idnits 2.17.1 draft-ietf-cdni-framework-07.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 : ---------------------------------------------------------------------------- == There are 79 instances of lines with non-RFC2606-compliant FQDNs in the document. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (November 25, 2013) is 3799 days in the past. Is this intentional? -- Found something which looks like a code comment -- if you have code sections in the document, please surround them with '' and '' lines. Checking references for intended status: Informational ---------------------------------------------------------------------------- == Missing Reference: 'RI REPLY' is mentioned on line 472, but not defined == Unused Reference: 'I-D.lefaucheur-cdni-logging-delivery' is defined on line 2381, but no explicit reference was found in the text == Unused Reference: 'I-D.seedorf-alto-for-cdni' is defined on line 2392, but no explicit reference was found in the text == Unused Reference: 'RFC3466' is defined on line 2407, but no explicit reference was found in the text == Unused Reference: 'RFC5693' is defined on line 2411, but no explicit reference was found in the text == Unused Reference: 'RFC6707' is defined on line 2415, but no explicit reference was found in the text == Unused Reference: 'RFC6770' is defined on line 2419, but no explicit reference was found in the text == Unused Reference: 'RFC6983' is defined on line 2423, but no explicit reference was found in the text == Outdated reference: A later version (-21) exists of draft-ietf-cdni-metadata-03 == Outdated reference: A later version (-17) exists of draft-ietf-cdni-requirements-12 -- Obsolete informational reference (is this intentional?): RFC 3466 (Obsoleted by RFC 7336) Summary: 0 errors (**), 0 flaws (~~), 12 warnings (==), 3 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group L. Peterson, Ed. 3 Internet-Draft Akamai Technologies, Inc. 4 Obsoletes: 3466 (if approved) B. Davie 5 Intended status: Informational VMware, Inc. 6 Expires: May 29, 2014 November 25, 2013 8 Framework for CDN Interconnection 9 draft-ietf-cdni-framework-07 11 Abstract 13 This document presents a framework for Content Distribution Network 14 Interconnection (CDNI). The purpose of the framework is to provide 15 an overall picture of the problem space of CDNI and to describe the 16 relationships among the various components necessary to interconnect 17 CDNs. CDN Interconnection requires the specification of interfaces 18 and mechanisms to address issues such as request routing, 19 distribution metadata exchange, and logging information exchange 20 across CDNs. The intent of this document is to outline what each 21 interface needs to accomplish, and to describe how these interfaces 22 and mechanisms fit together, while leaving their detailed 23 specification to other documents. It obsoletes RFC 3466. 25 Status of This Memo 27 This Internet-Draft is submitted in full conformance with the 28 provisions of BCP 78 and BCP 79. 30 Internet-Drafts are working documents of the Internet Engineering 31 Task Force (IETF). Note that other groups may also distribute 32 working documents as Internet-Drafts. The list of current Internet- 33 Drafts is at http://datatracker.ietf.org/drafts/current/. 35 Internet-Drafts are draft documents valid for a maximum of six months 36 and may be updated, replaced, or obsoleted by other documents at any 37 time. It is inappropriate to use Internet-Drafts as reference 38 material or to cite them other than as "work in progress." 40 This Internet-Draft will expire on May 29, 2014. 42 Copyright Notice 44 Copyright (c) 2013 IETF Trust and the persons identified as the 45 document authors. All rights reserved. 47 This document is subject to BCP 78 and the IETF Trust's Legal 48 Provisions Relating to IETF Documents 49 (http://trustee.ietf.org/license-info) in effect on the date of 50 publication of this document. Please review these documents 51 carefully, as they describe your rights and restrictions with respect 52 to this document. Code Components extracted from this document must 53 include Simplified BSD License text as described in Section 4.e of 54 the Trust Legal Provisions and are provided without warranty as 55 described in the Simplified BSD License. 57 Table of Contents 59 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 60 1.1. Terminology . . . . . . . . . . . . . . . . . . . . . . . 3 61 1.2. Reference Model . . . . . . . . . . . . . . . . . . . . . 4 62 1.3. Structure Of This Document . . . . . . . . . . . . . . . 7 63 2. Building Blocks . . . . . . . . . . . . . . . . . . . . . . . 8 64 2.1. Request Redirection . . . . . . . . . . . . . . . . . . . 8 65 2.1.1. DNS Redirection . . . . . . . . . . . . . . . . . . . 8 66 2.1.2. HTTP Redirection . . . . . . . . . . . . . . . . . . 9 67 3. Overview of CDNI Operation . . . . . . . . . . . . . . . . . 10 68 3.1. Preliminaries . . . . . . . . . . . . . . . . . . . . . . 12 69 3.2. Iterative HTTP Redirect Example . . . . . . . . . . . . . 13 70 3.3. Recursive HTTP Redirection Example . . . . . . . . . . . 18 71 3.4. Iterative DNS-based Redirection Example . . . . . . 22 72 3.5. Dynamic Footprint Discovery Example . . . . . . . . . . . 25 73 3.6. Content Removal Example . . . . . . . . . . . . . . . . . 26 74 3.7. Pre-Positioned Content Acquisition Example . . . . . . . 27 75 3.8. Asynchronous CDNI Metadata Example . . . . . . . . . . . 28 76 3.9. Synchronous CDNI Metadata Acquisition Example . . . . . . 30 77 3.10. Content and Metadata Acquisition with Multiple Upstream 78 CDNs . . . . . . . . . . . . . . . . . . . . . . . . . . 