 the glossary as well as a lot of technical guidance, support and 2656 language suggestions. 2658 - Claudio Guarnieri for writing the first iterations of the case 2659 studies on VPN, HTTP, and Peer to Peer. 2661 - Will Scott for writing the first iterations of the case studies on 2662 DNS, IP, XMPP. 2664 - Avri Doria for proposing writing a glossary in the first place, 2665 help with writing the initial proposals and Internet Drafts, her 2666 reviews and contributions to the glossary. 2668 and Stephane Bortzmeyer, John Curran, Barry Shein, Joe Hall, Joss 2669 Wright, Harry Halpin, and Tim Sammut who made a lot of excellent 2670 suggestions, many of which found their way directly into the text. 2671 We want to thank Amelia Andersdotter, Stephen Farrell, Stephane 2672 Bortzemeyer, Shane Kerr, Giovane Moura, James Gannon, Alissa Cooper, 2673 Andrew Sullivan, S. Moonesamy, Roland Bless and Scott Craig for 2674 their reviews and testing the HRPC guidelines in the wild. We would 2675 also like to thank Molly Sauter, Arturo Filasto, Nathalie Marechal, 2676 Eleanor Saitta, Richard Hill and all others who provided input on the 2677 draft or the conceptualization of the idea. Thanks to Edward Snowden 2678 for his comments regarding the impact of protocols on the rights of 2679 users at IETF93. 2681 9. Security Considerations 2683 As this document concerns a research document, there are no security 2684 considerations. 2686 10. IANA Considerations 2688 This document has no actions for IANA. 2690 11. Research Group Information 2692 The discussion list for the IRTF Human Rights Protocol Considerations 2693 proposed working group is located at the e-mail address hrpc@ietf.org 2694 [3]. Information on the group and information on how to subscribe to 2695 the list is at https://www.irtf.org/mailman/listinfo/hrpc 2697 Archives of the list can be found at: https://www.irtf.org/mail- 2698 archive/web/hrpc/current/index.html 2700 12. References 2702 12.1. Informative References 2704 [Abbate] Abbate, J., "Inventing the Internet", MIT Press , 2000, 2705 . 2707 [Abibil] Danchev, D., "Dissecting 'Operation Ababil' - an OSINT 2708 Analysis", 2012, . 2711 [Adrian] Adrian, D., Bhargavan, K., Durumeric, Z., Gaudry, P., 2712 Green, M., Halderman, J., Heninger, N., Springall, D., 2713 Thome, E., Valenta, L., VanderSloot, B., Wustrow, E., 2714 Zanella Beguelin, S., and P. Zimmermann, "Imperfect 2715 Forward Secrecy: How Diffie-Hellman Fails in Practice", 2716 ACM Conference on Computer and Communications Security 2717 2015: 5-17 , 2015. 2719 [Alshalanetal] 2720 Alshalan, A., Pisharody, S., and D. Huang, "A Survey of 2721 Mobile VPN Technologies", 2016, 2722 . 2725 [APIP] Naylor, D., Mukerjee, M., and P. Steenkiste, "Balancing 2726 accountability and privacy in the network", SIGCOMM '14 2727 Proceedings of the 2014 ACM conference on SIGCOMM Pages 2728 75-86 , 2014, . 2731 [Appelbaum] 2732 Appelbaum, J., Gibson, A., Kabish, V., Kampf, L., and L. 2733 Ryge, "NSA targets the privacy-conscious", 2015, 2734 . 2737 [ars] Anderson, N., "P2P researchers - use a blocklist or you 2738 will be tracked... 100% of the time", 2007, 2739 . 2743 [Aryanetall] 2744 Aryan, S., Aryan, H., and J. Alex Halderman, "Internet 2745 Censorship in Iran: A First Look", 2013, 2746 . 