idnits 2.17.1 draft-mm-wg-effect-encrypt-24.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 : ---------------------------------------------------------------------------- No issues found here. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (March 2, 2018) is 2219 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- == Missing Reference: 'Ben17a' is mentioned on line 1615, but not defined == Missing Reference: 'Ben17b' is mentioned on line 1615, but not defined == Missing Reference: 'Res17a' is mentioned on line 1615, but not defined == Missing Reference: 'Res17b' is mentioned on line 1615, but not defined == Unused Reference: 'TLS100Proceedings' is defined on line 2412, but no explicit reference was found in the text == Outdated reference: A later version (-25) exists of draft-ietf-dots-use-cases-09 == Outdated reference: A later version (-09) exists of draft-ietf-tls-sni-encryption-00 == Outdated reference: A later version (-06) exists of draft-mglt-nvo3-geneve-security-requirements-03 -- Obsolete informational reference (is this intentional?): RFC 3315 (Obsoleted by RFC 8415) -- Obsolete informational reference (is this intentional?): RFC 7230 (Obsoleted by RFC 9110, RFC 9112) -- Obsolete informational reference (is this intentional?): RFC 7234 (Obsoleted by RFC 9111) -- Obsolete informational reference (is this intentional?): RFC 7525 (Obsoleted by RFC 9325) -- Obsolete informational reference (is this intentional?): RFC 7540 (Obsoleted by RFC 9113) Summary: 0 errors (**), 0 flaws (~~), 9 warnings (==), 6 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group K. Moriarty, Ed. 3 Internet-Draft Dell EMC 4 Intended status: Informational A. Morton, Ed. 5 Expires: September 3, 2018 AT&T Labs 6 March 2, 2018 8 Effects of Pervasive Encryption on Operators 9 draft-mm-wg-effect-encrypt-24 11 Abstract 13 Pervasive Monitoring (PM) attacks on the privacy of Internet users 14 are of serious concern to both the user and the operator communities. 15 RFC7258 discussed the critical need to protect users' privacy when 16 developing IETF specifications and also recognized making networks 17 unmanageable to mitigate PM is not an acceptable outcome; an 18 appropriate balance is needed. This document discusses current 19 security and network operations and management practices that may be 20 impacted by the shift to increased use of encryption to help guide 21 protocol development in support of manageable and secure networks. 23 Status of This Memo 25 This Internet-Draft is submitted in full conformance with the 26 provisions of BCP 78 and BCP 79. 28 Internet-Drafts are working documents of the Internet Engineering 29 Task Force (IETF). Note that other groups may also distribute 30 working documents as Internet-Drafts. The list of current Internet- 31 Drafts is at https://datatracker.ietf.org/drafts/current/. 33 Internet-Drafts are draft documents valid for a maximum of six months 34 and may be updated, replaced, or obsoleted by other documents at any 35 time. It is inappropriate to use Internet-Drafts as reference 36 material or to cite them other than as "work in progress." 38 This Internet-Draft will expire on September 3, 2018. 40 Copyright Notice 42 Copyright (c) 2018 IETF Trust and the persons identified as the 43 document authors. All rights reserved. 45 This document is subject to BCP 78 and the IETF Trust's Legal 46 Provisions Relating to IETF Documents 47 (https://trustee.ietf.org/license-info) in effect on the date of 48 publication of this document. Please review these documents 49 carefully, as they describe your rights and restrictions with respect 50 to this document. Code Components extracted from this document must 51 include Simplified BSD License text as described in Section 4.e of 52 the Trust Legal Provisions and are provided without warranty as 53 described in the Simplified BSD License. 55 Table of Contents 57 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 58 1.1. Additional Background on Encryption Changes . . . . . . . 4 59 1.2. Examples of Attempts to Preserve Functions . . . . . . . 6 60 2. Network Service Provider Monitoring . . . . . . . . . . . . . 7 61 2.1. Passive Monitoring . . . . . . . . . . . . . . . . . . . 8 62 2.1.1. Traffic Surveys . . . . . . . . . . . . . . . . . . . 8 63 2.1.2. Troubleshooting . . . . . . . . . . . . . . . . . . . 8 64 2.1.3. Traffic Analysis Fingerprinting . . . . . . . . . . . 11 65 2.2. Traffic Optimization and Management . . . . . . . . . . . 12 66 2.2.1. Load Balancers . . . . . . . . . . . . . . . . . . . 12 67 2.2.2. Differential Treatment based on Deep Packet 68 Inspection (DPI) . . . . . . . . . . . . . . . . . . 14 69 2.2.3. Network Congestion Management . . . . . . . . . . . . 15 70 2.2.4. Performance-enhancing Proxies . . . . . . . . . . . . 15 71 2.2.5. Caching and Content Replication Near the Network Edge 16 72 2.2.6. Content Compression . . . . . . . . . . . . . . . . . 17 73 2.2.7. Service Function Chaining . . . . . . . . . . . . . . 18 74 2.3. Content Filtering, Network Access, and Accounting . . . . 18 75 2.3.1. Content Filtering . . . . . . . . . . . . . . . . . . 19 76 2.3.2. Network Access and Data Usage . . . . . . . . . . . . 20 77 2.3.3. Application Layer Gateways . . . . . . . . . . . . . 21 78 2.3.4. HTTP Header Insertion . . . . . . . . . . . . . . . . 22 79 3. Encryption in Hosting and Application SP Environments . . . . 22 80 3.1. Management Access Security . . . . . . . . . . . . . . . 22 81 3.1.1. Customer Access Monitoring . . . . . . . . . . . . . 23 82 3.1.2. SP Content Monitoring of Applications . . . . . . . . 24 83 3.2. Hosted Applications . . . . . . . . . . . . . . . . . . . 26 84 3.2.1. Monitoring Managed Applications . . . . . . . . . . . 26 85 3.2.2. Mail Service Providers . . . . . . . . . . . . . . . 27 86 3.3. Data Storage . . . . . . . . . . . . . . . . . . . . . . 27 87 3.3.1. Object-level Encryption . . . . . . . . . . . . . . . 27 88 3.3.2. Disk Encryption, Data at Rest . . . . . . . . . . . . 28 89 3.3.3. Cross Data Center Replication Services . . . . . . . 29 90 4. Encryption for Enterprises . . . . . . . . . . . . . . . . . 29 91 4.1. Monitoring Practices of the Enterprise . . . . . . . . . 30 92 4.1.1. Security Monitoring in the Enterprise . . . . . . . . 30 93 4.1.2. Application Performance Monitoring in the Enterprise 31 94 4.1.3. Enterprise Network Diagnostics and Troubleshooting . 32 95 4.2. Techniques for Monitoring Internet Session Traffic . . . 34 96 5. Security Monitoring for Specific Attack Types . . . . . . . . 36 97 5.1. Mail Abuse and spam . . . . . . . . . . . . . . . . . . . 36 98 5.2. Denial of Service . . . . . . . . . . . . . . . . . . . . 37 99 5.3. Phishing . . . . . . . . . . . . . . . . . . . . . . . . 37 100 5.4. Botnets . . . . . . . . . . . . . . . . . . . . . . . . . 38 101 5.5. Malware . . . . . . . . . . . . . . . . . . . . . . . . . 38 102 5.6. Spoofed Source IP Address Protection . . . . . . . . . . 39 103 5.7. Further work . . . . . . . . . . . . . . . . . . . . . . 39 104 6. Application-based Flow Information Visible to a Network . . . 39 105 6.1. IP Flow Information Export . . . . . . . . . . . . . . . 39 106 6.2. TLS Server Name Indication . . . . . . . . . . . . . . . 40 107 6.3. Application Layer Protocol Negotiation (ALPN) . . . . . . 41 108 6.4. Content Length, BitRate and Pacing . . . . . . . . . . . 41 109 7. Effect of Encryption on Mobile Network Evolution . . . . . . 41 110 8. Response to Increased Encryption and Looking Forward . . . . 42 111 9. Security Considerations . . . . . . . . . . . . . . . . . . . 43 112 10. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 43 113 11. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 43 114 12. Informative References . . . . . . . . . . . . . . . . . . . 43 115 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 52 117 1. Introduction 119 In response to pervasive monitoring revelations and the IETF 120 consensus that Pervasive Monitoring is an Attack [RFC7258], efforts 121 are underway to increase encryption of Internet traffic. Pervasive 122 Monitoring (PM) is of serious concern to users, operators, and 123 application providers. RFC7258 discussed the critical need to 124 protect users' privacy when developing IETF specifications and also 125 recognized that making networks unmanageable to mitigate PM is not an 126 acceptable outcome, but rather that an appropriate balance would 127 emerge over time. 129 This document describes practices currently used by network operators 130 to manage, operate, and secure their networks and how those practices 131 may be impacted by a shift to increased use of encryption. It 132 provides network operators' perspectives about the motivations and 133 objectives of those practices as well as effects anticipated by 134 operators as use of encryption increases. It is a summary of 135 concerns of the operational community as they transition to managing 136 networks with less visibility. The document does not endorse the use 137 of the practices described herein. Nor does it aim to provide a 138 comprehensive treatment of the effects of current practices, some of 139 which have been considered controversial from a technical or business 140 perspective or contradictory to previous IETF statements (e.g., 141 [RFC1958], [RFC1984], [RFC2804]). The informational documents 142 consider the end to end (e2e) architectural principle to be a guiding 143 principle for the development of Internet protocols [RFC2775] 144 [RFC3724] [RFC7754]. 146 This document aims to help IETF participants understand network 147 operators' perspectives about the impact of pervasive encryption, 148 both opportunistic and strong end-to-end encryption, on operational 149 practices. The goal is to help inform future protocol development to 150 ensure that operational impact is part of the conversation. Perhaps, 151 new methods could be developed to accomplish some of the goals of 152 current practices despite changes in the extent to which cleartext 153 will be available to network operators (including methods that rely 154 on network endpoints where applicable). Discussion of current 155 practices and the potential future changes is provided as a 156 prerequisite to potential future cross-industry and cross-layer work 157 to support the ongoing evolution towards a functional Internet with 158 pervasive encryption. 160 Traditional network management, planning, security operations, and 161 performance optimization have been developed in an Internet where a 162 large majority of data traffic flows without encryption. While 163 unencrypted traffic has made information that aids operations and 164 troubleshooting at all layers accessible, it has also made pervasive 165 monitoring by unseen parties possible. With broad support and 166 increased awareness of the need to consider privacy in all aspects 167 across the Internet, it is important to catalog existing management, 168 operational, and security practices that have depended upon the 169 availability of cleartext to function and to explore if critical 170 operational practices can be met by less invasive means. 172 This document refers to several different forms of service providers, 173 distinguished with adjectives. For example, network service 174 providers (or network operators) provide IP-packet transport 175 primarily, though they may bundle other services with packet 176 transport. Alternatively, application service providers primarily 177 offer systems that participate as an end-point in communications with 178 the application user, and hosting service providers lease computing, 179 storage, and communications systems in datacenters. In practice, 180 many companies perform two or more service provider roles, but may be 181 historically associated with one. 183 This document includes a sampling of current practices and does not 184 attempt to describe every nuance. Some sections cover technologies 185 used over a broad spectrum of devices and use cases. 187 1.1. Additional Background on Encryption Changes 189 Pervasive encryption in this document refers to all types of session 190 encryption including Transport Layer Security (TLS), IP security 191 (IPsec), TCPcrypt [TCPcrypt], QUIC [QUIC] and others that are 192 increasing in deployment usage. It is well understood that session 193 encryption helps to prevent both passive and active attacks on 194 transport protocols; more on pervasive monitoring can be found in 195 Confidentiality in the Face of Pervasive Surveillance: A Threat Model 196 and Problem Statement [RFC7624]. Active attacks have long been a 197 motivation for increased encryption, and preventing pervasive 198 monitoring became a focus just a few years ago. As such, the 199 Internet Architecture Board (IAB) released a statement advocating for 200 increased use of encryption in November 2014. Perspectives on 201 encryption paradigms have shifted over time to incorporporate ease of 202 deployment as a high priority, and balance that against providing the 203 maximum possible level of security regardless of deployment 204 considerations. 206 One such shift is documented in "Opportunistic Security" (OS) 207 [RFC7435], which suggests that when use of authenticated encryption 208 is not possible, cleartext sessions should be upgraded to 209 unauthenticated session encryption, rather than no encryption. OS 210 encourages upgrading from cleartext, but cannot require or guarantee 211 such upgrades. Once OS is used, it allows for an evolution to 212 authenticated encryption. These efforts are necessary to improve end 213 user's expectation of privacy, making pervasive monitoring cost 214 prohibitive. With OS in use, active attacks are still possible on 215 unauthenticated sessions. OS has been implemented as NULL 216 Authentication with IPsec [RFC7619] and there are a number of 217 infrastructure use cases such as server to server encryption where 218 this mode is deployed. While OS is helpful in reducing pervasive 219 monitoring by increasing the cost to monitor, it is recognized that 220 risk profiles for some applications require authenticated and secure 221 session encryption as well to prevent active attacks. IPsec, and 222 other session encryption protocols, with authentication has many 223 useful applications and usage has increased for infrastructure 224 applications such as for virtual private networks between data 225 centers. OS as well as other protocol developments, like the 226 Automated Certificate Management Environment (ACME), have increased 227 the usage of session encryption on the Internet. 229 Risk profiles vary and so do the types of session encryption 230 deployed. To understand the scope of changes in visibility a few 231 examples are highlighted. Work continues to improve the 232 implementation, development and configuration of TLS and DTLS 233 sessions to prevent active attacks used to monitor or intercept 234 session data. The changes from TLS 1.2 to 1.3 enhance the security 235 of TLS, while hiding more of the session negotiation and providing 236 less visibility on the wire. The Using TLS in Applications (UTA) 237 working group has been publishing documentation to improve the 238 security of TLS and DTLS sessions. They have documented the known 239 attack vectors in [RFC7457] and have documented Best Practices for 240 TLS and DTLS in [RFC7525] and have other documents in the queue. The 241 recommendations from these documents were built upon for TLS 1.3 to 242 provide a more inherently secure end-to-end protocol. 244 In addition to encrypted web site access (HTTP over TLS), there are 245 other well-deployed application level transport encryption efforts 246 such as mail transfer agent (MTA)-to-MTA session encryption transport 247 for email (SMTP over TLS) and gateway-to-gateway for instant 248 messaging (Extensible Messaging and Presence Protocol (XMPP) over 249 TLS). Although this does provide protection from transport layer 250 attacks, the servers could be a point of vulnerability if user-to- 251 user encryption is not provided for these messaging protocols. User- 252 to-user content encryption schemes, such as S/MIME and PGP for email 253 and Off-the-Record (OTR) encryption for XMPP are used by those 254 interested to protect their data as it crosses intermediary servers, 255 preventing transport layer attacks by providing an end-to-end 256 solution. User-to-user schemes are under review and additional 257 options will emerge to ease the configuration requirements, making 258 this type of option more accessible to non-technical users interested 259 in protecting their privacy. 