32 79 4. Main Interfaces . . . . . . . . . . . . . . . . . . . . . . . 33 80 4.1. In-Band versus Out-of-Band Interfaces . . . . . . . . . . 34 81 4.2. Cross Interface Concerns . . . . . . . . . . . . . . . . 34 82 4.3. Request Routing Interfaces . . . . . . . . . . . . . . . 35 83 4.4. CDNI Logging Interface . . . . . . . . . . . . . . . . . 36 84 4.5. CDNI Control Interface . . . . . . . . . . . . . . . . . 38 85 4.6. CDNI Metadata Interface . . . . . . . . . . . . . . . . . 38 86 4.7. HTTP Adaptive Streaming Concerns . . . . . . . . . . . . 39 87 4.8. URI Rewriting . . . . . . . . . . . . . . . . . . . . . . 40 88 5. Deployment Models . . . . . . . . . . . . . . . . . . . . . . 41 89 5.1. Meshed CDNs . . . . . . . . . . . . . . . . . . . . . . . 42 90 5.2. CSP combined with CDN . . . . . . . . . . . . . . . . . . 43 91 5.3. CSP using CDNI Request Routing Interface . . . . . . . . 43 92 5.4. CDN Federations and CDN Exchanges . . . . . . . . . . . . 44 93 6. Trust Model . . . . . . . . . . . . . . . . . . . . . . . . . 46 94 7. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 48 95 8. Security Considerations . . . . . . . . . . . . . . . . . . . 48 96 8.1. Security of CDNI Interfaces . . . . . . . . . . . . . . . 49 97 8.2. Digital Rights Management . . . . . . . . . . . . . . . . 49 98 9. Contributors . . . . . . . . . . . . . . . . . . . . . . . . 50 99 10. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 50 100 11. Informative References . . . . . . . . . . . . . . . . . . . 50 101 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 52 103 1. Introduction 105 This document provides an overview of the various components 106 necessary to interconnect CDNs, expanding on the problem statement 107 and use cases introduced in RFCs 6770 and 6707. It describes the 108 necessary interfaces and mechanisms in general terms and outlines how 109 they fit together to form a complete system for CDN Interconnection. 110 Detailed specifications are left to other documents. This document 111 makes extensive use of message flow examples to illustrate the 112 operation of interconnected CDNs, but these examples should be 113 considered illustrative rather than prescriptive. 115 RFC 3466 uses different terminology and models for "Content 116 Internetworking (CDI)". It is also less prescriptive in terms of 117 interfaces. To avoid confusion, this document obsoletes RFC 3466. 119 1.1. Terminology 121 This document uses the core terminology defined in RFC 6707. It also 122 introduces the following terms: 124 CDN-Domain: a host name (FQDN) at the beginning of a URL, 125 representing a set of content that is served by a given CDN. For 126 example, in the URL http://cdn.csp.com/...rest of url..., the CDN 127 domain is cdn.csp.com. A major role of CDN-Domain is to identify a 128 region (subset) of the URI space relative to which various CDN 129 Interconnection rules and policies are to apply. For example, a 130 record of CDN Metadata might be defined for the set of resources 131 corresponding to some CDN-Domain. 133 Distinguished CDN-Domain: a CDN-Domain that is allocated by a CDN for 134 the purposes of communication with a peer CDN, but which is not found 135 in client requests. Such CDN-Domains may be used for inter-CDN 136 acquisition, or as redirection targets, and enable a CDN to 137 distinguish a request from a peer CDN from an end-user request. 139 Delivering CDN: the CDN that ultimately delivers a piece of content 140 to the end-user. The last in a potential sequence of downstream 141 CDNs. 143 Recursive CDNI Request Redirection: When an upstream CDN elects to 144 redirect a request towards a downstream CDN, the upstream CDN can 145 query the downstream CDN Request Routing system via the CDNI Request 146 Routing Redirection Interface (or use information cached from earlier 147 similar queries) to find out how the downstream CDN wants the request 148 to be redirected, which allows the upstream CDN to factor in the 149 downstream CDN response when redirecting the user agent. This 150 approach is referred to as "Recursive" CDNI Request Redirection. 151 Note that the downstream CDN may elect to have the request redirected 152 directly to a Surrogate inside the downstream CDN, to the Request 153 Routing System of the downstream CDN, to another CDN, or to whatever 154 system is necessary to handle the redirected request appropriately. 156 Iterative CDNI Request Redirection: When an upstream CDN elects to 157 redirect a request towards a downstream CDN, the upstream CDN can 158 base its redirection purely on a local decision (and without 159 attempting to take into account how the downstream CDN may in turn 160 redirect the user agent). In that case, the upstream CDN redirects 161 the request to the request routing system in the downstream CDN, 162 which in turn will decide how to redirect that request: this approach 163 is referred to as "Iterative" CDNI Request Redirection. 165 Synchronous CDNI operations: operations between CDNs that happen 166 during the process of servicing a user request, i.e. between the time 167 that the user agent begins its attempt to obtain content and the time 168 at which that request is served. 170 Asynchronous CDNI operations: operations between CDNs that happen 171 independently of any given user request, such as advertisement of 172 footprint information or pre-positioning of content for later 173 delivery. 175 Trigger Interface: a subset of the CDNI Control interface that 176 includes operations to pre-position, revalidate, and purge both 177 metadata and content. These operations are typically called in 178 response to some action (Trigger) by the CSP on the upstream CDN. 180 We also sometimes use uCDN and dCDN as shorthand for upstream CDN and 181 downstream CDN, respectively. 