2748 [Babbie] Babbie, E., "The Basics of Social Research", Belmont CA 2749 Cengage , 2010. 2751 [bbc-wikileaks] 2752 BBC, "Whistle-blower site taken offline", 2008, 2753 . 2755 [BCP72] IETF, "Guidelines for Writing RFC Text on Security 2756 Considerations", 2003, . 2759 [Benkler] Benkler, Y., "The wealth of Networks - How social 2760 production transforms markets and freedom", New Haven and 2761 London - Yale University Press , 2006, 2762 . 2764 [Berners-Lee] 2765 Berners-Lee, T. and M. Fischetti, "Weaving the Web,", 2766 HarperCollins p 208, 1999. 2768 [BernersLeeHalpin] 2769 Berners-Lee, T. and H. Halpin, "Defend the Web", 2012, 2770 . 2773 [Bhargavan] 2774 Bhargavan, K., Delignat-Lavaud, A., Fournet, C., Pironti, 2775 A., and P. Strub, "Triple Handshakes and Cookie Cutters: 2776 Breaking and Fixing Authentication over TLS", IEEE 2777 Symposium on Security and Privacy 2014: 98-113 , 2014. 2779 [bitmessage] 2780 Bitmessage, "Bitmessage Wiki?", 2014, 2781 . 2783 [Bless] Bless, R. and C. Orwat, "Values and Networks", 2015. 2785 [Broeders] 2786 Broeders, D., "The public core of the Internet", WRR , 2787 2015, 2788 . 2791 [Brown] Brown, I. and M. Ziewitz, "A Prehistory of Internet 2792 Governance", Research Handbook on Governance of the 2793 Internet. Cheltenham, Edward Elgar. , 2013. 2795 [Brownetal] 2796 Brown, I., Clark, D., and D. Trossen, "Should specific 2797 values be embedded in the Internet Architecture?", 2798 Sigcomm , 2010, . 2801 [BrownMarsden] 2802 Brown, I. and C. Marsden, "Regulating code", MIT Press , 2803 2013, . 2805 [caida] Dainotti, A., Squarcella, C., Aben, E., Claffy, K., 2806 Chiesa, M., Russo, M., and A. Pescape, "Analysis of 2807 Country-wide Internet Outages Caused by", 2013, 2808 . 2811 [Cath] Cath, C., "A Case Study of Coding Rights: Should Freedom 2812 of Speech Be Instantiated in the Protocols and Standards 2813 Designed by the Internet Engineering Task Force?", 2015, 2814 . 2817 [CathFloridi] 2818 Cath, C. and L. Floridi, "The Design of the Internet's 2819 Architecture by the Internet Engineering Task Force (IETF) 2820 and Human Rights", February 2017. 2822 [Clark] Clark, D., "The Design Philosophy of the DARPA Internet 2823 Protocols", Proc SIGCOMM 88, ACM CCR Vol 18, Number 4, 2824 August 1988, pp. 106-114. , 1988. 2826 [Clarketal] 2827 Clark, D., Wroclawski, J., Sollins, K., and R. Braden, 2828 "Tussle in cyberspace - defining tomorrow's Internet", ACM 2829 Digital Library , 2005, . 2832 [CoE] Council of Europe, "Applications to ICANN for community- 2833 based new generic top level domains: Opportunities and 2834 challenges from a human rights perspective", 2016, 2835 . 2838 [Collins] Collins, K., "Hacking Team's oppressive regimes customer 2839 list revealed in hack", 2015, 2840 . 2843 [Davidsonetal] 2844 Davidson, A., Morris, J., and R. Courtney, "Strangers in a 2845 strange land", Telecommunications Policy Research 2846 Conference , 2002, 2847 . 2849 [Denardis14] 2850 Denardis, L., "The Global War for Internet Governance", 2851 Yale University Press , 2014, 2852 . 2854 [Denardis15] 2855 Denardis, L., "The Internet Design Tension between 2856 Surveillance and Security", IEEE Annals of the History of 2857 Computing (volume 37-2) , 2015, . 2859 [Denzin] Denzin, N. and Y. Lincoln, "Handbook of Qualitative 2860 Research", Thousand Oaks CA Sage , 2000, 2861 . 