261 Increased use of encryption, either opportunistic or authenticated, 262 at the transport, network or application layer, impacts how networks 263 are operated, managed, and secured. In some cases, new methods to 264 operate, manage, and secure networks will evolve in response. In 265 other cases, currently available capabilities for monitoring or 266 troubleshooting networks could become unavailable. This document 267 lists a collection of functions currently employed by network 268 operators that may be impacted by the shift to increased use of 269 encryption. This draft does not attempt to specify responses or 270 solutions to these impacts, but rather documents the current state. 272 1.2. Examples of Attempts to Preserve Functions 274 Following the Snowden [Snowden] revelations, application service 275 providers responded by encrypting traffic between their data centers 276 (IPsec) to prevent passive monitoring from taking place unbeknownst 277 to them (Yahoo, Google, etc.). Infrastructure traffic carried over 278 the public Internet has been encrypted for some time, this change for 279 universal encryption was specific to their private backbones. Large 280 mail service providers also began to encrypt session transport (TLS) 281 to hosted mail services. This and other increases in the use of 282 encryption had the immediate effect of providing confidentiality and 283 integrity for protected data, but created a problem for some network 284 management functions. Operators could no longer gain access to some 285 session streams resulting in actions by several to regain their 286 operational practices that previously depended on cleartext data 287 sessions. 289 The EFF reported [EFF2014] several network service providers using a 290 downgrade attack to prevent the use of SMTP over TLS by breaking 291 STARTTLS (section 3.2 of [RFC7525]), essentially preventing the 292 negotiation process resulting in fallback to the use of clear text. 293 There has already been documented cases of service providers 294 preventing STARTTLS to prevent session encryption negotiation on some 295 session to inject a super cookie to enable tracking of users for 296 advertisers, also considered an attack. These serves as examples of 297 undesirable behavior that could be prevented through upfront 298 discussions in protocol work for operators and protocol designers to 299 understand the implications of such actions. In other cases, some 300 service providers and enterprises have relied on middleboxes having 301 access to clear text for the purposes of load balancing, monitoring 302 for attack traffic, meeting regulatory requirements, or for other 303 purposes. The implications for enterprises, who own the data on 304 their networks is very different from service providers who may be 305 accessing content that violates privacy considerations. 306 Additionally, service provider equipment is designed for accessing 307 only the headers exposed for the data-link, network, and transport 308 layers. Delving deeper into packets is possible, but there is 309 typically a high degree of accuracy from the header information and 310 packet sizes when limited to header information from these three 311 layers. Service providers also have the option of adding routing 312 overlay protocols to traffic. These middlebox implementations, 313 whether performing functions considered legitimate by the IETF or 314 not, have been impacted by increases in encrypted traffic. Only 315 methods keeping with the goal of balancing network management and PM 316 mitigation in [RFC7258] should be considered in solution work 317 resulting from this document. 319 It is well known that national surveillance programs monitor traffic 320 [JNSLP] [RFC2804] [RFC7258] monitor for criminal activities. 321 Governments vary on their balance between monitoring versus the 322 protection of user privacy, data, and assets. Those that favor 323 unencrypted access to data ignore the real need to protect users' 324 identity, financial transactions and intellectual property, which 325 requires security and encryption to prevent crime. A clear 326 understanding of technology, encryption, and monitoring goals will 327 aid in the development of solutions as work continues towards finding 328 an appropriate balance allowing for management while protecting users 329 privacy with strong encryption solutions. 331 2. Network Service Provider Monitoring 333 Network Service Providers (SP) for this definition include the 334 backbone Internet Service providers as well as those providing 335 infrastructure at scale for core Internet use (hosted infrastructure 336 and services such as email). 338 Network service providers use various techniques to operate, manage, 339 and secure their networks. The following subsections detail the 340 purpose of several techniques and which protocol fields are used to 341 accomplish each task. In response to increased encryption of these 342 fields, some network service providers may be tempted to undertake 343 undesirable security practices in order to gain access to the fields 344 in unencrypted data flows. To avoid this situation, new methods 345 could be developed to accomplish the same goals without service 346 providers having the ability to see session data. 348 2.1. Passive Monitoring 350 2.1.1. Traffic Surveys 352 Internet traffic surveys are useful in many pursuits, such as input 353 for Center for Applied Internet Data Analysis (CAIDA) studies 354 [CAIDA], network planning and optimization. Tracking the trends in 355 Internet traffic growth, from earlier peer-to-peer communication to 356 the extensive adoption of unicast video streaming applications, has 357 relied on a view of traffic composition with a particular level of 358 assumed accuracy, based on access to cleartext by those conducting 359 the surveys. 361 Passive monitoring makes inferences about observed traffic using the 362 maximal information available, and is subject to inaccuracies 363 stemming from incomplete sampling (of packets in a stream) or loss 364 due to monitoring system overload. When encryption conceals more 365 layers in each packet, reliance on pattern inferences and other 366 heuristics grows, and accuracy suffers. For example, the traffic 367 patterns between server and browser are dependent on browser supplier 368 and version, even when the sessions use the same server application 369 (e.g., web e-mail access). It remains to be seen whether more 370 complex inferences can be mastered to produce the same monitoring 371 accuracy. 373 2.1.2. Troubleshooting 375 Network operators use protocol-dissecting analyzers when responding 376 to customer problems, to identify the presence of attack traffic, and 377 to identify root causes of the problem such as misconfiguration. In 378 limited cases, packet captures may also be used when a customer 379 approves of access to their packets or provides packet captures close 380 to the endpoint. The protocol dissection is generally limited to 381 supporting protocols (e.g., DNS, DHCP), network and transport (e.g., 382 IP, TCP), and some higher layer protocols (e.g., RTP, RTCP). 383 Troubleshooting will move closer to the endpoint with increased 384 encryption and adjustments in practices to effectively troubleshoot 385 using a 5-tuple may require education. Packet loss investigations, 386 and those where access is limited to a 2-tuple (IPsec tunnel mode), 387 rely on network and transport layer headers taken at the endpoint. 388 In this case, captures on intermediate nodes are not reliable as 389 there are far too many cases of aggregate interfaces and alternate 390 paths in service provider networks. 392 Network operators are often the first ones called upon to investigate 393 application problems (e.g., "my HD video is choppy"), to first rule 394 out network and network services as a cause for the underlying issue. 395 When diagnosing a customer problem, the starting point may be a 396 particular application that isn't working. The ability to identify 397 the problem application's traffic is important and packet capture 398 provided from the customer close to the edge may be used for this 399 purpose; IP address filtering is not useful for applications using 400 content delivery networks (CDNs) or cloud providers. After 401 identifying the traffic, an operator may analyze the traffic 402 characteristics and routing of the traffic. This diagnostic step is 403 important to help determine the root cause before exploring if the 404 issue is directly with the application. 406 For example, by investigating packet loss (from TCP sequence and 407 acknowledgement numbers), round-trip-time (from TCP timestamp options 408 or application-layer transactions, e.g., DNS or HTTP response time), 409 TCP receive-window size, packet corruption (from checksum 410 verification), inefficient fragmentation, or application-layer 411 problems, the operator can narrow the problem to a portion of the 412 network, server overload, client or server misconfiguration, etc. 413 Network operators may also be able to identify the presence of attack 414 traffic as not conforming to the application the user claims to be 415 using. In many instances, the exposed packet header is sufficient 416 for this type of troubleshooting. 418 One way of quickly excluding the network as the bottleneck during 419 troubleshooting is to check whether the speed is limited by the 420 endpoints. For example, the connection speed might instead be 421 limited by suboptimal TCP options, the sender's congestion window, 422 the sender temporarily running out of data to send, the sender 423 waiting for the receiver to send another request, or the receiver 424 closing the receive window. All this information can be derived from 425 the cleartext TCP header. 427 Packet captures and protocol-dissecting analyzers have been important 428 tools. Automated monitoring has also been used to proactively 429 identify poor network conditions, leading to maintenance and network 430 upgrades before user experience declines. For example, findings of 431 loss and jitter in VoIP traffic can be a predictor of future customer 432 dissatisfaction (supported by metadata from the RTP/RTCP protocol ) 433 [RFC3550], or increases in DNS response time can generally make 434 interactive web browsing appear sluggish. But to detect such 435 problems, the application or service stream must first be 436 distinguished from others. 438 When increased encryption is used, operators lose a source of data 439 that may be used to debug user issues. For example, IPsec obscures 440 TCP and RTP header information, while TLS and SRTP do not. Because 441 of this, application server operators using increased encryption 442 might be called upon more frequently to assist with debugging and 443 troubleshooting, and thus may want to consider what tools can be put 444 in the hands of their clients or network operators. 446 Further, the performance of some services can be more efficiently 447 managed and repaired when information on user transactions is 448 available to the service provider. It may be possible to continue 449 transaction monitoring activities without clear text access to the 450 application layers of interest, but inaccuracy will increase and 451 efficiency of repair activities will decrease. For example, an 452 application protocol error or failure would be opaque to network 453 troubleshooters when transport encryption is applied, making root 454 cause location more difficult and therefore increasing the time-to- 455 repair. Repair time directly reduces the availability of the 456 service, and most network operators have made availability a key 457 metric in their Service Level Agreements and/or subscription rebates. 458 Also, there may be more cases of user communication failures when the 459 additional encryption processes are introduced (e.g., key management 460 at large scale), leading to more customer service contacts and (at 461 the same time) less information available to network operations 462 repair teams. 464 In mobile networks, knowledge about TCP's stream transfer progress 465 (by observing ACKs, retransmissions, packet drops, and the Sector 466 Utilization Level etc.) is further used to measure the performance of 467 Network Segments (Sector, eNodeB (eNB) etc.). This information is 468 used as key performance indicators (KPIs) and for the estimation of 469 user/service key quality indicators at network edges for circuit 470 emulation (CEM) as well as input for mitigation methods. If the 471 make-up of active services per user and per sector are not visible to 472 a server that provides Internet Access Point Names (APN), it cannot 473 perform mitigation functions based on network segment view. 475 It is important to note that the push for encryption by application 476 providers has been motivated by the application of the described 477 techniques. Although network operators have noted performance 478 improvements with network-based optimization or enhancement of user 479 traffic (otherwise, deployment would not have occurred), application 480 providers have likewise noted some degraded performance and/or user 481 experience, and such cases may result in additional operator 482 troubleshooting. Further, encrypted application streams might avoid 483 outdated optimization or enhancement techniques, where they exist. 485 A gap exists for vendors where built-in diagnostics and 486 serviceability is not adequate to provide detailed logging and 487 debugging capabilities that, when possible, can access cleartext 488 network parameters. In addition to traditional logging and debugging 489 methods, packet tracing and inspection along the service path 490 provides operators the visibility to continue to diagnose problems 491 reported both internally and by their customers. Logging of service 492 path upon exit for routing overlay protocols will assist with policy 493 management and troubleshooting capabilities for traffic flows on 494 encrypted networks. Protocol trace logging and protocol data unit 495 (PDU) logging should also be considered to improve visibility to 496 monitor and troubleshoot application level traffic. Additional work 497 on this gap would assist network operators to better troubleshoot and 498 manage networks with increasing amounts of encrypted traffic. 500 2.1.3. Traffic Analysis Fingerprinting 502 Fingerprinting is used in traffic analysis and monitoring to identify 503 traffic streams that match certain patterns. This technique can be 504 used with both clear text or encrypted sessions. Some Distributed 505 Denial of Service (DDoS) prevention techniques at the network 506 provider level rely on the ability to fingerprint traffic in order to 507 mitigate the effect of this type of attack. Thus, fingerprinting may 508 be an aspect of an attack or part of attack countermeasures. 510 A common, early trigger for DDoS mitigation includes observing 511 uncharacteristic traffic volumes or sources; congestion; or 512 degradation of a given network or service. One approach to mitigate 513 such an attack involves distinguishing attacker traffic from 514 legitimate user traffic. The ability to examine layers and payloads 515 above transport provides an increased range of filtering 516 opportunities at each layer in the clear. If fewer layers are in the 517 clear, this means that there are reduced filtering opportunities 518 available to mitigate attacks. However, fingerprinting is still 519 possible. 521 Passive monitoring of network traffic can lead to invasion of privacy 522 by external actors at the endpoints of the monitored traffic. 523 Encryption of traffic end-to-end is one method to obfuscate some of 524 the potentially identifying information. For example, browser 525 fingerprints are comprised of many characteristics, including User 526 Agent, HTTP Accept headers, browser plug-in details, screen size and 527 color details, system fonts and time zone. A monitoring system could 528 easily identify a specific browser, and by correlating other 529 information, identify a specific user. 531 2.2. Traffic Optimization and Management 533 2.2.1. Load Balancers 535 A standalone load balancer is a function one can take off the shelf, 536 place in front of a pool of servers, configure appropriately, and it 537 will balance the traffic load among servers in the pool. This is a 538 typical setup for load balancers. Standalone load balancers rely on 539 the plainly observable information in the packets they are forwarding 540 and rely on industry-accepted standards in interpreting the plainly 541 observable information. Typically, this is a 5-tuple of the 542 connection. This type of configuration terminates TLS sessions at 543 the load balancer, making it the end point instead of the server. 