183 1.2. Reference Model 185 This document uses the reference model in Figure 1, which expands the 186 reference model originally defined in RFC 6707. (The difference is 187 that the expanded model splits the Request Routing Interface into its 188 two distinct parts: the Request Routing Redirection interface and the 189 Footprint and Capabilities Advertisement interface, as described 190 below.) 191 -------- 192 / \ 193 | CSP | 194 \ / 195 -------- 196 * 197 * 198 * /\ 199 * / \ 200 ---------------------- |CDNI| ---------------------- 201 / Upstream CDN \ | | / Downstream CDN \ 202 | +-------------+ | | CI | | +-------------+ | 203 |******* Control |<======|====|=======>| Control *******| 204 |* +------*----*-+ | | | | +-*----*------+ *| 205 |* * * | | | | * * *| 206 |* +------*------+ | | LI | | +------*------+ *| 207 |* ***** Logging |<======|====|=======>| Logging ***** *| 208 |* * +-*-----------+ | | | | +-----------*-+ * *| 209 |* * * * | | | | * * * *| 210 .....*...+-*---------*-+ | | RI | | +-*---------*-+...*.*... 211 . |* * | |<======|====|=======>| | * *| . 212 . |* * | Req-Routing | | |FCI | | | Req-Routing | * *| . 213 . |* * *** |<======|====|=======>| |** * *| . 214 . |* * * +-------------+.| | | | +-------------+ * * *| . 215 . |* * * . | | | * * *| . 216 . |* * * +-------------+ |. | MI | | +-------------+ * * *| . 217 . |* * * | Distribution|<==.===|====|=======>| Distribution| * * *| . 218 . |* * * | | | . \ / | | | * * *| . 219 . |* * * |+---------+ | | . \/ | | +---------+| * * *| . 220 . |* * ***| +---------+| | ...Request......+---------+ |*** * *| . 221 . |* *****+-|Surrogate|***********************|Surrogate|-+***** *| . 222 . |******* +---------+| | Acquisition | |+----------+ *******| . 223 . | +-------------+ | | +-------*-----+ | . 224 . \ / \ * / . 225 . ---------------------- ---------*------------ . 226 . * . 227 . * Delivery . 228 . * . 229 . +--*---+ . 230 ...............Request............................| User |..Request.. 231 | Agent| 232 +------+ 234 <==> interfaces inside the scope of CDNI 236 **** and .... interfaces outside the scope of CDNI 238 Figure 1: CDNI Expanded Model and CDNI Interfaces 240 We note that while some interfaces in the reference model are "out of 241 scope" for the CDNI WG (in the sense that there is no need to define 242 new protocols for those interfaces) we still need to refer to them in 243 this document to explain the overall operation of CDNI. 245 We also note that, while we generally show only one upstream CDN 246 serving a given CSP, it is entirely possible that multiple uCDNs can 247 serve a single CSP. In fact, this situation effectively exists today 248 in the sense that a single CSP can currently delegate its content 249 delivery to more than one CDN. 251 The following briefly describes the five CDNI interfaces, 252 paraphrasing the definitions given in RFC 6707. We discuss these 253 interfaces in more detail in Section 4. 255 o CDNI Control interface (CI): Operations to bootstrap and 256 parameterize the other CDNI interfaces, as well as operations to 257 pre-position, revalidate, and purge both metadata and content. 258 The latter subset of operations is sometimes collectively called 259 the "Trigger interface." 261 o CDNI Request Routing interface: Operations to determine what CDN 262 (and optionally what surrogate within a CDN) is to serve end- 263 user's requests. Is actually a logical bundling of two separate 264 but related interfaces: 266 * CDNI Footprint & Capabilities Advertisement interface (FCI): 267 Asynchronous operations to exchange routing information (e.g., 268 the network footprint and capabilities served by a given CDN) 269 that enables CDN selection for subsequent user requests; and 271 * CDNI Request Routing Redirection interface (RI): Synchronous 272 operations to select a delivery CDN (surrogate) for a given 273 user request. 275 o CDNI Metadata interface (MI): Operations to communicate metadata 276 that governs how the content is delivered by interconnected CDNs. 277 Examples of CDNI metadata include geo-blocking directives, 278 availability windows, access control mechanisms, and purge 279 directives. It may include a combination of: 281 * Asynchronous operations to exchange metadata that govern 282 subsequent user requests for content; and 284 * Synchronous operations that govern behavior for a given user 285 request for content. 287 o CDNI Logging interface (LI): Operations that allow interconnected 288 CDNs to exchange relevant activity logs. May include a 289 combination of: 291 * Real-time exchanges, suitable for runtime traffic monitoring; 292 and 294 * Offline exchanges, suitable for analytics and billing. 296 There is some potential overlap between the set of Trigger-based 297 operations in the CDNI Control interface and the CDNI Metadata 298 interface. For both cases, the information passed from the upstream 299 CDN to the downstream CDN can broadly be viewed as metadata that 300 describes how content is to be managed by the downstream CDN. For 301 example, the information conveyed by CI to pre-position, revalidate 302 or purge metadata is similar to the information conveyed by posting 303 updated metadata via the MI. Even the CI operation to purge content 304 could be viewed as an metadata update for that content: purge simply 305 says that the availability window for the named content ends now. 306 The two interfaces share much in common, so minimally, there will 307 need to be a consistent data model that spans both. 309 The distinction we draw has to do with what the uCDN knows about the 310 successful application of the metadata by the dCDN. In the case of 311 the CI, the downstream CDN returning a successful status message 312 guarantees that the operation has been successfully completed; e.g., 313 the content has been purged or pre-positioned. This implies that the 314 downstream CDN accepts responsibility for having successfully 315 completed the requested operation. In contrast, metadata passed 316 between CDNs via the MI carries no such completion guarantee. 