2864 [dict] BusinessDictionary.com. WebFinance, Inc., "Reliability 2865 (dictionary entry)", 2016, 2866 . 2869 [Doty] Doty, N., "Automated text analysis of Requests for Comment 2870 (RFCs)", 2014, . 2872 [Douceur] Douceur, J., "The Sybil Attack", 2002, 2873 . 2876 [Dutton] Dutton, W., "Freedom of Connection, Freedom of Expression: 2877 the Changing legal and regulatory Ecology Shaping the 2878 Internet.", 2011, . 2881 [Farrow] Farrow, R., "Source Address Spoofing", 2016, 2882 . 2884 [FIArch] "Future Internet Design Principles", January 2012, 2885 . 2888 [FOC] Ministers of the Freedom Online Coalition, "The Tallinn 2889 Agenda - Recommendations for Freedom Online", 2014, 2890 . 2894 [FRAMEWORK] 2895 ISO/IEC, ., "Information technology - Framework for 2896 internationalization, prepared by ISO/IEC JTC 1/SC 22/WG 2897 20 ISO/IEC TR 11017", 1997. 2899 [Franklin] 2900 Franklin, U., "The Real World of Technology", 1999, 2901 . 2904 [freenet1] 2905 Freenet, "What is Freenet?", n.d., 2906 . 2908 [freenet2] 2909 Ian Clarke, ., "The Philosphy behind Freenet?", n.d., 2910 . 2912 [geekfeminism] 2913 Geek Feminism Wiki, "Pseudonymity", 2015, 2914 . 2916 [Geertz] Clifford, G., "Kinship in Bali", Chicago University of 2917 Chicago Press. , 1975, 2918 . 2921 [Googlepatent] 2922 Google, ., "Method and device for network traffic 2923 manipulation", 2012, . 2926 [greatfirewall] 2927 Anonymous, ., "Towards a Comprehensive Picture of the 2928 Great Firewall's DNS Censorship", 2014, 2929 . 2932 [GreenMovement] 2933 Villeneuve, N., "Iran DDoS", 2009, 2934 . 2936 [Greenwald] 2937 Greenwald, G., "XKeyscore: NSA tool collects 'nearly 2938 everything a user does on the internet'", 2013, 2939 . 2942 [Haagsma] Haagsma, L., "Deep dive into QUANTUM INSERT", 2015, 2943 . 2946 [hall] Hall, J., Aaron, M., and B. Jones, "A Survey of Worldwide 2947 Censorship Techniques", 2015, 2948 . 2951 [Hill2014] 2952 Hill, R., "Partial Catalog of Human Rights Related to ICT 2953 Activities", 2014, 2954 . 2956 [Hornet] Chen, C., Asoni, D., Barrera, D., Danezis, G., and A. 2957 Perrig, "HORNET: High-speed Onion Routing at the Network 2958 Layer", CCS '15 Proceedings of the 22nd ACM SIGSAC 2959 Conference on Computer and Communications Security Pages 2960 1441-1454 , 2015, . 2963 [HRC2012] United Nations Human Rights Council, "UN General Assembly 2964 Resolution "The right to privacy in the digital age" 2965 (A/C.3/68/L.45)", 2011, 2966 . 2968 [HTML5] W3C, "HTML5", 2014, . 2970 [ICCPR] United Nations General Assembly, "International Covenant 2971 on Civil and Political Rights", 1976, 2972 . 2975 [ICESCR] United Nations General Assembly, "International Covenant 2976 on Economic, Social and Cultural Rights", 1966, 2977 . 2980 [Insinuator] 2981 Schiess, N., "Vulnerabilities & attack vectors of VPNs (Pt 2982 1)", 2013, . 2985 [IRP] Internet Rights and Principles Dynamic Coalition, "10 2986 Internet Rights & Principles", 2014, 2987 . 2991 [Jabri] Jabri, V., "Discourses on Violence - conflict analysis 2992 reconsidered", Manchester University Press , 1996. 2994 [Kaye] Kaye, D., "Report of the Special Rapporteur on the 2995 promotion and protection of the right to freedom of 2996 opinion and expression", 2016, 2997 . 3000 [King] King, C., "Power, Social Violence and Civil Wars", 3001 Washington D.C. United States Institute of Peace Press , 3002 2007. 