544 Standalone load balancers are considered middleboxes, but are an 545 integral part of server infrastructure that scales. 547 In contrast, an integrated load balancer is developed to be an 548 integral part of the service provided by the server pool behind that 549 load balancer. These load balancers can communicate state with their 550 pool of servers to better route flows to the appropriate servers. 551 They rely on non-standard system-specific information and operational 552 knowledge shared between the load balancer and its servers. 554 Both standalone and integrated load balancers can be deployed in 555 pools for redundancy and load sharing. For high availability, it is 556 important that when packets belonging to a flow start to arrive at a 557 different load balancer in the load balancer pool, the packets 558 continue to be forwarded to the original server in the server pool. 559 The importance of this requirement increases as the chances of such 560 load balancer change event increases. 562 Mobile operators deploy integrated load balancers to assist with 563 maintaining connection state as devices migrate. With the 564 proliferation of mobile connected devices, there is an acute need for 565 connection-oriented protocols that maintain connections after a 566 network migration by an endpoint. This connection persistence 567 provides an additional challenge for multi-homed anycast-based 568 services typically employed by large content owners and Content 569 Distribution Networks (CDNs). The challenge is that a migration to a 570 different network in the middle of the connection greatly increases 571 the chances of the packets routed to a different anycast point-of- 572 presence (POP) due to the new network's different connectivity and 573 Internet peering arrangements. The load balancer in the new POP, 574 potentially thousands of miles away, will not have information about 575 the new flow and would not be able to route it back to the original 576 POP. 578 To help with the endpoint network migration challenges, anycast 579 service operations are likely to employ integrated load balancers 580 that, in cooperation with their pool servers, are able to ensure that 581 client-to-server packets contain some additional identification in 582 plainly-observable parts of the packets (in addition to the 5-tuple). 583 As noted in Section 2 of [RFC7258], careful consideration in protocol 584 design to mitigate PM is important, while ensuring manageability of 585 the network. 587 An area for further research includes end-to-end solutions that would 588 provide a simpler architecture and may solve the issue with CDN 589 anycast. In this case, connections would be migrated to a CDN 590 unicast address. 592 Current protocols, such as TCP, allow the development of stateless 593 integrated load balancers by availing such load balancers of 594 additional plain text information in client-to-server packets. In 595 case of TCP, such information can be encoded by having server- 596 generated sequence numbers (that are ACK'd by the client), segment 597 values, lengths of the packet sent, etc. The use of some of these 598 mechanisms for load balancing negates some of the security 599 assumptions associated with those primitives (e.g., that an off-path 600 attacker guessing valid sequence numbers for a flow is hard). 601 Another possibility is a dedicated mechanism for storing load 602 balancer state, such as QUIC's proposed connection ID to provide 603 visibility to the load balancer. An identifier could be used for 604 tracking purposes, but this may provide an option that is an 605 improvement from bolting it on to an unrelated transport signal. 606 This method allows for tight control by one of the endpoints and can 607 be rotated to avoid roving client linkability: in other words, being 608 a specific, separate signal, it can be governed in a way that is 609 finely targeted at that specific use-case. 611 Some integrated load balancers have the ability to use additional 612 plainly observable information even for today's protocols that are 613 not network migration tolerant. This additional information allows 614 for improved availability and scalability of the load balancing 615 operation. For example, BGP reconvergence can cause a flow to switch 616 anycast POPs even without a network change by any endpoint. 617 Additionally, a system that is able to encode the identity of the 618 pool server in plain text information available in each incoming 619 packet is able to provide stateless load balancing. This ability 620 confers great reliability and scalability advantages even if the flow 621 remains in a single POP, because the load balancing system is not 622 required to keep state of each flow. Even more importantly, there's 623 no requirement to continuously synchronize such state among the pool 624 of load balancers. An integrated load balancer repurposing limited 625 existing bits in transport flow state must maintain and synchronize 626 per-flow state occasionally: using the sequence number as a cookie 627 only works for so long given that there aren't that many bits 628 available to divide across a pool of machines. 630 Mobile operators apply Self Organizing Networks (3GPP SON) for 631 intelligent workflows such as content-aware MLB (Mobility Load 632 Balancing). Where network load balancers have been configured to 633 route according to application-layer semantics, an encrypted payload 634 is effectively invisible. This has resulted in practices of 635 intercepting TLS in front of load balancers to regain that 636 visibility, but at a cost to security and privacy. 638 In future Network Function Virtualization (NFV) architectures, load 639 balancing functions are likely to be more prevalent (deployed at 640 locations throughout operators' networks). NFV environments will 641 require some type of identifier (IPv6 flow identifiers, the proposed 642 QUIC connection ID, etc.) for managing traffic using encrypted 643 tunnels. The shift to increased encryption will have an impact to 644 visibility of flow information and will require adjustments to 645 perform similar load balancing functions within an NFV. 647 2.2.2. Differential Treatment based on Deep Packet Inspection (DPI) 649 Data transfer capacity resources in cellular radio networks tend to 650 be more constrained than in fixed networks. This is a result of 651 variance in radio signal strength as a user moves around a cell, the 652 rapid ingress and egress of connections as users hand off between 653 adjacent cells, and temporary congestion at a cell. Mobile networks 654 alleviate this by queuing traffic according to its required bandwidth 655 and acceptable latency: for example, a user is unlikely to notice a 656 20ms delay when receiving a simple Web page or email, or an instant 657 message response, but will very likely notice a re-buffering pause in 658 a video playback or a VoIP call de-jitter buffer. Ideally, the 659 scheduler manages the queue so that each user has an acceptable 660 experience as conditions vary, but inferences of the traffic type 661 have been used to make bearer assignments and set scheduler priority. 663 Deep Packet Inspection (DPI) allows identification of applications 664 based on payload signatures, in contrast to trusting well-known port 665 numbers. Application and transport layer encryption make the traffic 666 type estimation more complex and less accurate, and therefore it may 667 not be effectual to use this information as input for queue 668 management. With the use of WebSockets [RFC6455], for example, many 669 forms of communications (from isochronous/real-time to bulk/elastic 670 file transfer) will take place over HTTP port 80 or port 443, so only 671 the messages and higher-layer data will make application 672 differentiation possible. If the monitoring system sees only "HTTP 673 port 443", it cannot distinguish application streams that would 674 benefit from priority queueing from others that would not. 676 Mobile networks especially rely on content/application based 677 prioritization of Over-the-Top (OTT) services - each application-type 678 or service has different delay/loss/throughput expectations, and each 679 type of stream will be unknown to an edge device if encrypted; this 680 impedes dynamic-QoS adaptation. An alternate way to achieve 681 encrypted application separation is possible when the User Equipment 682 (UE) requests a dedicated bearer for the specific application stream 683 (known by the UE), using a mechanism such as the one described in 684 Section 6.5 of 3GPP TS 24.301 [TS3GPP]. The UE's request includes 685 the Quality Class Indicator (QCI) appropriate for each application, 686 based on their different delay/loss/throughput expectations. 687 However, UE requests for dedicated bearers and QCI may not be 688 supported at the subscriber's service level, or in all mobile 689 networks. 691 These effects and potential alternative solutions have been discussed 692 at the accord BoF [ACCORD] at IETF95. 694 This section does not consider traffic discrimination by service 695 providers related to NetNeutrality, where traffic may be favored 696 according to the service provider preference as opposed to the user's 697 preference. These use cases are considered out-of-scope for this 698 document as controversial practices. 700 2.2.3. Network Congestion Management 702 For User Plane Congestion Management (3GPP UPCON) [UPCON], the 703 ability to understand content and manage networks during periods of 704 congestion is the focus of this 3GPP work item. Mitigating 705 techniques such as deferred download, off-peak acceleration, and 706 outbound roamers are a few examples of the areas explored in the 707 associated 3GPP documents. The documents describe the issues, the 708 data utilized in managing congestion, and make policy 709 recommendations. 711 2.2.4. Performance-enhancing Proxies 713 Performance-enhancing TCP proxies may perform local retransmission at 714 the network edge; this also applies to mobile networks. In TCP, 715 duplicated ACKs are detected and potentially concealed when the proxy 716 retransmits a segment that was lost on the mobile link without 717 involvement of the far end (see section 2.1.1 of [RFC3135] and 718 section 3.5 of [I-D.dolson-plus-middlebox-benefits]). 720 Operators report that this optimization at network edges improves 721 real-time transmission over long delay Internet paths or networks 722 with large capacity-variation (such as mobile/cellular networks). 723 However, such optimizations can also cause problems with performance, 724 for example if the characteristics of some packet streams begin to 725 vary significantly from those considered in the proxy design. 727 In general some operators have stated that performance-enhancing 728 proxies have a lower Round-Trip Time (RTT) to the client and 729 therefore determine the responsiveness of flow control. A lower RTT 730 makes the flow control loop more responsive to changes in the mobile 731 network conditions and enables faster adaptation in a delay and 732 capacity varying network due to user mobility. 734 Further, some use service-provider-operated proxies to reduce the 735 control delay between the sender and a receiver on a mobile network 736 where resources are limited. The RTT determines how quickly a user's 737 attempt to cancel a video is recognized and therefore how quickly the 738 traffic is stopped, thus keeping un-wanted video packets from 739 entering the radio scheduler queue. If impacted by encryption, 740 performance enhancing proxies could make use of routing overlay 741 protocols to accomplish the same task, but this results in additional 742 overhead. 744 An application-type-aware network edge (middlebox) can further 745 control pacing, limit simultaneous HD videos, or prioritize active 746 videos against new videos, etc. Services at this more granular level 747 are limited with the use of encryption. 749 Performance enhancing proxies are primarily used on long delay links 750 (satellite) with access to the TCP header to provide an early ACK and 751 make the long delay link of the path seem shorter. With some 752 specific forms of flow control, TCP can be more efficient than 753 alternatives such as proxies. The editors cannot cite research on 754 this point specific to the performance enhancing proxies described, 755 but agree this area could be explored to determine if flow-control 756 modifications could preserve the end-to-end performance on long delay 757 paths session where the TCP header is exposed. 759 2.2.5. Caching and Content Replication Near the Network Edge 761 The features and efficiency of some Internet services can be 762 augmented through analysis of user flows and the applications they 763 provide. For example, network caching of popular content at a 764 location close to the requesting user can improve delivery efficiency 765 (both in terms of lower request response times and reduced use of 766 International Internet links when content is remotely located), and 767 the service provider through an authorized agreement acting on their 768 behalf use DPI in combination with content distribution networks to 769 determine if they can intervene effectively. Encryption of packet 770 contents at a given protocol layer usually makes DPI processing of 771 that layer and higher layers impossible. That being said, it should 772 be noted that some content providers prevent caching to control 773 content delivery through the use of encrypted end-to-end sessions. 774 CDNs vary in their deployment options of end-to-end encryption. The 775 business risk of losing control of content is a motivation outside of 776 privacy and pervasive monitoring that are driving end-to-end 777 encryption for these content providers. 779 It should be noted that caching was first supported in [RFC1945] and 780 continued in the recent update of "Hypertext Transfer Protocol 781 (HTTP/1.1): Caching" in [RFC7234]. Some operators also operate 782 transparent caches which neither the user nor the origin opt-in. The 783 use of these caches is controversial within IETF and is generally 784 precluded by the use of HTTPS. 786 Content replication in caches (for example live video, Digital Rights 787 Management (DRM) protected content) is used to most efficiently 788 utilize the available limited bandwidth and thereby maximize the 789 user's Quality of Experience (QoE). Especially in mobile networks, 790 duplicating every stream through the transit network increases 791 backhaul cost for live TV. The Enhanced Multimedia Broadcast/ 792 Multicast Services (3GPP eMBMS) utilizes trusted edge proxies to 793 facilitate delivering the same stream to different users, using 794 either unicast or multicast depending on channel conditions to the 795 user. There are on-going efforts to support multicast inside carrier 796 networks while preserving end-to-end security: Automatic Multicast 797 Tunneling (AMT), for instance, allows CDNs to deliver a single 798 (potentially encrypted) copy of a live stream to a carrier network 799 over the public internet and for the carrier to then distribute that 800 live stream as efficiently as possible within its own network using 801 multicast. 803 Alternate approaches are in the early phase of being explored to 804 allow caching of encrypted content. These solutions require 805 cooperation from content owners and fall outside the scope of what is 806 covered in this document. Content delegation allows for replication 807 with possible benefits, but any form of delegation has the potential 808 to affect the expectation of client-server confidentiality. 