317 Returning success implies successful receipt of the metadata, but 318 nothing can be inferred about precisely when the metadata will take 319 effect in the downstream CDN, only that it will take effect 320 eventually. This is because of the challenge in globally 321 synchronizing updates to metadata with end-user requests that are 322 currently in progress (or indistinguishable from currently being in 323 progress). Clearly, a CDN will not be viewed as a trusted peer if 324 "eventually" often becomes an indefinite period of time, but the 325 acceptance of responsibility cannot be as crisply defined for the MI. 327 Finally, there is a practical issue that impacts all of the CNDI 328 interfaces, and that is whether or not to optimize CDNI for HTTP 329 Adaptive Streaming (HAS). We highlight specific issues related to 330 delivering HAS content throughout this document, but for a more 331 thorough treatment of the topic, see RFC 6983. 333 1.3. Structure Of This Document 334 The remainder of this document is organized as follows: 336 o Section 2 describes some essential building blocks for CDNI, 337 notably the various options for redirecting user requests to a 338 given CDN. 340 o Section 3 provides a number of illustrative examples of various 341 CDNI operations. 343 o Section 4 describes the functionality of the main CDNI interfaces. 345 o Section 5 shows how various deployment models of CDNI may be 346 achieved using the defined interfaces. 348 o Section 6 describes the trust model of CDNI and the issues of 349 transitive trust in particular that CDNI raises. 351 2. Building Blocks 353 2.1. Request Redirection 355 At its core, CDN Interconnection requires the redirection of requests 356 from one CDN to another. For any given request that is received by 357 an upstream CDN, it will either respond to the request directly, or 358 somehow redirect the request to a downstream CDN. Two main 359 mechanisms are available for redirecting a request to a downstream 360 CDN. The first leverages the DNS name resolution process and the 361 second uses in-protocol redirection mechanisms such as the HTTP 302 362 or 307 redirection response. We discuss these below as background 363 before discussing some examples of their use in Section 3. 365 2.1.1. DNS Redirection 367 DNS redirection is based on returning different IP addresses for the 368 same DNS name, for example, to balance server load or to account for 369 the client's location in the network. A DNS server, sometimes called 370 the Local DNS (LDNS), resolves DNS names on behalf of an end-user. 371 The LDNS server in turn queries other DNS servers until it reaches 372 the authoritative DNS server for the CDN-Domain. The network 373 operator typically provides the LDNS server, although the user is 374 free to choose other DNS servers (e.g., OpenDNS, Google Public DNS). 375 This latter possibility is important because the authoritative DNS 376 server sees only the IP address of the DNS server that queries it, 377 not the IP address of the original end-user. 379 The advantage of DNS redirection is that it is completely transparent 380 to the end user; the user sends a DNS name to the LDNS server and 381 gets back an IP address. On the other hand, DNS redirection is 382 problematic because the DNS request comes from the LDNS server, not 383 the end-user. This may affect the accuracy of server selection that 384 is based on the user's location. The transparency of DNS redirection 385 is also a problem in that there is no opportunity to take the 386 attributes of the user agent or the URI path component into account. 387 We consider two main forms of DNS redirection: simple and CNAME- 388 based. 390 In simple DNS redirection, the authoritative DNS server for the name 391 simply returns an IP address from a set of possible IP addresses. 392 The answer is chosen from the set based on characteristics of the set 393 (e.g., the relative loads on the servers) or characteristics of the 394 client (e.g., the location of the client relative to the servers). 395 Simple redirection is straightforward. The only caveats are (1) 396 there is a limit to the number of alternate IP addresses a single DNS 397 server can manage; and (2) DNS responses are cached by downstream 398 servers so the TTL on the response must be set to an appropriate 399 value so as to preserve the fresheness of the redirection. 401 In CNAME-based DNS redirection, the authoritative server returns a 402 CNAME response to the DNS request, telling the LDNS server to restart 403 the name lookup using a new name. A CNAME is essentially a symbolic 404 link in the DNS namespace, and like a symbolic link, redirection is 405 transparent to the client; the LDNS server gets the CNAME response 406 and re-executes the lookup. Only when the name has been resolved to 407 an IP address does it return the result to the user. Note that DNAME 408 would be preferable to CNAME if it becomes widely supported. 