3004 [Lessig] Lessig, L., "Code - And Other Laws of Cyberspace, Version 3005 2.0.", New York Basic Books , 2006, . 3007 [Marcak] Marcak, B., Weaver, N., Dalek, J., Ensafi, R., Fifield, 3008 D., McKune, S., Rey, A., Scott-Railton, J., Deibert, R., 3009 and V. Paxson, "China's Great Fire Cannon", 2015, 3010 . 3012 [Marquis-Boire] 3013 Marquis-Boire, M., "Schrodinger's Cat Video and the Death 3014 of Clear-Text", 2014, . 3017 [Meyer] Meyer, J., "Defining and Evaluating Resilience: A 3018 Performability Perspective, presentation at International 3019 Workshop on Performability Modeling of Computer and 3020 Communication Systems.", 2009. 3022 [Mueller] Mueller, M., "Networks and States", MIT Press , 2010, 3023 . 3025 [Musiani] Musiani, F., "Giants, Dwarfs and Decentralized 3026 Alternatives to Internet-based Services - An Issue of 3027 Internet Governance", Westminister Papers in Communication 3028 and Culture , 2015, . 3030 [namecoin] 3031 Namecoin, "Namecoin - Decentralized secure names", 2015, 3032 . 3034 [natusage] 3035 Maier, G., Schneider, F., and A. Feldmann, "NAT usage in 3036 Residential Broadband networks", 2011, 3037 . 3040 [NETmundial] 3041 NETmundial, "NETmundial Multistakeholder Statement", 2014, 3042 . 3045 [newegg] Mullin, J., "Newegg on trial: Mystery company TQP rewrites 3046 the history of encryption", 2013, . 3050 [notewell] 3051 IETF, "Note Well", 2015, . 3054 [patentpolicy] 3055 W3C, "W3C Patent Policy", 2004, 3056 . 3058 [Penney] Penney, J., "Chilling Effects: Online Surveillance and 3059 Wikipedia Use", 2016, . 3062 [Peterson] 3063 Peterson, A., Gellman, B., and A. Soltani, "Yahoo to make 3064 SSL encryption the default for Webmail users. Finally.", 3065 2013, . 3068 [PETS2015VPN] 3069 Pera, V., Barbera, M., Tyson, G., Haddadi, H., and A. Mei, 3070 "A Glance through the VPN Looking Glass", 2015, 3071 . 3074 [pidgin] js, . and Pidgin Developers, "-XMPP- Invisible mode 3075 violating standard", July 2015, 3076 . 3078 [Pouwelse] 3079 Pouwelse, Ed, J., "Media without censorship", 2012, 3080 . 3083 [Rachovitsa] 3084 Rachovitsa, A., "Engineering 'Privacy by Design' in the 3085 Internet Protocols - Understanding Online Privacy both as 3086 a Technical and a Human Rights Issue in the Face of 3087 Pervasive Monitoring", International Journal of Law and 3088 Information Technology , 2015, . 3091 [RFC0226] Karp, P., "Standardization of host mnemonics", RFC 226, 3092 DOI 10.17487/RFC0226, September 1971, 3093 . 3095 [RFC0760] Postel, J., "DoD standard Internet Protocol", RFC 760, 3096 DOI 10.17487/RFC0760, January 1980, 3097 . 3099 [RFC0791] Postel, J., "Internet Protocol", STD 5, RFC 791, 3100 DOI 10.17487/RFC0791, September 1981, 3101 . 3103 [RFC0793] Postel, J., "Transmission Control Protocol", STD 7, 3104 RFC 793, DOI 10.17487/RFC0793, September 1981, 3105 . 3107 [RFC0894] Hornig, C., "A Standard for the Transmission of IP 3108 Datagrams over Ethernet Networks", STD 41, RFC 894, 3109 DOI 10.17487/RFC0894, April 1984, 3110 . 3112 [RFC1035] Mockapetris, P., "Domain names - implementation and 3113 specification", STD 13, RFC 1035, DOI 10.17487/RFC1035, 3114 November 1987, . 3116 [RFC1122] Braden, R., Ed., "Requirements for Internet Hosts - 3117 Communication Layers", STD 3, RFC 1122, 3118 DOI 10.17487/RFC1122, October 1989, 3119 . 