810 2.2.6. Content Compression 812 In addition to caching, various applications exist to provide data 813 compression in order to conserve the life of the user's mobile data 814 plan or make delivery over the mobile link more efficient. The 815 compression proxy access can be built into a specific user level 816 application, such as a browser, or it can be available to all 817 applications using a system level application. The primary method is 818 for the mobile application to connect to a centralized server as a 819 transparent proxy (user does not opt-in), with the data channel 820 between the client application and the server using compression to 821 minimize bandwidth utilization. The effectiveness of such systems 822 depends on the server having access to unencrypted data flows. 824 Aggregated data stream content compression that spans objects and 825 data sources that can be treated as part of a unified compression 826 scheme (e.g., through the use of a shared segment store) is often 827 effective at providing data offload when there is a network element 828 close to the receiver that has access to see all the content. 830 2.2.7. Service Function Chaining 832 Service Function Chaining (SFC) has been defined in RFC7665 [RFC7665] 833 and RFC8300 [RFC8300]. As discussed in RFC7498 [RFC7498], common SFC 834 deployments may use classifiers to direct traffic into VLANs instead 835 of using NSH, as defined in RFC8300 [RFC8300]. As described in 836 RFC7665 [RFC7665], the ordered steering of traffic to support 837 specific optimizations depends upon the ability of a Classifier to 838 determine the microflows. RFC2474 [RFC2474] defines "Microflow: a 839 single instance of an application-to-application flow of packets 840 which is identified by source address, destination address, protocol 841 id, and source port, destination port (where applicable)." SFC 842 currently depends upon a classifier to at least identify the 843 microflow. As the classifier's visibility is reduced from a 5-tuple 844 to a 2-tuple, or if information above the transport layer becomes 845 inaccessible, then the SFC Classifier is not able to perform its job 846 and the service functions of the path may be adversely affected. 848 There are also mechanisms provided to protect security and privacy. 849 In the SFC case, the layer below a network service header can be 850 protected with session encryption. A goal is protecting end-user 851 data, while retaining the intended functions of RFC7665 [RFC7665] at 852 the same time. 854 2.3. Content Filtering, Network Access, and Accounting 856 Mobile Networks and many ISPs operate under the regulations of their 857 licensing government authority. These regulations include Lawful 858 Intercept, adherence to Codes of Practice on content filtering, and 859 application of court order filters. Such regulations assume network 860 access to provide content filtering and accounting, as discussed 861 below. As previously stated, the intent of this document is to 862 document existing practices; the development of IETF protocols 863 follows the guiding principles of [RFC1984] and [RFC2804] and 864 explicitly do not support tools and methods that could be used for 865 wiretapping and censorship. 867 2.3.1. Content Filtering 869 There are numerous reasons why service providers might block content: 870 to comply with requests from law enforcement or regulatory 871 authorities, to effectuate parental controls, to enforce content- 872 based billing, or for other reasons, possibly considered 873 inappropriate by some. See RFC7754 [RFC7754] for a survey of 874 Internet filtering techniques and motivations and the IAB consensus 875 on those mechanisms. This section is intended to document a 876 selection of current content blocking practices by operators and the 877 effects of encryption on those practices. Content blocking may also 878 happen at endpoints or at the edge of enterprise networks, but those 879 are not addressed in this section. 881 In a mobile network content filtering usually occurs in the core 882 network. With other networks, content filtering could occur in the 883 core network or at the edge. A proxy is installed which analyses the 884 transport metadata of the content users are viewing and either 885 filters content based on a blacklist of sites or based on the user's 886 pre-defined profile (e.g. for age sensitive content). Although 887 filtering can be done by many methods, one commonly used method 888 involves a trigger based on the proxy identifying a DNS lookup of a 889 host name in a URL which appears on a blacklist being used by the 890 operator. The subsequent requests to that domain will be re-routed 891 to a proxy which checks whether the full URL matches a blocked URL on 892 the list, and will return a 404 if a match is found. All other 893 requests should complete. This technique does not work in situations 894 where DNS traffic is encrypted (e.g., by employing [RFC7858] ). This 895 method is also used by other types of network providers enabling 896 traffic inspection, but not modification. 898 Content filtering via a proxy can also utilize an intercepting 899 certificate where the client's session is terminated at the proxy 900 enabling for cleartext inspection of the traffic. A new session is 901 created from the intercepting device to the client's destination; 902 this is an opt-in strategy for the client, where the endpoint is 903 configured to trust the intercepting certificate. Changes to TLSv1.3 904 do not impact this more invasive method of interception, that has the 905 potential to expose every HTTPS session to an active man in the 906 middle (MitM). 908 Another form of content filtering is called parental control, where 909 some users are deliberately denied access to age-sensitive content as 910 a feature to the service subscriber. Some sites involve a mixture of 911 universal and age-sensitive content and filtering software. In these 912 cases, more granular (application layer) metadata may be used to 913 analyze and block traffic. Methods that accessed cleartext 914 application-layer metadata no longer work when sessions are 915 encrypted. This type of granular filtering could occur at the 916 endpoint or as a proxy service. However, the lack of ability to 917 efficiently manage endpoints as a service reduces network service 918 providers' ability to offer parental control. 920 2.3.2. Network Access and Data Usage 922 Approved access to a network is a prerequisite to requests for 923 Internet traffic. 925 However, there are cases (beyond parental control) when a network 926 service provider currently redirects customer requests for content 927 (affecting content accessibility): 929 1. The network service provider is performing the accounting and 930 billing for the content provider, and the customer has not (yet) 931 purchased the requested content. 933 2. Further content may not be allowed as the customer has reached 934 their usage limit and needs to purchase additional data service, 935 which is the usual billing approach in mobile networks. 937 Currently, some network service providers redirect the customer using 938 HTTP redirect to a captive portal page that explains to those 939 customers the reason for the blockage and the steps to proceed. 940 [RFC6108] describes one viable web notification system. When the 941 HTTP headers and content are encrypted, this appropriately prevents 942 mobile carriers from intercepting the traffic and performing an HTTP 943 redirect. As a result, some mobile carriers block customer's 944 encrypted requests, which impacts customer experience because the 945 blocking reason must be conveyed by some other means. The customer 946 may need to call customer care to find out the reason and/or resolve 947 the issue, possibly extending the time needed to restore their 948 network access. While there are well deployed alternate SMS-based 949 solutions that do not involve out of specification protocol 950 interception, this is still an unsolved problem for non-SMS users. 952 Further, when the requested service is about to consume the remainder 953 of the user's plan limits, the transmission could be terminated and 954 advance notifications may be sent to the user by their service 955 provider to warn the user ahead of the exhausted plan. If web 956 content is encrypted, the network provider cannot know the data 957 transfer size at request time. Lacking this visibility of the 958 application type and content size, the network would continue the 959 transmission and stop the transfer when the limit was reached. A 960 partial transfer may not be usable by the client wasting both network 961 and user resources, possibly leading to customer complaints. The 962 content provider does not know user's service plans or current usage, 963 and cannot warn the user of plan exhaustion. 965 In addition, some mobile network operators sell tariffs that allow 966 free-data access to certain sites, known as 'zero rating'. A session 967 to visit such a site incurs no additional cost or data usage to the 968 user. For some implementations, zero rating is impacted if 969 encryption hides the details of the content domain from the network. 971 2.3.3. Application Layer Gateways 973 Application Layer Gateways (ALG) assist applications to set 974 connectivity across Network Address Translators (NAT), Firewalls, 975 and/or Load Balancers for specific applications running across mobile 976 networks. Section 2.9 of [RFC2663] describes the role of ALGs and 977 their interaction with NAT and/or application payloads. ALG are 978 deployed with an aim to improve connectivity. However, it is an IETF 979 Best Common Practice recommendation that ALGs for UDP-based protocols 980 should be turned off [RFC4787]. 982 One example of an ALG in current use is aimed at video applications 983 that use the Real Time Session Protocol (RTSP) [RFC7826] primary 984 stream as a means to identify related Real Time Protocol/Real Time 985 Control Protocol (RTP/RTCP) [RFC3550] flows at set-up. The ALG in 986 this case relies on the 5-tuple flow information derived from RTSP to 987 provision NAT or other middleboxes and provide connectivity. 988 Implementations vary, and two examples follow: 990 1. Parse the content of the RTSP stream and identify the 5-tuple of 991 the supporting streams as they are being negotiated. 993 2. Intercept and modify the 5-tuple information of the supporting 994 media streams as they are being negotiated on the RTSP stream, 995 which is more intrusive to the media streams. 997 When RTSP stream content is encrypted, the 5-tuple information within 998 the payload is not visible to these ALG implementations, and 999 therefore they cannot provision their associated middleboxes with 1000 that information. 1002 The deployment of IPv6 may well reduce the need for NAT, and the 1003 corresponding requirement for Application Layer Gateways. 1005 2.3.4. HTTP Header Insertion 1007 Some mobile carriers use HTTP header insertion (see section 3.2.1 of 1008 [RFC7230]) to provide information about their customers to third 1009 parties or to their own internal systems [Enrich]. Third parties use 1010 the inserted information for analytics, customization, advertising, 1011 cross-site tracking of users, to bill the customer, or to selectively 1012 allow or block content. HTTP header insertion is also used to pass 1013 information internally between a mobile service provider's sub- 1014 systems, thus keeping the internal systems loosely coupled. When 1015 HTTP connections are encrypted to protect users privacy, mobile 1016 network service providers cannot insert headers to accomplish the, 1017 sometimes considered controversial, functions above. 1019 Guidance from the Internet Architecture Board has been provided in 1020 RFC8165 [RFC8165] on Design Considerations for Metadata Insertion. 1021 The guidance asserts that designs that share metadata only by 1022 explicit actions at the host are preferable to designs in which 1023 middleboxes insert metadata. Alternate notification methods that 1024 follow this and other guidance would be helpful to mobile carriers. 1026 3. Encryption in Hosting and Application SP Environments 1028 Hosted environments have had varied requirements in the past for 1029 encryption, with many businesses choosing to use these services 1030 primarily for data and applications that are not business or privacy 1031 sensitive. A shift prior to the revelations on surveillance/passive 1032 monitoring began where businesses were asking for hosted environments 1033 to provide higher levels of security so that additional applications 1034 and service could be hosted externally. Businesses understanding the 1035 threats of monitoring in hosted environments increased that pressure 1036 to provide more secure access and session encryption to protect the 1037 management of hosted environments as well as for the data and 1038 applications. 1040 3.1. Management Access Security 1042 Hosted environments may have multiple levels of management access, 1043 where some may be strictly for the Hosting SP (infrastructure that 1044 may be shared among customers) and some may be accessed by a specific 1045 customer for application management. In some cases, there are 1046 multiple levels of hosting service providers, further complicating 1047 the security of management infrastructure and the associated 1048 requirements. 1050 Hosting service provider management access is typically segregated 1051 from other traffic with a control channel and may or may not be 1052 encrypted depending upon the isolation characteristics of the 1053 management session. Customer access may be through a dedicated 1054 connection, but discussion for that connection method is out-of-scope 1055 for this document. 1057 In overlay networks (e.g. VXLAN, Geneve, etc.) that are used to 1058 provide hosted services, management access for a customer to support 1059 application management may depend upon the security mechanisms 1060 available as part of that overlay network. While overlay network 1061 data encapsulations may be used to indicate the desired isolation, 1062 this is not sufficient to prevent deliberate attacks that are aware 1063 of the use of the overlay network. 1064 [I-D.mglt-nvo3-geneve-security-requirements] describes requirements 1065 to handle attacks. It is possible to use an overlay header in 1066 combination with IPsec or other encrypted traffic sessions, but this 1067 adds the requirement for authentication infrastructure and may reduce 1068 packet transfer performance. The use of an overlay header may also 1069 be deployed as a mechanism to manage encrypted traffic streams on the 1070 network by network service providers. Additional extension 1071 mechanisms to provide integrity and/or privacy protections are being 1072 investigated for overlay encapsulations. Section 7 of [RFC7348] 1073 describes some of the security issues possible when deploying VXLAN 1074 on Layer 2 networks. Rogue endpoints can join the multicast groups 1075 that carry broadcast traffic, for example. 1077 3.1.1. Customer Access Monitoring 1079 Hosted applications that allow some level of customer management 1080 access may also require monitoring by the hosting service provider. 1081 Monitoring could include access control restrictions such as 1082 authentication, authorization, and accounting for filtering and 1083 firewall rules to ensure they are continuously met. Customer access 1084 may occur on multiple levels, including user-level and administrative 1085 access. The hosting service provider may need to monitor access 1086 either through session monitoring or log evaluation to ensure 1087 security service level agreements (SLA) for access management are 1088 met. The use of session encryption to access hosted environments 1089 limits access restrictions to the metadata described below. 1090 Monitoring and filtering may occur at an: 1092 2-tuple IP-level with source and destination IP addresses alone, or 1094 5-tuple IP and protocol-level with source IP address, destination IP 1095 address, protocol number, source port number, and destination port 1096 number. 1098 Session encryption at the application level, TLS for example, 1099 currently allows access to the 5-tuple. IP-level encryption, such as 1100 IPsec in tunnel mode prevents access to the original 5-tuple and may 1101 limit the ability to restrict traffic via filtering techniques. This 1102 shift may not impact all hosting service provider solutions as 1103 alternate controls may be used to authenticate sessions or access may 1104 require that clients access such services by first connecting to the 1105 organization before accessing the hosted application. Shifts in 1106 access may be required to maintain equivalent access control 1107 management. Logs may also be used for monitoring that access control 1108 restrictions are met, but would be limited to the data that could be 1109 observed due to encryption at the point of log generation. Log 1110 analysis is out of scope for this document. 