410 2.1.2. HTTP Redirection 412 HTTP redirection makes use of the redirection response of the HTTP 413 protocol (e.g.,"302" or "307"). This response contains a new URL 414 that the application should fetch instead of the original URL. By 415 changing the URL appropriately, the server can cause the user to 416 redirect to a different server. The advantages of HTTP redirection 417 are that (1) the server can change the URL fetched by the client to 418 include, for example, both the DNS name of the particular server to 419 use, as well as the original HTTP server that was being accessed; (2) 420 the client sends the HTTP request to the server, so that its IP 421 address is known and can be used in selecting the server; and (3) 422 other attributes (e.g., content type, user agent type) are visible to 423 the redirection mechanism. 425 The disadvantages of HTTP redirection are (1) it is visible to the 426 application, so it requires application support and may affect the 427 application behavior (e.g., web browsers will not send cookies if the 428 URL changes to a different domain); (2) HTTP is a heavyweight 429 protocol layered on TCP so it has relatively high overhead; and (3) 430 the results of HTTP redirection are not cached so that all 431 redirections must go through to the server. 433 3. Overview of CDNI Operation 435 To provide a big picture overview of the various components of CDN 436 Interconnection, we walk through a "day in the life" of a content 437 item that is made available via a pair of interconnected CDNs. This 438 will serve to illustrate many of the functions that need to be 439 supported in a complete CDNI solution. We give examples using both 440 DNS-based and HTTP-based redirection. We begin with very simple 441 examples and then how additional capabilities, such as recursive 442 request redirection and content removal, might be added. 444 Before walking through some specific examples, we present a high- 445 level view of the operations that may take place. This high-level 446 overview is illustrated in Figure 2. Note that most operations will 447 involve only a subset of all the messages shown below, and that the 448 order and number of operations may vary considerably, as more 449 detailed examples illustrate below. 451 The following shows Operator A as the upstream CDN (uCDN) and 452 Operator B as the downstream CDN (dCDN), where the former has a 453 relationship with a content provider and the latter being the CDN 454 selected by Operator A to deliver content to the end-user. The 455 interconnection relationship may be symmetric between these two CDN 456 operators, but each direction can be considered as operating 457 independently of the other so for simplicity we show the interaction 458 in one direction only. 460 End-User Operator B Operator A 461 | | | 462 | | | 463 | | [Async FCI Push] | (1) 464 | | | 465 | | [MI pre-positioning] | (2) 466 | | | 467 | CONTENT REQUEST | | 468 |-------------------------------------------------->| (3) 469 | | | 470 | | [Sync RI Pull] | (4) 471 | | | 472 | [RI REPLY] | | 473 |<--------------------------------------------------| (5) 474 | | | 475 | | | 476 | CONTENT REQUEST | | 477 |------------------------>| | (6) 478 | | | 479 | | [Sync MI Pull] | (7) 480 | | | 481 | | ACQUISITION REQUEST | 482 | X------------------------>| (8) 483 | X | 484 | X CONTENT DATA | 485 | X<------------------------| (9) 486 | | | 487 | CONTENT DATA | | 488 |<------------------------| | (10) 489 | | | 490 : : : 491 : [Other content requests] : 492 : : : 493 | | [CI: Content Purge] | (11) 494 : : : 495 | | [LI: Log exchange] | (12) 496 | | | 498 Figure 2: Overview of Operation 500 The operations shown in the Figure are as follows: 502 1. dCDN uses the FCI to advertise information relevant to its 503 delivery footprint and capabilities prior to any content 504 requests being redirected. 506 2. Prior to any content request, the uCDN uses the MI to 507 pre=position CDNI metadata to the dCDN, thereby making that 508 metadata available in readiness for later content requests. 510 3. A content request from a user agent arrives at uCDN. 512 4. uCDN may use the RI to synchronously request information from 513 dCDN regarding its delivery capabilities to decide if dCDN is a 514 suitable target for redirection of this request. 516 5. uCDN redirects the request to dCDN by sending some response 517 (DNS, HTTP) to the user agent. 519 6. The user agent requests the content from dCDN. 521 7. dCDN may use the MI to synchronously request metadata related to 522 this content from uCDN, e.g. to decide whether to serve it. 524 8. If the content is not already in a suitable cache in dCDN, dCDN 525 may acquire it from uCDN. 527 9. The content is delivered to dCDN from uCDN. 529 10. The content is delivered to the user agent by dCDN. 531 11. Some time later, perhaps at the request of the CSP (not shown) 532 uCDN may use the CI to instruct dCDN to purge the content, 533 thereby ensuring it is not delivered again. 