3121 [RFC1958] Carpenter, B., Ed., "Architectural Principles of the 3122 Internet", RFC 1958, DOI 10.17487/RFC1958, June 1996, 3123 . 3125 [RFC1984] IAB and IESG, "IAB and IESG Statement on Cryptographic 3126 Technology and the Internet", BCP 200, RFC 1984, 3127 DOI 10.17487/RFC1984, August 1996, 3128 . 3130 [RFC2026] Bradner, S., "The Internet Standards Process -- Revision 3131 3", BCP 9, RFC 2026, DOI 10.17487/RFC2026, October 1996, 3132 . 3134 [RFC2277] Alvestrand, H., "IETF Policy on Character Sets and 3135 Languages", BCP 18, RFC 2277, DOI 10.17487/RFC2277, 3136 January 1998, . 3138 [RFC2460] Deering, S. and R. Hinden, "Internet Protocol, Version 6 3139 (IPv6) Specification", RFC 2460, DOI 10.17487/RFC2460, 3140 December 1998, . 3142 [RFC2775] Carpenter, B., "Internet Transparency", RFC 2775, 3143 DOI 10.17487/RFC2775, February 2000, 3144 . 3146 [RFC3022] Srisuresh, P. and K. Egevang, "Traditional IP Network 3147 Address Translator (Traditional NAT)", RFC 3022, 3148 DOI 10.17487/RFC3022, January 2001, 3149 . 3151 [RFC3365] Schiller, J., "Strong Security Requirements for Internet 3152 Engineering Task Force Standard Protocols", BCP 61, 3153 RFC 3365, DOI 10.17487/RFC3365, August 2002, 3154 . 3156 [RFC3536] Hoffman, P., "Terminology Used in Internationalization in 3157 the IETF", RFC 3536, DOI 10.17487/RFC3536, May 2003, 3158 . 3160 [RFC3724] Kempf, J., Ed., Austein, R., Ed., and IAB, "The Rise of 3161 the Middle and the Future of End-to-End: Reflections on 3162 the Evolution of the Internet Architecture", RFC 3724, 3163 DOI 10.17487/RFC3724, March 2004, 3164 . 3166 [RFC3935] Alvestrand, H., "A Mission Statement for the IETF", 3167 BCP 95, RFC 3935, DOI 10.17487/RFC3935, October 2004, 3168 . 3170 [RFC3979] Bradner, S., Ed., "Intellectual Property Rights in IETF 3171 Technology", BCP 79, RFC 3979, DOI 10.17487/RFC3979, March 3172 2005, . 3174 [RFC4033] Arends, R., Austein, R., Larson, M., Massey, D., and S. 3175 Rose, "DNS Security Introduction and Requirements", 3176 RFC 4033, DOI 10.17487/RFC4033, March 2005, 3177 . 3179 [RFC4084] Klensin, J., "Terminology for Describing Internet 3180 Connectivity", BCP 104, RFC 4084, DOI 10.17487/RFC4084, 3181 May 2005, . 3183 [RFC4101] Rescorla, E. and IAB, "Writing Protocol Models", RFC 4101, 3184 DOI 10.17487/RFC4101, June 2005, 3185 . 3187 [RFC4646] Phillips, A. and M. Davis, "Tags for Identifying 3188 Languages", RFC 4646, DOI 10.17487/RFC4646, September 3189 2006, . 3191 [RFC4941] Narten, T., Draves, R., and S. Krishnan, "Privacy 3192 Extensions for Stateless Address Autoconfiguration in 3193 IPv6", RFC 4941, DOI 10.17487/RFC4941, September 2007, 3194 . 3196 [RFC4949] Shirey, R., "Internet Security Glossary, Version 2", 3197 FYI 36, RFC 4949, DOI 10.17487/RFC4949, August 2007, 3198 . 3200 [RFC5246] Dierks, T. and E. Rescorla, "The Transport Layer Security 3201 (TLS) Protocol Version 1.2", RFC 5246, 3202 DOI 10.17487/RFC5246, August 2008, 3203 . 3205 [RFC5321] Klensin, J., "Simple Mail Transfer Protocol", RFC 5321, 3206 DOI 10.17487/RFC5321, October 2008, 3207 . 3209 [RFC5694] Camarillo, G., Ed. and IAB, "Peer-to-Peer (P2P) 3210 Architecture: Definition, Taxonomies, Examples, and 3211 Applicability", RFC 5694, DOI 10.17487/RFC5694, November 3212 2009, . 3214 [RFC5944] Perkins, C., Ed., "IP Mobility Support for IPv4, Revised", 3215 RFC 5944, DOI 10.17487/RFC5944, November 2010, 3216 . 