1112 3.1.2. SP Content Monitoring of Applications 1114 The following observations apply to any IT organization that is 1115 responsible for delivering services, whether to third-parties, for 1116 example as a web based service, or to internal customers in an 1117 enterprise, e.g. a data processing system that forms a part of the 1118 enterprise's business. 1120 Organizations responsible for the operation of a data center have 1121 many processes which access the contents of IP packets (passive 1122 methods of measurement, as defined in [RFC7799]). These processes 1123 are typically for service assurance or security purposes as part of 1124 their data center operations. 1126 Examples include: 1128 - Network Performance Monitoring/Application Performance 1129 Monitoring 1131 - Intrusion defense/prevention systems 1133 - Malware detection 1135 - Fraud Monitoring 1137 - Application DDOS protection 1139 - Cyber-attack investigation 1141 - Proof of regulatory compliance 1143 - Data Leakage Prevention 1145 Many application service providers simply terminate sessions to/from 1146 the Internet at the edge of the data center in the form of SSL/TLS 1147 offload in the load balancer. Not only does this reduce the load on 1148 application servers, it simplifies the processes to enable monitoring 1149 of the session content. 1151 However, in some situations, encryption deeper in the data center may 1152 be necessary to protect personal information or in order to meet 1153 industry regulations, e.g. those set out by the Payment Card Industry 1154 (PCI). In such situations, various methods have been used to allow 1155 service assurance and security processes to access unencrypted data. 1156 These include SSL/TLS decryption in dedicated units, which then 1157 forward packets to SP-controlled tools, or by real-time or post- 1158 capture decryption in the tools themselves. The use of passive tools 1159 that perform SSL/TLS decryption are impacted by the increased use of 1160 encryption that prevents monitoring via interception, while providing 1161 forward secrecy. 1163 Data center operators may also maintain packet recordings in order to 1164 be able to investigate attacks, breach of internal processes, etc. 1165 In some industries, organizations may be legally required to maintain 1166 such information for compliance purposes. Investigations of this 1167 nature have used access to the unencrypted contents of the packet. 1168 Alternate methods to investigate attacks or breach of process will 1169 rely on endpoint information, such as logs. As previously noted, 1170 logs often lack complete information, and this is seen as a concern 1171 resulting in some relying on session access for additional 1172 information. 1174 Application Service Providers may offer content-level monitoring 1175 options to detect intellectual property leakage, or other attacks. 1176 In service provider environments where Data Loss Prevention (DLP) has 1177 been implemented on the basis of the service provider having 1178 cleartext access to session streams, the use of encrypted streams 1179 prevents these implementations from conducting content searches for 1180 the keywords or phrases configured in the DLP system. DLP is often 1181 used to prevent the leakage of Personally Identifiable Information 1182 (PII) as well as financial account information, Personal Health 1183 Information (PHI), and Payment Card Information (PCI). If session 1184 encryption is terminated at a gateway prior to accessing these 1185 services, DLP on session data can still be performed. The decision 1186 of where to terminate encryption to hosted environments will be a 1187 risk decision made between the application service provider and 1188 customer organization according to their priorities. DLP can be 1189 performed at the server for the hosted application and on an end 1190 user's system in an organization as alternate or additional 1191 monitoring points of content; however, this is not frequently done in 1192 a service provider environment. 1194 Application service providers, by their very nature, control the 1195 application endpoint. As such, much of the information gleaned from 1196 sessions are still available on that endpoint. However, when a gap 1197 exists in the application's logging and debugging capabilities, this 1198 has led the application service provider to access data-in-transport 1199 for monitoring and debugging. 1201 3.2. Hosted Applications 1203 Organizations are increasingly using hosted applications rather than 1204 in-house solutions that require maintenance of equipment and 1205 software. Examples include Enterprise Resource Planning (ERP) 1206 solutions, payroll service, time and attendance, travel and expense 1207 reporting among others. Organizations may require some level of 1208 management access to these hosted applications and will typically 1209 require session encryption or a dedicated channel for this activity. 1211 In other cases, hosted applications may be fully managed by a hosting 1212 service provider with service level agreement expectations for 1213 availability and performance as well as for security functions 1214 including malware detection. Due to the sensitive nature of these 1215 hosted environments, the use of encryption is already prevalent. Any 1216 impact may be similar to an enterprise with tools being used inside 1217 of the hosted environment to monitor traffic. Additional concerns 1218 were not reported in the call for contributions. 1220 3.2.1. Monitoring Managed Applications 1222 Performance, availability, and other aspects of a SLA are often 1223 collected through passive monitoring. For example: 1225 o Availability: ability to establish connections with hosts to 1226 access applications, and discern the difference between network or 1227 host-related causes of unavailability. 1229 o Performance: ability to complete transactions within a target 1230 response time, and discern the difference between network or host- 1231 related causes of excess response time. 1233 Here, as with all passive monitoring, the accuracy of inferences are 1234 dependent on the cleartext information available, and encryption 1235 would tend to reduce the information and therefore, the accuracy of 1236 each inference. Passive measurement of some metrics will be 1237 impossible with encryption that prevents inferring packet 1238 correspondence across multiple observation points, such as for packet 1239 loss metrics. 1241 Application logging currently lacks detail sufficient to make 1242 accurate inferences in an environment with increased encryption, and 1243 so this constitutes a gap for passive performance monitoring (which 1244 could be closed if log details are enhanced in the future). 1246 3.2.2. Mail Service Providers 1248 Mail (application) service providers vary in what services they 1249 offer. Options may include a fully hosted solution where mail is 1250 stored external to an organization's environment on mail service 1251 provider equipment or the service offering may be limited to monitor 1252 incoming mail to remove spam [Section 5.1], malware [Section 5.6], 1253 and phishing attacks [Section 5.3] before mail is directed to the 1254 organization's equipment. In both of these cases, content of the 1255 messages and headers is monitored to detect spam, malware, phishing, 1256 and other messages that may be considered an attack. 1258 STARTTLS should have zero effect on anti-spam efforts for SMTP 1259 traffic. Anti-spam services could easily be performed on an SMTP 1260 gateway, eliminating the need for TLS decryption services. The 1261 impact to anti-spam service providers should be limited to a change 1262 in tools, where middleboxes were deployed to perform these functions. 1264 Many efforts are emerging to improve user-to-user encryption, 1265 including promotion of PGP and newer efforts such as Dark Mail 1266 [DarkMail]. Of course, content-based spam filtering will not be 1267 possible on encrypted content. 1269 3.3. Data Storage 1271 Numerous service offerings exist that provide hosted storage 1272 solutions. This section describes the various offerings and details 1273 the monitoring for each type of service and how encryption may impact 1274 the operational and security monitoring performed. 1276 Trends in data storage encryption for hosted environments include a 1277 range of options. The following list is intentionally high-level to 1278 describe the types of encryption used in coordination with data 1279 storage that may be hosted remotely, meaning the storage is 1280 physically located in an external data center requiring transport 1281 over the Internet. Options for monitoring will vary with each 1282 encryption approach described below. In most cases, solutions have 1283 been identified to provide encryption while ensuring management 1284 capabilities were maintained through logging or other means. 1286 3.3.1. Object-level Encryption 1288 For higher security and/or privacy of data and applications, options 1289 that provide end-to-end encryption of the data from the user's 1290 desktop or server to the storage platform may be preferred. This 1291 description includes any solution that encrypts data at the object 1292 level, not transport level. Encryption of data may be performed with 1293 libraries on the system or at the application level, which includes 1294 file encryption services via a file manager. Object-level encryption 1295 is useful when data storage is hosted, or scenarios when the storage 1296 location is determined based on capacity or based on a set of 1297 parameters to automate decisions. This could mean that large data 1298 sets accessed infrequently could be sent to an off-site storage 1299 platform at an external hosting service, data accessed frequently may 1300 be stored locally, or the decision could be based on the transaction 1301 type. Object-level encryption is grouped separately for the purpose 1302 of this document since data may be stored in multiple locations 1303 including off-site remote storage platforms. If session encryption 1304 is also used, the protocol is likely to be TLS. 1306 Impacts to monitoring may include access to content inspection for 1307 data leakage prevention and similar technologies, depending on their 1308 placement in the network. 1310 3.3.1.1. Monitoring for Hosted Storage 1312 Monitoring of hosted storage solutions that use host-level (object) 1313 encryption is described in this subsection. Host-level encryption 1314 can be employed for backup services, and occasionally for external 1315 storage services (operated by a third party) when internal storage 1316 limits are exceeded. 1318 Monitoring of data flows to hosted storage solutions is performed for 1319 security and operational purposes. The security monitoring may be to 1320 detect anomalies in the data flows that could include changes to 1321 destination, the amount of data transferred, or alterations in the 1322 size and frequency of flows. Operational considerations include 1323 capacity and availability monitoring. 1325 3.3.2. Disk Encryption, Data at Rest 1327 There are multiple ways to achieve full disk encryption for stored 1328 data. Encryption may be performed on data to be stored while in 1329 transit close to the storage media with solutions like Controller 1330 Based Encryption (CBE) or in the drive system with Self-Encrypting 1331 Drives (SED). Session encryption is typically coupled with 1332 encryption of these data at rest (DAR) solutions to also protect data 1333 in transit. Transport encryption is likely via TLS. 1335 3.3.2.1. Monitoring Session Flows for Data at Rest (DAR) Solutions 1337 Monitoring for transport of data to storage platforms, where object 1338 level encryption is performed close to or on the storage platform are 1339 similar to those described in the section on Monitoring for Hosted 1340 Storage. The primary difference for these solutions is the possible 1341 exposure of sensitive information, which could include privacy 1342 related data, financial information, or intellectual property if 1343 session encryption via TLS is not deployed. Session encryption is 1344 typically used with these solutions, but that decision would be based 1345 on a risk assessment. There are use cases where DAR or disk-level 1346 encryption is required. Examples include preventing exposure of data 1347 if physical disks are stolen or lost. In the case where TLS is in 1348 use, monitoring and the exposure of data is limited to a 5-tuple. 1350 3.3.3. Cross Data Center Replication Services 1352 Storage services also include data replication which may occur 1353 between data centers and may leverage Internet connections to tunnel 1354 traffic. The traffic may use iSCSI [RFC7143] or FC/IP [RFC7146] 1355 encapsulated in IPsec. Either transport or tunnel mode may be used 1356 for IPsec depending upon the termination points of the IPsec session, 1357 if it is from the storage platform itself or from a gateway device at 1358 the edge of the data center respectively. 1360 3.3.3.1. Monitoring Of IPsec for Data Replication Services 1362 Monitoring of data flows between data centers (for data replication) 1363 may be performed for security and operational purposes and would 1364 typically concentrate more on operational aspects since these flows 1365 are essentially virtual private networks (VPN) between data centers. 1366 Operational considerations include capacity and availability 1367 monitoring. The security monitoring may be to detect anomalies in 1368 the data flows, similar to what was described in the "Monitoring for 1369 Hosted Storage Section". If IPsec tunnel mode is in use, monitoring 1370 is limited to a 2-tuple, or with transport mode, a 5-tuple. 1372 4. Encryption for Enterprises 1374 Encryption of network traffic within the private enterprise is a 1375 growing trend, particularly in industries with audit and regulatory 1376 requirements. Some enterprise internal networks are almost 1377 completely TLS and/or IPsec encrypted. 1379 For each type of monitoring, different techniques and access to parts 1380 of the data stream are part of current practice. As we transition to 1381 an increased use of encryption, alternate methods of monitoring for 1382 operational purposes may be necessary to reduce the practice of 1383 breaking encryption (other policies may apply in some enterprise 1384 settings). 1386 4.1. Monitoring Practices of the Enterprise 1388 Large corporate enterprises are the owners of the platforms, data, 1389 and network infrastructure that provide critical business services to 1390 their user communities. As such, these enterprises are responsible 1391 for all aspects of the performance, availability, security, and 1392 quality of experience for all user sessions. In many such 1393 enterprises, users are required to consent to the enterprise 1394 monitoring all their activities as a condition of employment. 1395 Subsections of 4. Encryption for Enterprises may discuss techniques 1396 that access data beyond the data-link, network, and transport level 1397 headers typically used in SP networks since the corporate enterprise 1398 owns the data. These responsibilities break down into three basic 1399 areas: 1401 1. Security Monitoring and Control 1403 2. Application Performance Monitoring and Reporting 1405 3. Network Diagnostics and Troubleshooting 1407 In each of the above areas, technical support teams utilize 1408 collection, monitoring, and diagnostic systems. Some organizations 1409 currently use attack methods such as replicated TLS server RSA 1410 private keys to decrypt passively monitored copies of encrypted TLS 1411 packet streams. 