535 12. After one or more content delivery actions by dCDN, a log of 536 delivery actions may be provided to uCDN using the LI. 538 The following sections show some more specific examples of how these 539 operations may be combined to perform various delivery, control and 540 logging operations across a pair of CDNs. 542 3.1. Preliminaries 544 Initially, we assume that there is at least one CSP that has 545 contracted with an upstream CDN (uCDN) to deliver content on its 546 behalf. We are not particularly concerned with the interface between 547 the CSP and uCDN, other than to note that it is expected to be the 548 same as in the "traditional" (non-interconnected) CDN case. Existing 549 mechanisms such as DNS CNAMEs or HTTP redirects (Section 2) can be 550 used to direct a user request for a piece of content from the CSP 551 towards the CSP's chosen upstream CDN. 553 We assume Operator A provides an upstream CDN that serves content on 554 behalf of a CSP with CDN-Domain cdn.csp.com. We assume that Operator 555 B provides a downstream CDN. An end user at some point makes a 556 request for URL 558 http://cdn.csp.com/...rest of url... 560 It may well be the case that cdn.csp.com is just a CNAME for some 561 other CDN-Domain (such as csp.op-a.net). Nevertheless, the HTTP 562 request in the examples that follow is assumed to be for the example 563 URL above. 565 Our goal is to enable content identified by the above URL to be 566 served by the CDN of operator B. In the following sections we will 567 walk through some scenarios in which content is served, as well as 568 other CDNI operations such as the removal of content from a 569 downstream CDN. 571 3.2. Iterative HTTP Redirect Example 573 In this section we walk through a simple, illustrative example using 574 HTTP redirection from uCDN to dCDN. The example also assumes the use 575 of HTTP redirection inside uCDN and dCDN; however, this is 576 independent of the choice of redirection approach across CDNs, so an 577 alternative example could be constructed still showing HTTP 578 redirection from uCDN to dCDN but using DNS for handling of request 579 inside each CDN. 581 We assume for this example that Operators A and B have established an 582 agreement to interconnect their CDNs, with A being upstream and B 583 being downstream. 585 The operators agree that a CDN-Domain peer-a.op-b.net will be used as 586 the target of redirections from uCDN to dCDN. We assume the name of 587 this domain is communicated by some means to each CDN. (This could 588 be established out-of-band or via a CDNI interface.) We refer to 589 this domain as a "distinguished" CDN-Domain to convey the fact that 590 its use is limited to the interconnection mechanism; such a domain is 591 never used directly by a CSP. 593 We assume the operators also agree on some distinguished CDN-Domain 594 that will be used for inter-CDN acquisition of CSP's content from 595 uCDN by dCDN. In this example, we'll use op-b-acq.op-a.net. 597 We assume the operators also exchange information regarding which 598 requests dCDN is prepared to serve. For example, dCDN may be 599 prepared to serve requests from clients in a given geographical 600 region or a set of IP address prefixes. This information may again 601 be provided out of band or via a defined CDNI interface. 603 We assume DNS is configured in the following way: 605 o The content provider is configured to make operator A the 606 authoritative DNS server for cdn.csp.com (or to return a CNAME for 607 cdn.csp.com for which operator A is the authoritative DNS server). 609 o Operator A is configured so that a DNS request for 610 op-b-acq.op-a.net returns a request router in Operator A. 612 o Operator B is configured so that a DNS request for peer-a.op-b.net 613 /cdn.csp.com returns a request router in Operator B. 615 Figure 3 illustrates how a client request for 617 http://cdn.csp.com/...rest of url... 619 is handled. 621 End-User Operator B Operator A 622 |DNS cdn.csp.com | | 623 |-------------------------------------------------->| 624 | | |(1) 625 |IPaddr of A's Request Router | 626 |<--------------------------------------------------| 627 |HTTP cdn.csp.com | | 628 |-------------------------------------------------->| 629 | | |(2) 630 |302 peer-a.op-b.net/cdn.csp.com | 631 |<--------------------------------------------------| 632 |DNS peer-a.op-b.net | | 633 |------------------------>| | 634 | |(3) | 635 |IPaddr of B's Request Router | 636 |<------------------------| | 637 | | | 638 |HTTP peer-a.op-b.net/cdn.csp.com | 639 |------------------------>| | 640 | |(4) | 641 |302 node1.peer-a.op-b.net/cdn.csp.com | 642 |<------------------------| | 643 |DNS node1.peer-a.op-b.net| | 644 |------------------------>| | 645 | |(5) | 646 |IPaddr of B's Delivery Node | 647 |<------------------------| | 648 | | | 649 |HTTP node1.peer-a.op-b.net/cdn.csp.com | 650 |------------------------>| | 651 | |(6) | 652 | |DNS op-b-acq.op-a.net | 653 | |------------------------>| 654 | | |(7) 655 | |IPaddr of A's Request Router 656 | |<------------------------| 657 | |HTTP op-b-acq.op-a.net | 658 | |------------------------>| 659 | | |(8) 660 | |302 node2.op-b.acq.op-A.net 661 | |<------------------------| 662 | |DNS node2.op-b-acq.op-a.net 663 | |------------------------>| 664 | | |(9) 665 | |IPaddr of A's Delivery Node 666 | |<------------------------| 667 | | | 668 | |HTTP node2.op-b-acq.op-a.net 669 | |------------------------>| 670 | | |(10) 671 | |Data | 672 | |<------------------------| 673 |Data | | 674 |<------------------------| | 676 Figure 3: Message Flow for Iterative HTTP Redirection 678 The steps illustrated in the figure are as follows: 680 1. A DNS resolver for Operator A processes the DNS request for its 681 customer based on CDN-Domain cdn.csp.com. It returns the IP 682 address of a request router in Operator A. 684 2. A Request Router for Operator A processes the HTTP request and 685 recognizes that the end-user is best served by another CDN, 686 specifically one provided by Operator B, and so it returns a 302 687 redirect message for a new URL constructed by "stacking" 688 Operator B's distinguished CDN-Domain (peer-a.op-b.net) on the 689 front of the original URL. (Note that more complex URL 690 manipulations are possible, such as replacing the initial CDN- 691 Domain by some opaque handle.) 