3218 [RFC6101] Freier, A., Karlton, P., and P. Kocher, "The Secure 3219 Sockets Layer (SSL) Protocol Version 3.0", RFC 6101, 3220 DOI 10.17487/RFC6101, August 2011, 3221 . 3223 [RFC6108] Chung, C., Kasyanov, A., Livingood, J., Mody, N., and B. 3224 Van Lieu, "Comcast's Web Notification System Design", 3225 RFC 6108, DOI 10.17487/RFC6108, February 2011, 3226 . 3228 [RFC6120] Saint-Andre, P., "Extensible Messaging and Presence 3229 Protocol (XMPP): Core", RFC 6120, DOI 10.17487/RFC6120, 3230 March 2011, . 3232 [RFC6365] Hoffman, P. and J. Klensin, "Terminology Used in 3233 Internationalization in the IETF", BCP 166, RFC 6365, 3234 DOI 10.17487/RFC6365, September 2011, 3235 . 3237 [RFC6698] Hoffman, P. and J. Schlyter, "The DNS-Based Authentication 3238 of Named Entities (DANE) Transport Layer Security (TLS) 3239 Protocol: TLSA", RFC 6698, DOI 10.17487/RFC6698, August 3240 2012, . 3242 [RFC6701] Farrel, A. and P. Resnick, "Sanctions Available for 3243 Application to Violators of IETF IPR Policy", RFC 6701, 3244 DOI 10.17487/RFC6701, August 2012, 3245 . 3247 [RFC6797] Hodges, J., Jackson, C., and A. Barth, "HTTP Strict 3248 Transport Security (HSTS)", RFC 6797, 3249 DOI 10.17487/RFC6797, November 2012, 3250 . 3252 [RFC6973] Cooper, A., Tschofenig, H., Aboba, B., Peterson, J., 3253 Morris, J., Hansen, M., and R. Smith, "Privacy 3254 Considerations for Internet Protocols", RFC 6973, 3255 DOI 10.17487/RFC6973, July 2013, 3256 . 3258 [RFC7258] Farrell, S. and H. Tschofenig, "Pervasive Monitoring Is an 3259 Attack", BCP 188, RFC 7258, DOI 10.17487/RFC7258, May 3260 2014, . 3262 [RFC7469] Evans, C., Palmer, C., and R. Sleevi, "Public Key Pinning 3263 Extension for HTTP", RFC 7469, DOI 10.17487/RFC7469, April 3264 2015, . 3266 [RFC7540] Belshe, M., Peon, R., and M. Thomson, Ed., "Hypertext 3267 Transfer Protocol Version 2 (HTTP/2)", RFC 7540, 3268 DOI 10.17487/RFC7540, May 2015, 3269 . 3271 [RFC7574] Bakker, A., Petrocco, R., and V. Grishchenko, "Peer-to- 3272 Peer Streaming Peer Protocol (PPSPP)", RFC 7574, 3273 DOI 10.17487/RFC7574, July 2015, 3274 . 3276 [RFC7624] Barnes, R., Schneier, B., Jennings, C., Hardie, T., 3277 Trammell, B., Huitema, C., and D. Borkmann, 3278 "Confidentiality in the Face of Pervasive Surveillance: A 3279 Threat Model and Problem Statement", RFC 7624, 3280 DOI 10.17487/RFC7624, August 2015, 3281 . 3283 [RFC7626] Bortzmeyer, S., "DNS Privacy Considerations", RFC 7626, 3284 DOI 10.17487/RFC7626, August 2015, 3285 . 3287 [RFC7725] Bray, T., "An HTTP Status Code to Report Legal Obstacles", 3288 RFC 7725, DOI 10.17487/RFC7725, February 2016, 3289 . 3291 [RFC7754] Barnes, R., Cooper, A., Kolkman, O., Thaler, D., and E. 3292 Nordmark, "Technical Considerations for Internet Service 3293 Blocking and Filtering", RFC 7754, DOI 10.17487/RFC7754, 3294 March 2016, . 3296 [RFC7858] Hu, Z., Zhu, L., Heidemann, J., Mankin, A., Wessels, D., 3297 and P. Hoffman, "Specification for DNS over Transport 3298 Layer Security (TLS)", RFC 7858, DOI 10.17487/RFC7858, May 3299 2016, . 3301 [Richie] Richie, J. and J. Lewis, "Qualitative Research Practice - 3302 A Guide for Social Science Students and Researchers", 3303 London Sage , 2003, . 3307 [Rideout] Rideout, A., "Making security easier", 2008, 3308 . 