1413 For an enterprise to avoid costly application down time and deliver 1414 expected levels of performance, protection, and availability, some 1415 forms of traffic analysis, sometimes including examination of packet 1416 payloads, are currently used. 1418 4.1.1. Security Monitoring in the Enterprise 1420 Enterprise users are subject to the policies of their organization 1421 and the jurisdictions in which the enterprise operates. As such, 1422 proxies may be in use to: 1424 1. intercept outbound session traffic to monitor for intellectual 1425 property leakage (by users, malware, and trojans), 1427 2. detect viruses/malware entering the network via email or web 1428 traffic, 1430 3. detect malware/Trojans in action, possibly connecting to remote 1431 hosts, 1433 4. detect attacks (Cross site scripting and other common web related 1434 attacks), 1436 5. track misuse and abuse by employees, 1438 6. restrict the types of protocols permitted to/from the entire 1439 corporate environment, 1441 7. detect and defend against Internet DDoS attacks, including both 1442 volumetric and layer 7 attacks. 1444 A significant portion of malware hides its activity within TLS or 1445 other encryption protocols. This includes lateral movement, Command 1446 and Control, and Data Exfiltration. 1448 The impact to a fully encrypted internal network would include cost 1449 and possible loss of detection capabilities associated with the 1450 transformation of the network architecture and tools for monitoring. 1451 The capabilities of detection through traffic fingerprinting, logs, 1452 host-level transaction monitoring, and flow analysis would vary 1453 depending on access to a 2-tuple or 5-tuple in the network as well. 1455 Security monitoring in the enterprise may also be performed at the 1456 endpoint with numerous current solutions that mitigate the same 1457 problems as some of the above mentioned solutions. Since the 1458 software agents operate on the device, they are able to monitor 1459 traffic before it is encrypted, monitor for behavior changes, and 1460 lock down devices to use only the expected set of applications. 1461 Session encryption does not affect these solutions. Some might argue 1462 that scaling is an issue in the enterprise, but some large 1463 enterprises have used these tools effectively. 1465 Use of Bring-your-own-device (BYOD) policies within organizations may 1466 limit the scope of monitoring permitted with these alternate 1467 solutions. Network endpoint assessment (NEA) or the use of virtual 1468 hosts could help to bridge the monitoring gap. 1470 4.1.2. Application Performance Monitoring in the Enterprise 1472 There are two main goals of monitoring: 1474 1. Assess traffic volume on a per-application basis, for billing, 1475 capacity planning, optimization of geographical location for 1476 servers or proxies, and other goals. 1478 2. Assess performance in terms of application response time and user 1479 perceived response time. 1481 Network-based Application Performance Monitoring tracks application 1482 response time by user and by URL, which is the information that the 1483 application owners and the lines of business request. Content 1484 Delivery Networks (CDNs) add complexity in determining the ultimate 1485 endpoint destination. By their very nature, such information is 1486 obscured by CDNs and encrypted protocols -- adding a new challenge 1487 for troubleshooting network and application problems. URL 1488 identification allows the application support team to do granular, 1489 code level troubleshooting at multiple tiers of an application. 1491 New methodologies to monitor user perceived response time and to 1492 separate network from server time are evolving. For example, the 1493 IPv6 Destination Option Header (DOH) implementation of Performance 1494 and Diagnostic Metrics (PDM) will provide this [RFC8250]. Using PDM 1495 with IPsec Encapsulating Security Payload (ESP) Transport Mode 1496 requires placement of the PDM DOH within the ESP encrypted payload to 1497 avoid leaking timing and sequence number information that could be 1498 useful to an attacker. Use of PDM DOH also may introduce some 1499 security weaknesses, including a timing attack, as described in 1500 Section 7 of [RFC8250]. For these and other reasons, [RFC8250] 1501 requires that the PDM DOH option be explicitly turned on by 1502 administrative action in each host where this measurement feature 1503 will be used. 1505 4.1.3. Enterprise Network Diagnostics and Troubleshooting 1507 One primary key to network troubleshooting is the ability to follow a 1508 transaction through the various tiers of an application in order to 1509 isolate the fault domain. A variety of factors relating to the 1510 structure of the modern data center and multi-tiered application have 1511 made it difficult to follow a transaction in network traces without 1512 the ability to examine some of the packet payload. Alternate 1513 methods, such as log analysis need improvement to fill this gap. 1515 4.1.3.1. Address Sharing (NAT) 1517 Content Delivery Networks (CDNs) and NATs and Network Address and 1518 Port Translators (NAPT) obscure the ultimate endpoint designation 1519 (See [RFC6269] for types of address sharing and a list of issues). 1520 Troubleshooting a problem for a specific end user requires finding 1521 information such as the IP address and other identifying information 1522 so that their problem can be resolved in a timely manner. 1524 NAT is also frequently used by lower layers of the data center 1525 infrastructure. Firewalls, Load Balancers, Web Servers, App Servers, 1526 and Middleware servers all regularly NAT the source IP of packets. 1527 Combine this with the fact that users are often allocated randomly by 1528 load balancers to all these devices, the network troubleshooter is 1529 often left with very few options in today's environment due to poor 1530 logging implementations in applications. As such, network 1531 troubleshooting is used to trace packets at a particular layer, 1532 decrypt them, and look at the payload to find a user session. 1534 This kind of bulk packet capture and bulk decryption is frequently 1535 used when troubleshooting a large and complex application. Endpoints 1536 typically don't have the capacity to handle this level of network 1537 packet capture, so out-of-band networks of robust packet brokers and 1538 network sniffers that use techniques such as copies of TLS RSA 1539 private keys accomplish this task today. 1541 4.1.3.2. TCP Pipelining/Session Multiplexing 1543 TCP pipelining/session multiplexing used mainly by middleboxes today 1544 allows for multiple end user sessions to share the same TCP 1545 connection. This raises several points of interest with an increased 1546 use of encryption. TCP session multiplexing should still be possible 1547 when TLS or TCPcrypt is in use since the TCP header information is 1548 exposed leaving the 5-tuple accessible. The use of TCP session 1549 multiplexing of an IP layer encryption, e.g. IPsec, that only 1550 exposes a 2-tuple would not be possible. Troubleshooting 1551 capabilities with encrypted sessions from the middlebox may limit 1552 troubleshooting to the use of logs from the end points performing the 1553 TCP multiplexing or from the middleboxes prior to any additional 1554 encryption that may be added to tunnel the TCP multiplexed traffic. 1556 Increased use of HTTP/2 will likely further increase the prevalence 1557 of session multiplexing, both on the Internet and in the private data 1558 center. HTTP pipelining requires both the client and server to 1559 participate; visibility of packets once encrypted will hide the use 1560 of HTTP pipelining for any monitoring that takes place outside of the 1561 endpoint or proxy solution. Since HTTP pipelining is between a 1562 client and server, logging capabilities may require improvement in 1563 some servers and clients for debugging purposes if this is not 1564 already possible. Visibility for middleboxes includes anything 1565 exposed by TLS and the 5-tuple. 1567 4.1.3.3. HTTP Service Calls 1569 When an application server makes an HTTP service call to back end 1570 services on behalf of a user session, it uses a completely different 1571 URL and a completely different TCP connection. Troubleshooting via 1572 network trace involves matching up the user request with the HTTP 1573 service call. Some organizations do this today by decrypting the TLS 1574 packet and inspecting the payload. Logging has not been adequate for 1575 their purposes. 1577 4.1.3.4. Application Layer Data 1579 Many applications use text formats such as XML to transport data or 1580 application level information. When transaction failures occur and 1581 the logs are inadequate to determine the cause, network and 1582 application teams work together, each having a different view of the 1583 transaction failure. Using this troubleshooting method, the network 1584 packet is correlated with the actual problem experienced by an 1585 application to find a root cause. The inability to access the 1586 payload prevents this method of troubleshooting. 1588 4.2. Techniques for Monitoring Internet Session Traffic 1590 Corporate networks commonly monitor outbound session traffic to 1591 detect or prevent attacks as well as to guarantee service level 1592 expectations. In some cases, alternate options are available when 1593 encryption is in use through a proxy or a shift to monitoring at the 1594 endpoint. In both cases, scaling is a concern and advancements to 1595 support this shift in monitoring practices will assist the deployment 1596 of end-to-end encryption. 1598 Some DLP tools intercept traffic at the Internet gateway or proxy 1599 services with the ability to man-in-the-middle (MiTM) encrypted 1600 session traffic (HTTP/TLS). These tools may monitor for key words 1601 important to the enterprise including business sensitive information 1602 such as trade secrets, financial data, personally identifiable 1603 information (PII), or personal health information (PHI). Various 1604 techniques are used to intercept HTTP/TLS sessions for DLP and other 1605 purposes, and can be misused as described in "Summarizing Known 1606 Attacks on TLS and DTLS" [RFC7457] Section 2.8. Note: many corporate 1607 policies allow access to personal financial and other sites for users 1608 without interception. Another option is to terminate a TLS session 1609 prior to the point where monitoring is performed. Aside from 1610 exposing user information to the enterprise MITM devices often are 1611 subject to severe security defects which can lead to exposure of user 1612 data to attackers outside the enterprise UserData [UserData]. In 1613 addition, implementation errors in middleboxes have led to major 1614 difficulties in deploying new versions of security protocols such as 1615 TLS [Ben17a][Ben17b][Res17a][Res17b] 1617 Monitoring traffic patterns for anomalous behavior such as increased 1618 flows of traffic that could be bursty at odd times or flows to 1619 unusual destinations (small or large amounts of traffic) is common. 1620 This traffic may or may not be encrypted and various methods of 1621 encryption or just obfuscation may be used. 1623 Web filtering devices are sometimes used to allow only access to 1624 well-known sites found to be legitimate and free of malware on last 1625 check by a web filtering service company. One common example of web 1626 filtering in a corporate environment is blocking access to sites that 1627 are not well-known to these tools for the purpose of blocking 1628 malware; this may be noticeable to those in research who are unable 1629 to access colleague's individual sites or new web sites that have not 1630 yet been screened. In situations where new sites are required for 1631 access, they can typically be added after notification by the user or 1632 log alerts and review. Home mail account access may be blocked in 1633 corporate settings to prevent another vector for malware to enter as 1634 well as for intellectual property to leak out of the network. This 1635 method remains functional with increased use of encryption and may be 1636 more effective at preventing malware from entering the network. Some 1637 enterprises may be more aggessive in their filtering and monitoring 1638 policy, causing undesirable outcomes. Web filtering solutions 1639 monitor and potentially restrict access based on the destination URL 1640 when available, server name, IP address, or the DNS name. A complete 1641 URL may be used in cases where access restrictions vary for content 1642 on a particular site or for the sites hosted on a particular server. 1643 In some cases, the enterprise may use a proxy to access this 1644 additional information based on their policy. This type of 1645 restriction is intended to be transparent to users in a corporate 1646 setting as the typical corporate user does not access sites which are 1647 not well-known to these tools. However, the mechanisms which these 1648 web filters use to do monitoring and enforcement have the potential 1649 to cause access issues or other user-visible failures. 1651 Desktop DLP tools are used in some corporate environments as well. 1652 Since these tools reside on the desktop, they can intercept traffic 1653 before it is encrypted and may provide a continued method of 1654 monitoring intellectual property leakage from the desktop to the 1655 Internet or attached devices. 1657 DLP tools can also be deployed by Network Service providers, as they 1658 have the vantage point of monitoring all traffic paired with 1659 destinations off the enterprise network. This makes an effective 1660 solution for enterprises that allow "bring-your-own" devices when the 1661 traffic is not encrypted, and for devices outside the desktop 1662 category (such as mobile phones) that are used on corporate networks 1663 nonetheless. 1665 Enterprises may wish to reduce the traffic on their Internet access 1666 facilities by monitoring requests for within-policy content and 1667 caching it. In this case, repeated requests for Internet content 1668 spawned by URLs in e-mail trade newsletters or other sources can be 1669 served within the enterprise network. Gradual deployment of end to 1670 end encryption would tend to reduce the cacheable content over time, 1671 owing to concealment of critical headers and payloads. Many forms of 1672 enterprise performance management may be similarly affected. It 1673 should be noted that transparent caching is considered an anti- 1674 pattern. 1676 5. Security Monitoring for Specific Attack Types 1678 Effective incident response today requires collaboration at Internet 1679 scale. This section will only focus on efforts of collaboration at 1680 Internet scale that are dedicated to specific attack types. They may 1681 require new monitoring and detection techniques in an increasingly 1682 encrypted Internet. As mentioned previously, some service providers 1683 have been interfering with STARTTLS to prevent session encryption to 1684 be able to perform functions they are used to (injecting ads, 1685 monitoring, etc.). By detailing the current monitoring methods used 1686 for attack detection and response, this information can be used to 1687 devise new monitoring methods that will be effective in the changed 1688 Internet via collaboration and innovation. 1690 Changes to improve encryption or to deploy OS methods have little 1691 impact on the detection of malicious actors. Malicious actors have 1692 had access to strong encryption for quite some time. Incident 1693 responders, in many cases, have developed techniques to locate 1694 malicious traffic within encrypted sessions. The following section 1695 will note some examples where detection and mitigation of such 1696 traffic has been successful. 1698 5.1. Mail Abuse and spam 1700 The largest operational effort to prevent mail abuse is through the 1701 Messaging, Malware, Mobile Anti-Abuse Working Group (M3AAWG)[M3AAWG]. 1702 Mail abuse is combatted directly with mail administrators who can 1703 shut down or stop continued mail abuse originating from large scale 1704 providers that participate in using the Abuse Reporting Format (ARF) 1705 agents standardized in the IETF [RFC5965], [RFC6430], [RFC6590], 1706 [RFC6591], [RFC6650], [RFC6651], and [RFC6652]. The ARF agent 1707 directly reports abuse messages to the appropriate service provider 1708 who can take action to stop or mitigate the abuse. Since this 1709 technique uses the actual message, the use of SMTP over TLS between 1710 mail gateways will not affect its usefulness. As mentioned 1711 previously, SMTP over TLS only protects data while in transit and the 1712 messages may be exposed on mail servers or mail gateways if a user- 1713 to-user encryption method is not used. Current user-to-user message 1714 encryption methods on email (S/MIME and PGP) do not encrypt the email 1715 header information used by ARF and the service provider operators in 1716 their abuse mitigation efforts. 