693 3. The end-user does a DNS lookup using Operator B's distinguished 694 CDN-Domain (peer-a.op-b.net). B's DNS resolver returns the IP 695 address of a request router for Operator B. Note that if request 696 routing within dCDN was performed using DNS instead of HTTP 697 redirection, B's DNS resolver would also behave as the request 698 router and directly return the IP address of a delivery node. 700 4. The request router for Operator B processes the HTTP request and 701 selects a suitable delivery node to serve the end-user request, 702 and returns a 302 redirect message for a new URL constructed by 703 replacing the hostname with a subdomain of the Operator B's 704 distinguished CDN-Domain that points to the selected delivery 705 node. 707 5. The end-user does a DNS lookup using Operator B's delivery node 708 subdomain (node1.peer-a.op-b.net). B's DNS resolver returns the 709 IP address of the delivery node. 711 6. The end-user requests the content from B's delivery node. In 712 the case of a cache hit, steps 6, 7, 8, 9 and 10 below do not 713 happen, and the content data is directly returned by the 714 delivery node to the end-user. In the case of a cache miss, the 715 content needs to be acquired by dCDN from uCDN (not the CSP). 716 The distinguished CDN-Domain peer-a.op-b.net indicates to dCDN 717 that this content is to be acquired from uCDN; stripping the 718 CDN-Domain reveals the original CDN-Domain cdn.csp.com and dCDN 719 may verify that this CDN-Domain belongs to a known peer (so as 720 to avoid being tricked into serving as an open proxy). It then 721 does a DNS request for an inter-CDN acquisition CDN-Domain as 722 agreed above (in this case, op-b-acq.op-a.net). 724 7. Operator A's DNS resolver processes the DNS request and returns 725 the IP address of a request router in operator A. 727 8. The request router for Operator A processes the HTTP request 728 from Operator B delivery node. Operator A request router 729 recognizes that the request is from a peer CDN rather than an 730 end-user because of the dedicated inter-CDN acquisition domain 731 (op-b-acq.op-a.net). (Note that without this specially defined 732 inter-CDN acquisition domain, operator A would be at risk of 733 redirecting the request back to operator B, resulting in an 734 infinite loop). The request router for Operator A selects a 735 suitable delivery node in uCDN to serve the inter-CDN 736 acquisition request and returns a 302 redirect message for a new 737 URL constructed by replacing the hostname with a subdomain of 738 the Operator A's distinguished inter-CDN acquisition domain that 739 points to the selected delivery node. 741 9. Operator A DNS resolver processes the DNS request and returns 742 the IP address of the delivery node in operator A. 744 10. Operator B requests (acquires) the content from Operator A. 745 Although not shown, Operator A processes the rest of the URL: it 746 extracts information identifying the origin server, validates 747 that this server has been registered, and determines the content 748 provider that owns the origin server. It may also perform its 749 own content acquisition steps if needed before returning the 750 content to dCDN. 752 The main advantage of this design is that it is simple: each CDN need 753 only know the distinguished CDN-Domain for each peer, with the 754 upstream CDN "pushing" the downstream CDN-Domain onto the URL as part 755 of its redirect (step 2) and the downstream CDN "popping" its CDN- 756 Domain off the URL to expose a CDN-Domain that the upstream CDN can 757 correctly process. Neither CDN needs to be aware of the internal 758 structure of the other's URLs. Moreover, the inter-CDN redirection 759 is entirely supported by a single HTTP redirect; neither CDN needs to 760 be aware of the other's internal redirection mechanism (i.e., whether 761 it is DNS or HTTP based). 763 One disadvantage is that the end-user's browser is redirected to a 764 new URL that is not in the same domain of the original URL. This has 765 implications on a number of security or validation mechanisms 766 sometimes used on endpoints. For example, it is important that any 767 redirected URL be in the same domain (e.g., csp.com) if the browser 768 is expected to send any cookies associated with that domain. As 769 another example, some video players enforce validation of a cross 770 domain policy that needs to allow for the domains involved in the CDN 771 redirection. These problems are generally soluble, but the solutions 772 complicate the example, so we do not discuss them further in this 773 version of the draft. 775 We note that this example begins to illustrate some of the interfaces 776 that may be required for CDNI, but does not require all of them. For 777 example, obtaining information from dCDN regarding the set of client 778 IP addresses or geographic regions it might be able to serve is an 779 aspect of request routing (specifically of the CDNI Footprint & 780 Capabilities Advertisement interface). Important configuration 781 information such as the distinguished names used for redirection and 782 inter-CDN acquisition could also be conveyed via a CDNI interface 783 (e.g., perhaps the CDNI Control interface). The example also shows 784 how existing HTTP-based methods suffice for the acquisition 785 interface. Arguably, the absolute minimum metadata required for CDNI 786 is the information required to acquire the content, and this 787 information was provided "in-band" in this example by means of the 788 URI handed to the client in the HTTP 302 response. The example also 789 assumes that the CSP does not require any distribution policy (e.g. 790 time window, geo-blocking) or delivery processing to be applied by 791 the interconnected CDNs. Hence, there is no explicit CDNI Metadata 792 interface invoked in this example. There is also no explicit CDNI 793 Logging interface discussed in this example. 