3311 [RSF] RSF, "Syria using 34 Blue Coat Servers to spy on Internet 3312 users", 2013, . 3315 [Saltzer] Saltzer, J., Reed, D., and D. Clark, "End-to-End Arguments 3316 in System Design", ACM TOCS, Vol 2, Number 4, November 3317 1984, pp 277-288. , 1984. 3319 [Sandvine] 3320 Sandvine, "Sandvine: Over 70% Of North American Traffic Is 3321 Now Streaming Video And Audio", 2015, 3322 . 3326 [Schillace] 3327 Schillace, S., "Default https access for Gmail", 2010, 3328 . 3331 [Schneier] 3332 Schneier, B., "Attacking Tor - how the NSA targets users' 3333 online anonymity", 2013, 3334 . 3337 [Schroeder] 3338 Schroeder, I. and B. Schmidt, "Introduction - Violent 3339 Imaginaries and Violent Practice", London and New York 3340 Routledge , 2001, . 3344 [spiegel] SPIEGEL, "Prying Eyes - Inside the NSA's War on Internet 3345 Security", 2014, 3346 . 3349 [sslstrip] 3350 Marlinspike, M., "Software >> sslstrip", 2011, 3351 . 3353 [techyum] Violet, ., "Official - vb.ly Link Shortener Seized by 3354 Libyan Government", 2010, . 3358 [torproject] 3359 The Tor Project, ., "Tor Project - Anonymity Online", 3360 2007, . 3362 [torrentfreak1] 3363 Van der Sar, E., "Proposal for research on human rights 3364 protocol considerations", 2015, . 3368 [torrentfreak2] 3369 Andy, ., "LAWYERS SENT 109,000 PIRACY THREATS IN GERMANY 3370 DURING 2013", 2014, . 3374 [tribler] Delft University of Technology, Department EWI/PDS/ 3375 Tribler, "About Tribler", 2013, . 3378 [UDHR] United Nations General Assembly, "The Universal 3379 Declaration of Human Rights", 1948, 3380 . 3382 [UNGA2013] 3383 United Nations General Assembly, "UN General Assembly 3384 Resolution "The right to privacy in the digital age" 3385 (A/C.3/68/L.45)", 2013, 3386 . 3388 [UNHRC2016] 3389 United Nations Human Rights Council, "UN Human Rights 3390 Council Resolution "The promotion, protection and 3391 enjoyment of human rights on the Internet" (A/HRC/32/ 3392 L.20)", 2016, . 3396 [ververis] 3397 Vasilis, V., Kargiotakis, G., Filasto, A., Fabian, B., and 3398 A. Alexandros, "Understanding Internet Censorship Policy - 3399 The Case of Greece", 2015, 3400 . 3403 [W3CAccessibility] 3404 W3C, "Accessibility", 2015, 3405 . 3407 [W3Ci18nDef] 3408 W3C, "Localization vs. Internationalization", 2010, 3409 . 3411 [wikileaks] 3412 Sladek, T. and E. Broese, "Market Survey : Detection & 3413 Filtering Solutions to Identify File Transfer of Copyright 3414 Protected Content for Warner Bros. and movielabs", 2011, 3415 . 3418 [WP-Tempora] 3419 Wikipedia, "Tempora", 2016, 3420 . 3422 [WSJ] Sonne, P. and M. Coker, "Firms Aided Libyan Spies", 2011, 3423 . 3426 [WynsbergheMoura] 3427 Wynsberghe, A. and G. Moura, "The concept of embedded 3428 values and the example of internet security", 2013, 3429 . 3431 [xmppmanifesto] 3432 Saint-Andre, P. and . XMPP Operators, "A Public Statement 3433 Regarding Ubiquitous Encryption on the XMPP Network", 3434 2014, 3435 . 3438 [Zittrain] 3439 Zittrain, J., "The Future of the Internet - And How to 3440 Stop It", Yale University Press , 2008, 3441 . 3444 12.2. URIs 3446 [1] mailto:node@domain/home 3448 [2] mailto:node@domain/work 3450 [3] mailto:hrpc@ietf.org 3452 Authors' Addresses 3454 Niels ten Oever 3455 ARTICLE 19 3457 EMail: niels@article19.org 3458 Corinne Cath 3459 Oxford Internet Institute 3461 EMail: corinnecath@gmail.com