1718 Another effort, Domain-based Message Authentication, Reporting, and 1719 Conformance (DMARC) [RFC7489] is a mechanism for policy distribution 1720 that enables increasingly strict handling of messages that fail 1721 authentication checks, ranging from no action, through altered 1722 delivery, up to message rejection. DMARC is also not affected by the 1723 use of STARTTLS. 1725 5.2. Denial of Service 1727 Response to Denial of Service (DoS) attacks are typically coordinated 1728 by the SP community with a few key vendors who have tools to assist 1729 in the mitigation efforts. Traffic patterns are determined from each 1730 DoS attack to stop or rate limit the traffic flows with patterns 1731 unique to that DoS attack. 1733 Data types used in monitoring traffic for DDoS are described in the 1734 DDoS Open Threat Signaling (DOTS) [DOTS] working group documents in 1735 development. The impact of encryption can be understood from their 1736 documented use cases[I-D.ietf-dots-use-cases]. 1738 Data types used in DDoS attacks have been detailed in the IODEF 1739 Guidance draft [RFC8274], Appendix A.2, with the help of several 1740 members of the service provider community. The examples provided are 1741 intended to help identify the useful data in detecting and mitigating 1742 these attacks independent of the transport and protocol descriptions 1743 in the drafts. 1745 5.3. Phishing 1747 Investigations and response to phishing attacks follow well-known 1748 patterns, requiring access to specific fields in email headers as 1749 well as content from the body of the message. When reporting 1750 phishing attacks, the recipient has access to each field as well as 1751 the body to make content reporting possible, even when end-to-end 1752 encryption is used. The email header information is useful to 1753 identify the mail servers and accounts used to generate or relay the 1754 attack messages in order to take the appropriate actions. The 1755 content of the message often contains an embedded attack that may be 1756 in an infected file or may be a link that results in the download of 1757 malware to the user's system. 1759 Administrators often find it helpful to use header information to 1760 track down similar message in their mail queue or users inboxes to 1761 prevent further infection. Combinations of To:, From:, Subject:, 1762 Received: from header information might be used for this purpose. 1763 Administrators may also search for document attachments of the same 1764 name, size, or containing a file with a matching hash to a known 1765 phishing attack. Administrators might also add URLs contained in 1766 messages to block lists locally or this may also be done by browser 1767 vendors through larger scales efforts like that of the Anti-Phishing 1768 Working Group (APWG). See the Coordinating Attack Response at 1769 Internet Scale (CARIS) workshop Report [RFC8073] for additional 1770 information and pointers to the APWG's efforts on anti- phishing. 1772 A full list of the fields used in phishing attack incident response 1773 can be found in RFC5901. Future plans to increase privacy 1774 protections may limit some of these capabilities if some email header 1775 fields are encrypted, such as To:, From:, and Subject: header fields. 1776 This does not mean that those fields should not be encrypted, only 1777 that we should be aware of how they are currently used. 1779 Some products protect users from phishing by maintaining lists of 1780 known phishing domains (such as misspelled bank names) and blocking 1781 access. This can be done by observing DNS, clear-text HTTP, or 1782 server name indication (SNI) in TLS, in addition to analyzing email. 1783 Alternate options to detect and prevent phishing attacks may be 1784 needed. More recent examples of data exchanged in spear phishing 1785 attacks has been detailed in the IODEF Guidance draft [RFC8274], 1786 Appendix A.3. 1788 5.4. Botnets 1790 Botnet detection and mitigation is complex as botnets may involve 1791 hundreds or thousands of hosts with numerous Command and Control 1792 (C&C) servers. The techniques and data used to monitor and detect 1793 each may vary. Connections to C&C servers are typically encrypted, 1794 therefore a move to an increasingly encrypted Internet may not affect 1795 the detection and sharing methods used. 1797 5.5. Malware 1799 Malware monitoring and detection techniques vary. As mentioned in 1800 the enterprise section, malware monitoring may occur at gateways to 1801 the organization analyzing email and web traffic. These services can 1802 also be provided by service providers, changing the scale and 1803 location of this type of monitoring. Additionally, incident 1804 responders may identify attributes unique to types of malware to help 1805 track down instances by their communication patterns on the Internet 1806 or by alterations to hosts and servers. 1808 Data types used in malware investigations have been summarized in an 1809 example of the IODEF Guidance draft [RFC8274], Appendix A.1. 1811 5.6. Spoofed Source IP Address Protection 1813 The IETF has reacted to spoofed source IP address-based attacks, 1814 recommending the use of network ingress filtering BCP 38 [RFC2827] 1815 and the unicast Reverse Path Forwarding (uRPF) mechanism [RFC2504]. 1816 But uRPF suffers from limitations regarding its granularity: a 1817 malicious node can still use a spoofed IP address included inside the 1818 prefix assigned to its link. The Source Address Validation 1819 Improvements (SAVI) mechanisms try to solve this issue. Basically, a 1820 SAVI mechanism is based on the monitoring of a specific address 1821 assignment/management protocol (e.g., SLAAC [RFC4862], SEND 1822 [RFC3971], DHCPv4/v6 [RFC2131][RFC3315]) and, according to this 1823 monitoring, set-up a filtering policy allowing only the IP flows with 1824 a correct source IP address (i.e., any packet with a source IP 1825 address, from a node not owning it, is dropped). The encryption of 1826 parts of the address assignment/management protocols, critical for 1827 SAVI mechanisms, can result in a dysfunction of the SAVI mechanisms. 1829 5.7. Further work 1831 Although incident response work will continue, new methods to prevent 1832 system compromise through security automation and continuous 1833 monitoring [SACM] may provide alternate approaches where system 1834 security is maintained as a preventative measure. 1836 6. Application-based Flow Information Visible to a Network 1838 This section describes specific techniques used in monitoring 1839 applications that is visible to the network if a 5-tuple is exposed 1840 and as such can potentially be used an input future network 1841 management approaches. It also includes an overview of IPFIX, a 1842 flow-based protocol used to export information about network flows. 1844 6.1. IP Flow Information Export 1846 Many of the accounting, monitoring and measurement tasks described in 1847 this document, especially Section 2.3.2, Section 3.1.1, 1848 Section 4.1.3, Section 4.2, and Section 5.2 use the IPFIX protocol 1849 [RFC7011] for export and storage of the monitored information. IPFIX 1850 evolved from the widely-deployed NetFlow protocol [RFC3954], which 1851 exports information about flows identified by 5-tuple. While NetFlow 1852 was largely concerned with exporting per-flow byte and packet counts 1853 for accounting purposes, IPFIX's extensible information model 1854 [RFC7012] provides a variety of Information Elements (IEs) 1855 [IPFIX-IANA] for representing information above and below the 1856 traditional network layer flow information. Enterprise-specific IEs 1857 allow exporter vendors to define their own non-standard IEs, as well, 1858 and many of these are driven by header and payload inspection at the 1859 metering process. 1861 While the deployment of encryption has no direct effect on the use of 1862 IPFIX, certain defined IEs may become unavailable when the metering 1863 process observing the traffic cannot decrypt formerly cleartext 1864 information. For example, HTTPS renders HTTP header analysis 1865 impossible, so IEs derived from the header (e.g. httpContentType, 1866 httpUserAgent) cannot be exported. 1868 The collection of IPFIX data itself, of course, provides a point of 1869 centralization for potentially business- and privacy-critical 1870 information. The IPFIX File Format specification [RFC5655] 1871 recommends encryption for this data at rest, and the IP Flow 1872 Anonymization specification [RFC6235] defines a metadata format for 1873 describing the anonymization functions applied to an IPFIX dataset, 1874 if anonymization is employed for data sharing of IPFIX information 1875 between enterprises or network operators. 1877 6.2. TLS Server Name Indication 1879 When initiating the TLS handshake, the Client may provide an 1880 extension field (server_name) which indicates the server to which it 1881 is attempting a secure connection. TLS SNI was standardized in 2003 1882 to enable servers to present the "correct TLS certificate" to clients 1883 in a deployment of multiple virtual servers hosted by the same server 1884 infrastructure and IP-address. Although this is an optional 1885 extension, it is today supported by all modern browsers, web servers 1886 and developer libraries. Akamai [Nygren] reports that many of their 1887 customer see client TLS SNI usage over 99%. It should be noted that 1888 HTTP/2 introduces the Alt-SVC method for upgrading the connection 1889 from HTTP/1 to either unencrypted or encrypted HTTP/2. If the 1890 initial HTTP/1 request is unencrypted, the destination alternate 1891 service name can be identified before the communication is 1892 potentially upgraded to encrypted HTTP/2 transport. HTTP/2 requires 1893 the TLS implementation to support the Server Name Indication (SNI) 1894 extension (see section 9.2 of [RFC7540]). It is also worth noting 1895 that [RFC7838] "allows an origin server to nominate additional means 1896 of interacting with it on the network", while [RFC8164] allows for a 1897 URI to be accessed with HTTP/2 and TLS using Opportunistic Security 1898 (on an experimental basis). 1900 This information is only available if the client populates the Server 1901 Name Indication extension. Doing so is an optional part of the TLS 1902 standard and as stated above this has been implemented by all major 1903 browsers. Due to its optional nature, though, existing network 1904 filters that examine a TLS ClientHello for a SNI extension cannot 1905 expect to always find one. The SNI Encryption in TLS Through 1906 Tunneling [I-D.ietf-tls-sni-encryption] draft has been adopted by the 1907 TLS working group, which provides solutions to encrypt SNI. As such, 1908 there will be an option to encrypt SNI in future versions of TLS. 1909 The per-domain nature of SNI may not reveal the specific service or 1910 media type being accessed, especially where the domain is of a 1911 provider offering a range of email, video, Web pages etc. For 1912 example, certain blog or social network feeds may be deemed 'adult 1913 content', but the Server Name Indication will only indicate the 1914 server domain rather than a URL path. 1916 There are additional issues for identification of content using SNI: 1917 [RFC7540] includes connection coalescing, 1918 [I-D.ietf-httpbis-origin-frame] defines the ORIGIN frame, and the 1919 [I-D.bishop-httpbis-http2-additional-certs] proposal will increase 1920 the difficulty of passive monitoring. 1922 6.3. Application Layer Protocol Negotiation (ALPN) 1924 ALPN is a TLS extension which may be used to indicate the application 1925 protocol within the TLS session. This is likely to be of more value 1926 to the network where it indicates a protocol dedicated to a 1927 particular traffic type (such as video streaming) rather than a 1928 multi-use protocol. ALPN is used as part of HTTP/2 'h2', but will 1929 not indicate the traffic types which may make up streams within an 1930 HTTP/2 multiplex. ALPN is sent clear text in the ClientHello and the 1931 server returns it in Encrypted Extensions in TLS 1.3. 1933 6.4. Content Length, BitRate and Pacing 1935 The content length of encrypted traffic is effectively the same as 1936 that of the cleartext. Although block ciphers utilize padding, this 1937 makes a negligible difference. Bitrate and pacing are generally 1938 application specific, and do not change much when the content is 1939 encrypted. Multiplexed formats (such as HTTP/2 and QUIC) may however incorporate several application 1941 streams over one connection, which makes the bitrate/pacing no longer 1942 application-specific. Also, packet padding is available in HTTP/2, 1943 TLS 1.3, and many other protocols. Traffic analysis is made more 1944 difficult by such countermeasures. 1946 7. Effect of Encryption on Mobile Network Evolution 1948 Transport header encryption prevents the use of transit proxies in 1949 center of the network and the use of some edge proxies by preventing 1950 the proxies from taking action on the stream. It may be that the 1951 claimed benefits of such proxies could be achieved by end-to-end 1952 client and server optimizations, distribution using CDNs, plus the 1953 ability to continue connections across different access technologies 1954 (across dynamic user IP addresses). The following aspects should be 1955 considered in this approach: 1957 1. In a wireless mobile network, the delay and channel capacity per 1958 user and sector varies due to coverage, contention, user 1959 mobility, scheduling balances fairness, capacity, and service 1960 QoE. If most users are at the cell edge, the controller cannot 1961 use more complex QAM, thus reducing total cell capacity; 1962 similarly if a UMTS edge is serving some number of CS-Voice 1963 Calls, the remaining capacity for packet services is reduced. 1965 2. Mobile wireless networks service in-bound roamers (Users of 1966 Operator A in a foreign operator Network B) by backhauling their 1967 traffic though Operator B's network to Operator A's Network and 1968 then serving through the P-Gateway (PGW), General GPRS Support 1969 Node (GGSN), Content Distribution Network (CDN) etc., of Operator 1970 A (User's Home Operator). Increasing window sizes to compensate 1971 for the path RTT will have the limitations outlined earlier for 1972 TCP. The outbound roamer scenario has a similar TCP performance 1973 impact. 1975 3. Issues in deploying CDNs in Radio Access Networks (RAN) include 1976 decreasing client-server control loop that requires deploying 1977 CDNs/Cloud functions that terminate encryption closer to the 1978 edge. In Cellular RAN, the user IP traffic is encapsulated into 1979 General Packet Radio Service (GPRS) Tunneling Protocol-User Plane 1980 (GTP-U in UMTS and LTE) tunnels to handle user mobility; the 1981 tunnels terminate in APN/GGSN/PGW that are in central locations. 1982 One user's traffic may flow through one or more APN's (for 1983 example Internet APN, Roaming APN for Operator X, Video-Service 1984 APN, OnDeckAPN etc.). The scope of operator private IP addresses 1985 may be limited to specific APNs. Since CDNs generally operate on 1986 user IP flows, deploying them would require enhancing them with 1987 tunnel translation, tunnel management functions etc.. 1989 4. While CDNs that de-encrypt flows or split-connection proxy 1990 (similar to split-tcp) could be deployed closer to the edges to 1991 reduce control loop RTT, with transport header encryption, such 1992 CDNs perform optimization functions only for partner client 1993 flows. Therefore, content from some Small-Medium Businesses 1994 (SMBs) would not get such CDN benefits. 1996 8. Response to Increased Encryption and Looking Forward 1998 As stated in [RFC7258], "an appropriate balance (between network 1999 management and PM mitigations) will emerge over time as real 2000 instances of this tension are considered." Numerous operators made 2001 it clear in their response to this document that they fully support 2002 strong encryption and providing privacy for end users, this is a 2003 common goal. Operators recognize not all the practices documented 2004 need to be supported going forward, either because of the risk to end 2005 user privacy or alternate technologies and tools have already 2006 emerged. This document is intended to support network engineers and 2007 other innovators to work toward solving network and security 2008 management problems with protocol designers and application 2009 developers in new ways that facilitate adoption of strong encryption 2010 rather than preventing the use of encryption. By having the 2011 discussions on network and security management practices with 2012 application developers and protocol designers, each side of the 2013 debate can understand each others goals, work toward alternate 2014 solutions, and disband with practices that should no longer be 2015 supported. A goal of this document is to assist the IETF to 2016 understand some of the current practices so as to identify new work 2017 items for IETF-related use cases which can help facilitate the 2018 adoption of strong session encryption and support network and 2019 security management. 2021 9. Security Considerations 2023 There are no additional security considerations as this is a summary 2024 and does not include a new protocol or functionality. 2026 10. IANA Considerations 2028 This memo makes no requests of IANA. 2030 11. Acknowledgements 2032 Thanks to our reviewers, Natasha Rooney, Kevin Smith, Ashutosh Dutta, 2033 Brandon Williams, Jean-Michel Combes, Nalini Elkins, Paul Barrett, 2034 Badri Subramanyan, Igor Lubashev, Suresh Krishnan, Dave Dolson, 2035 Mohamed Boucadair, Stephen Farrell, Warren Kumari, Alia Atlas, Roman 2036 Danyliw, Mirja Kuhlewind, Ines Robles, Joe Clarke, Kyle Rose, 2037 Christian Huitema, and Chris Morrow for their editorial and content 2038 suggestions. Surya K. Kovvali provided material for section 7. 2039 Chris Morrow and Nik Teague provided reviews and updates specific to 2040 the DoS fingerprinting text. Brian Trammell provided the IPFIX text. 2042 12. Informative References 2044 [ACCORD] "Acord BoF IETF95 2045 https://www.ietf.org/proceedings/95/accord.html". 2047 [CAIDA] "CAIDA *Anonymized Internet Traces* 2048 [http://www.caida.org/data/overview/ and 2049 http://www.caida.org/data/passive/ 2050 passive_2016_dataset.xml]". 2052 [DarkMail] 2053 "The Dark Mail Technical Aliance https://darkmail.info/". 2055 [DOTS] https://datatracker.ietf.org/wg/dots/charter/, "DDoS Open 2056 Threat Signaling IETF Working Group". 2058 [EFF2014] "EFF Report on STARTTLS Downgrade Attacks 2059 https://www.eff.org/deeplinks/2014/11/ 2060 starttls-downgrade-attacks". 2062 [Enrich] Narseo Vallina-Rodriguez, et al., "Header Enrichment or 2063 ISP Enrichment, Emerging Privacy Threats in Mobile 2064 Networks, Hot Middlebox, August 17-21 2015, London, United 2065 Kingdom", 2015. 2067 [I-D.bishop-httpbis-http2-additional-certs] 2068 Bishop, M., Sullivan, N., and M. Thomson, "Secondary 2069 Certificate Authentication in HTTP/2", draft-bishop- 2070 httpbis-http2-additional-certs-05 (work in progress), 2071 October 2017. 2073 [I-D.dolson-plus-middlebox-benefits] 2074 Dolson, D., Snellman, J., Boucadair, M., and C. Jacquenet, 2075 "Beneficial Functions of Middleboxes", draft-dolson-plus- 2076 middlebox-benefits-03 (work in progress), March 2017. 2078 [I-D.ietf-dots-use-cases] 2079 Dobbins, R., Migault, D., Fouant, S., Moskowitz, R., 2080 Teague, N., Xia, L., and K. Nishizuka, "Use cases for DDoS 2081 Open Threat Signaling", draft-ietf-dots-use-cases-09 (work 2082 in progress), November 2017. 2084 [I-D.ietf-httpbis-origin-frame] 2085 Nottingham, M. and E. Nygren, "The ORIGIN HTTP/2 Frame", 2086 draft-ietf-httpbis-origin-frame-06 (work in progress), 2087 January 2018. 2089 [I-D.ietf-tls-sni-encryption] 2090 Huitema, C. and E. Rescorla, "SNI Encryption in TLS 2091 Through Tunneling", draft-ietf-tls-sni-encryption-00 (work 2092 in progress), August 2017. 2094 [I-D.mglt-nvo3-geneve-security-requirements] 2095 Migault, D., Boutros, S., Wing, D., and S. Krishnan, 2096 "Geneve Protocol Security Requirements", draft-mglt-nvo3- 2097 geneve-security-requirements-03 (work in progress), 2098 February 2018. 2100 [IPFIX-IANA] 2101 "IP Flow Information Export (IPFIX) Entities 2102 https://www.iana.org/assignments/ipfix/". 2104 [JNSLP] Surveillance, Vol. 8 No. 3, "10 Standards for Oversight 2105 and Transparency of National Intelligence Services 2106 http://jnslp.com/". 2108 [M3AAWG] "Messaging, Malware, Mobile Anti-Abuse Working Group 2109 (M3AAWG) https://www.maawg.org/". 2111 [Nygren] https://blogs.akamai.com/2017/03/ reaching-toward- 2112 universal-tls-sni.html, "Erik Nygren, personal reference". 2114 [QUIC] https://datatracker.ietf.org/wg/quic/charter/, "QUIC 2115 (quic)". 2117 [RFC1945] Berners-Lee, T., Fielding, R., and H. Frystyk, "Hypertext 2118 Transfer Protocol -- HTTP/1.0", RFC 1945, 2119 DOI 10.17487/RFC1945, May 1996, 2120 . 2122 [RFC1958] Carpenter, B., Ed., "Architectural Principles of the 2123 Internet", RFC 1958, DOI 10.17487/RFC1958, June 1996, 2124 . 2126 [RFC1984] IAB and IESG, "IAB and IESG Statement on Cryptographic 2127 Technology and the Internet", BCP 200, RFC 1984, 2128 DOI 10.17487/RFC1984, August 1996, 2129 . 2131 [RFC2131] Droms, R., "Dynamic Host Configuration Protocol", 2132 RFC 2131, DOI 10.17487/RFC2131, March 1997, 2133 . 2135 [RFC2474] Nichols, K., Blake, S., Baker, F., and D. Black, 2136 "Definition of the Differentiated Services Field (DS 2137 Field) in the IPv4 and IPv6 Headers", RFC 2474, 2138 DOI 10.17487/RFC2474, December 1998, 2139 . 2141 [RFC2504] Guttman, E., Leong, L., and G. Malkin, "Users' Security 2142 Handbook", FYI 34, RFC 2504, DOI 10.17487/RFC2504, 2143 February 1999, . 2145 [RFC2663] Srisuresh, P. and M. Holdrege, "IP Network Address 2146 Translator (NAT) Terminology and Considerations", 2147 RFC 2663, DOI 10.17487/RFC2663, August 1999, 2148 . 2150 [RFC2775] Carpenter, B., "Internet Transparency", RFC 2775, 2151 DOI 10.17487/RFC2775, February 2000, 2152 . 2154 [RFC2804] IAB and IESG, "IETF Policy on Wiretapping", RFC 2804, 2155 DOI 10.17487/RFC2804, May 2000, 2156 . 2158 [RFC2827] Ferguson, P. and D. Senie, "Network Ingress Filtering: 2159 Defeating Denial of Service Attacks which employ IP Source 2160 Address Spoofing", BCP 38, RFC 2827, DOI 10.17487/RFC2827, 2161 May 2000, . 2163 [RFC3135] Border, J., Kojo, M., Griner, J., Montenegro, G., and Z. 2164 Shelby, "Performance Enhancing Proxies Intended to 2165 Mitigate Link-Related Degradations", RFC 3135, 2166 DOI 10.17487/RFC3135, June 2001, 2167 . 2169 [RFC3315] Droms, R., Ed., Bound, J., Volz, B., Lemon, T., Perkins, 2170 C., and M. Carney, "Dynamic Host Configuration Protocol 2171 for IPv6 (DHCPv6)", RFC 3315, DOI 10.17487/RFC3315, July 2172 2003, . 2174 [RFC3550] Schulzrinne, H., Casner, S., Frederick, R., and V. 2175 Jacobson, "RTP: A Transport Protocol for Real-Time 2176 Applications", STD 64, RFC 3550, DOI 10.17487/RFC3550, 2177 July 2003, . 2179 [RFC3724] Kempf, J., Ed., Austein, R., Ed., and IAB, "The Rise of 2180 the Middle and the Future of End-to-End: Reflections on 2181 the Evolution of the Internet Architecture", RFC 3724, 2182 DOI 10.17487/RFC3724, March 2004, 2183 . 2185 [RFC3954] Claise, B., Ed., "Cisco Systems NetFlow Services Export 2186 Version 9", RFC 3954, DOI 10.17487/RFC3954, October 2004, 2187 . 2189 [RFC3971] Arkko, J., Ed., Kempf, J., Zill, B., and P. Nikander, 2190 "SEcure Neighbor Discovery (SEND)", RFC 3971, 2191 DOI 10.17487/RFC3971, March 2005, 2192 . 2194 [RFC4787] Audet, F., Ed. and C. Jennings, "Network Address 2195 Translation (NAT) Behavioral Requirements for Unicast 2196 UDP", BCP 127, RFC 4787, DOI 10.17487/RFC4787, January 2197 2007, . 2199 [RFC4862] Thomson, S., Narten, T., and T. Jinmei, "IPv6 Stateless 2200 Address Autoconfiguration", RFC 4862, 2201 DOI 10.17487/RFC4862, September 2007, 2202 . 2204 [RFC5655] Trammell, B., Boschi, E., Mark, L., Zseby, T., and A. 2205 Wagner, "Specification of the IP Flow Information Export 2206 (IPFIX) File Format", RFC 5655, DOI 10.17487/RFC5655, 2207 October 2009, . 2209 [RFC5965] Shafranovich, Y., Levine, J., and M. Kucherawy, "An 2210 Extensible Format for Email Feedback Reports", RFC 5965, 2211 DOI 10.17487/RFC5965, August 2010, 2212 . 2214 [RFC6108] Chung, C., Kasyanov, A., Livingood, J., Mody, N., and B. 2215 Van Lieu, "Comcast's Web Notification System Design", 2216 RFC 6108, DOI 10.17487/RFC6108, February 2011, 2217 . 2219 [RFC6235] Boschi, E. and B. Trammell, "IP Flow Anonymization 2220 Support", RFC 6235, DOI 10.17487/RFC6235, May 2011, 2221 . 2223 [RFC6269] Ford, M., Ed., Boucadair, M., Durand, A., Levis, P., and 2224 P. Roberts, "Issues with IP Address Sharing", RFC 6269, 2225 DOI 10.17487/RFC6269, June 2011, 2226 . 2228 [RFC6430] Li, K. and B. Leiba, "Email Feedback Report Type Value: 2229 not-spam", RFC 6430, DOI 10.17487/RFC6430, November 2011, 2230 . 2232 [RFC6455] Fette, I. and A. Melnikov, "The WebSocket Protocol", 2233 RFC 6455, DOI 10.17487/RFC6455, December 2011, 2234 . 2236 [RFC6590] Falk, J., Ed. and M. Kucherawy, Ed., "Redaction of 2237 Potentially Sensitive Data from Mail Abuse Reports", 2238 RFC 6590, DOI 10.17487/RFC6590, April 2012, 2239 . 2241 [RFC6591] Fontana, H., "Authentication Failure Reporting Using the 2242 Abuse Reporting Format", RFC 6591, DOI 10.17487/RFC6591, 2243 April 2012, . 2245 [RFC6650] Falk, J. and M. Kucherawy, Ed., "Creation and Use of Email 2246 Feedback Reports: An Applicability Statement for the Abuse 2247 Reporting Format (ARF)", RFC 6650, DOI 10.17487/RFC6650, 2248 June 2012, . 2250 [RFC6651] Kucherawy, M., "Extensions to DomainKeys Identified Mail 2251 (DKIM) for Failure Reporting", RFC 6651, 2252 DOI 10.17487/RFC6651, June 2012, 2253 . 2255 [RFC6652] Kitterman, S., "Sender Policy Framework (SPF) 2256 Authentication Failure Reporting Using the Abuse Reporting 2257 Format", RFC 6652, DOI 10.17487/RFC6652, June 2012, 2258 . 2260 [RFC7011] Claise, B., Ed., Trammell, B., Ed., and P. Aitken, 2261 "Specification of the IP Flow Information Export (IPFIX) 2262 Protocol for the Exchange of Flow Information", STD 77, 2263 RFC 7011, DOI 10.17487/RFC7011, September 2013, 2264 . 2266 [RFC7012] Claise, B., Ed. and B. Trammell, Ed., "Information Model 2267 for IP Flow Information Export (IPFIX)", RFC 7012, 2268 DOI 10.17487/RFC7012, September 2013, 2269 . 2271 [RFC7143] Chadalapaka, M., Satran, J., Meth, K., and D. Black, 2272 "Internet Small Computer System Interface (iSCSI) Protocol 2273 (Consolidated)", RFC 7143, DOI 10.17487/RFC7143, April 2274 2014, . 2276 [RFC7146] Black, D. and P. Koning, "Securing Block Storage Protocols 2277 over IP: RFC 3723 Requirements Update for IPsec v3", 2278 RFC 7146, DOI 10.17487/RFC7146, April 2014, 2279 . 2281 [RFC7230] Fielding, R., Ed. and J. Reschke, Ed., "Hypertext Transfer 2282 Protocol (HTTP/1.1): Message Syntax and Routing", 2283 RFC 7230, DOI 10.17487/RFC7230, June 2014, 2284 . 2286 [RFC7234] Fielding, R., Ed., Nottingham, M., Ed., and J. Reschke, 2287 Ed., "Hypertext Transfer Protocol (HTTP/1.1): Caching", 2288 RFC 7234, DOI 10.17487/RFC7234, June 2014, 2289 . 2291 [RFC7258] Farrell, S. and H. Tschofenig, "Pervasive Monitoring Is an 2292 Attack", BCP 188, RFC 7258, DOI 10.17487/RFC7258, May 2293 2014, . 2295 [RFC7348] Mahalingam, M., Dutt, D., Duda, K., Agarwal, P., Kreeger, 2296 L., Sridhar, T., Bursell, M., and C. Wright, "Virtual 2297 eXtensible Local Area Network (VXLAN): A Framework for 2298 Overlaying Virtualized Layer 2 Networks over Layer 3 2299 Networks", RFC 7348, DOI 10.17487/RFC7348, August 2014, 2300 . 2302 [RFC7435] Dukhovni, V., "Opportunistic Security: Some Protection 2303 Most of the Time", RFC 7435, DOI 10.17487/RFC7435, 2304 December 2014, . 2306 [RFC7457] Sheffer, Y., Holz, R., and P. Saint-Andre, "Summarizing 2307 Known Attacks on Transport Layer Security (TLS) and 2308 Datagram TLS (DTLS)", RFC 7457, DOI 10.17487/RFC7457, 2309 February 2015, . 2311 [RFC7489] Kucherawy, M., Ed. and E. Zwicky, Ed., "Domain-based 2312 Message Authentication, Reporting, and Conformance 2313 (DMARC)", RFC 7489, DOI 10.17487/RFC7489, March 2015, 2314 . 2316 [RFC7498] Quinn, P., Ed. and T. Nadeau, Ed., "Problem Statement for 2317 Service Function Chaining", RFC 7498, 2318 DOI 10.17487/RFC7498, April 2015, 2319 . 2321 [RFC7525] Sheffer, Y., Holz, R., and P. Saint-Andre, 2322 "Recommendations for Secure Use of Transport Layer 2323 Security (TLS) and Datagram Transport Layer Security 2324 (DTLS)", BCP 195, RFC 7525, DOI 10.17487/RFC7525, May 2325 2015, . 2327 [RFC7540] Belshe, M., Peon, R., and M. Thomson, Ed., "Hypertext 2328 Transfer Protocol Version 2 (HTTP/2)", RFC 7540, 2329 DOI 10.17487/RFC7540, May 2015, 2330 . 2332 [RFC7619] Smyslov, V. and P. Wouters, "The NULL Authentication 2333 Method in the Internet Key Exchange Protocol Version 2 2334 (IKEv2)", RFC 7619, DOI 10.17487/RFC7619, August 2015, 2335 . 2337 [RFC7624] Barnes, R., Schneier, B., Jennings, C., Hardie, T., 2338 Trammell, B., Huitema, C., and D. Borkmann, 2339 "Confidentiality in the Face of Pervasive Surveillance: A 2340 Threat Model and Problem Statement", RFC 7624, 2341 DOI 10.17487/RFC7624, August 2015, 2342 . 2344 [RFC7665] Halpern, J., Ed. and C. Pignataro, Ed., "Service Function 2345 Chaining (SFC) Architecture", RFC 7665, 2346 DOI 10.17487/RFC7665, October 2015, 2347 . 2349 [RFC7754] Barnes, R., Cooper, A., Kolkman, O., Thaler, D., and E. 2350 Nordmark, "Technical Considerations for Internet Service 2351 Blocking and Filtering", RFC 7754, DOI 10.17487/RFC7754, 2352 March 2016, . 2354 [RFC7799] Morton, A., "Active and Passive Metrics and Methods (with 2355 Hybrid Types In-Between)", RFC 7799, DOI 10.17487/RFC7799, 2356 May 2016, . 2358 [RFC7826] Schulzrinne, H., Rao, A., Lanphier, R., Westerlund, M., 2359 and M. Stiemerling, Ed., "Real-Time Streaming Protocol 2360 Version 2.0", RFC 7826, DOI 10.17487/RFC7826, December 2361 2016, . 2363 [RFC7838] Nottingham, M., McManus, P., and J. Reschke, "HTTP 2364 Alternative Services", RFC 7838, DOI 10.17487/RFC7838, 2365 April 2016, . 2367 [RFC7858] Hu, Z., Zhu, L., Heidemann, J., Mankin, A., Wessels, D., 2368 and P. Hoffman, "Specification for DNS over Transport 2369 Layer Security (TLS)", RFC 7858, DOI 10.17487/RFC7858, May 2370 2016, . 2372 [RFC8073] Moriarty, K. and M. Ford, "Coordinating Attack Response at 2373 Internet Scale (CARIS) Workshop Report", RFC 8073, 2374 DOI 10.17487/RFC8073, March 2017, 2375 . 2377 [RFC8164] Nottingham, M. and M. Thomson, "Opportunistic Security for 2378 HTTP/2", RFC 8164, DOI 10.17487/RFC8164, May 2017, 2379 . 2381 [RFC8165] Hardie, T., "Design Considerations for Metadata 2382 Insertion", RFC 8165, DOI 10.17487/RFC8165, May 2017, 2383 . 2385 [RFC8250] Elkins, N., Hamilton, R., and M. Ackermann, "IPv6 2386 Performance and Diagnostic Metrics (PDM) Destination 2387 Option", RFC 8250, DOI 10.17487/RFC8250, September 2017, 2388 . 2390 [RFC8274] Kampanakis, P. and M. Suzuki, "Incident Object Description 2391 Exchange Format Usage Guidance", RFC 8274, 2392 DOI 10.17487/RFC8274, November 2017, 2393 . 2395 [RFC8300] Quinn, P., Ed., Elzur, U., Ed., and C. Pignataro, Ed., 2396 "Network Service Header (NSH)", RFC 8300, 2397 DOI 10.17487/RFC8300, January 2018, 2398 . 2400 [SACM] https://datatracker.ietf.org/wg/sacm/charter/, "Security 2401 Automation and Continuous Monitoring (sacm) IETF Working 2402 Group". 2404 [Snowden] http://www.jjsylvia.com/bigdatacourse/wp- 2405 content/uploads/2016/04/p14-verble-1.pdf, "The NSA and 2406 Edward Snowden: Surveillance In The 21st Century", 2014. 2408 [TCPcrypt] 2409 https://datatracker.ietf.org/wg/tcpinc/charter/, 2410 "TCPcrypt". 2412 [TLS100Proceedings] 2413 IETF 100, TLS Working Group Session, "Presentation before 2414 the TLS WG at IETF 2415 https://datatracker.ietf.org/meeting/100/materials/ 2416 slides-100-tls-sessa-tls13/", 2017. 2418 [TS3GPP] "3GPP TS 24.301, "Non-Access-Stratum (NAS) protocol for 2419 Evolved Packet System (EPS); Stage 3"", 2017. 2421 [UPCON] 3GPP, "User Plane Congestion Management 2422 http://www.3gpp.org/DynaReport/ 2423 FeatureOrStudyItemFile-570029.htm", 2014. 2425 [UserData] 2426 Network and Distributed Systems Symposium, The Internet 2427 Society, "The Security Impact of HTTPS Interception", 2428 2017. 2430 Authors' Addresses 2432 Kathleen Moriarty (editor) 2433 Dell EMC 2434 176 South St 2435 Hopkinton, MA 2436 USA 2438 Phone: +1 2439 Email: Kathleen.Moriarty@dell.com 2441 Al Morton (editor) 2442 AT&T Labs 2443 200 Laurel Avenue South 2444 Middletown,, NJ 07748 2445 USA 2447 Phone: +1 732 420 1571 2448 Fax: +1 732 368 1192 2449 Email: acmorton@att.com