795 We also note that the step of deciding when a request should be 796 redirected to dCDN rather than served by uCDN has been somewhat 797 glossed over. It may be as simple as checking the client IP address 798 against a list of prefixes, or it may be considerably more complex, 799 involving a wide range of factors, such as the geographic location of 800 the client (perhaps determined from a third party service), CDN load, 801 or specific business rules. 803 This example uses the "iterative" CDNI request redirection approach. 804 That is, uCDN performs part of the request redirection function by 805 redirecting the client to a request router in the dCDN, which then 806 performs the rest of the redirection function by redirecting to a 807 suitable surrogate. If request routing is performed in the dCDN 808 using HTTP redirection, this translates in the end-user experiencing 809 two successive HTTP redirections. By contrast, the alternative 810 approach of "recursive" CDNI request redirection effectively 811 coalesces these two successive HTTP redirections into a single one, 812 sending the end-user directly to the right delivery node in the dCDN. 813 This "recursive" CDNI request routing approach is discussed in the 814 next section. 816 3.3. Recursive HTTP Redirection Example 818 The following example builds on the previous one to illustrate the 819 use of the request routing interface (specifically the CDNI Request 820 Routing Redirection interface) to enable "recursive" CDNI request 821 routing. We build on the HTTP-based redirection approach because it 822 illustrates the principles and benefits clearly, but it is equally 823 possible to perform recursive redirection when DNS-based redirection 824 is employed. 826 In contrast to the prior example, the operators need not agree in 827 advance on a CDN-Domain to serve as the target of redirections from 828 uCDN to dCDN. We assume that the operators agree on some 829 distinguished CDN-Domain that will be used for inter-CDN acquisition 830 of CSP's content by dCDN. In this example, we'll use 831 op-b-acq.op-a.net. 833 We assume the operators also exchange information regarding which 834 requests dCDN is prepared to serve. For example, dCDN may be 835 prepared to serve requests from clients in a given geographical 836 region or a set of IP address prefixes. This information may again 837 be provided out of band or via a defined protocol. 839 We assume DNS is configured in the following way: 841 o The content provider is configured to make operator A the 842 authoritative DNS server for cdn.csp.com (or to return a CNAME for 843 cdn.csp.com for which operator A is the authoritative DNS server). 845 o Operator A is configured so that a DNS request for 846 op-b-acq.op-a.net returns a request router in Operator A. 848 o Operator B is configured so that a request for node1.op-b.net/ 849 cdn.csp.com returns the IP address of a delivery node. Note that 850 there might be a number of such delivery nodes. 852 Figure 3 illustrates how a client request for 854 http://cdn.csp.com/...rest of url... 856 is handled. 858 End-User Operator B Operator A 859 |DNS cdn.csp.com | | 860 |-------------------------------------------------->| 861 | | |(1) 862 |IPaddr of A's Request Router | 863 |<--------------------------------------------------| 864 |HTTP cdn.csp.com | | 865 |-------------------------------------------------->| 866 | | |(2) 867 | |RR/RI REQ cdn.csp.com | 868 | |<------------------------| 869 | | | 870 | |RR/RI RESP node1.op-b.net| 871 | |------------------------>| 872 | | |(3) 873 |302 node1.op-b.net/cdn.csp.com | 874 |<--------------------------------------------------| 875 |DNS node1.op-b.net | | 876 |------------------------>| | 877 | |(4) | 878 |IPaddr of B's Delivery Node | 879 |<------------------------| | 880 |HTTP node1.op-b.net/cdn.csp.com | 881 |------------------------>| | 882 | |(5) | 883 | |DNS op-b-acq.op-a.net | 884 | |------------------------>| 885 | | |(6) 886 | |IPaddr of A's Request Router 887 | |<------------------------| 888 | |HTTP op-b-acq.op-a.net | 889 | |------------------------>| 890 | | |(7) 891 | |302 node2.op-b.acq.op-A.net 892 | |<------------------------| 893 | |DNS node2.op-b-acq.op-a.net 894 | |------------------------>| 895 | | |(8) 896 | |IPaddr of A's Delivery Node 897 | |<------------------------| 898 | | | 899 | |HTTP node2.op-b-acq.op-a.net 900 | |------------------------>| 901 | | |(9) 902 | |Data | 903 | |<------------------------| 904 |Data | | 905 |<------------------------| | 907 Figure 4: Message Flow for Recursive HTTP Redirection 909 The steps illustrated in the figure are as follows: 911 1. A DNS resolver for Operator A processes the DNS request for its 912 customer based on CDN-Domain cdn.csp.com. It returns the IP 913 address of a Request Router in Operator A. 915 2. A Request Router for Operator A processes the HTTP request and 916 recognizes that the end-user is best served by another 917 CDN-\u002Dspecifically one provided by Operator B-\u002Dand so it 918 queries the CDNI Request Routing Redirection interface of 919 Operator B, providing a set of information about the request 920 including the URL requested. Operator B replies with the DNS 921 name of a delivery node. 923 3. Operator A returns a 302 redirect message for a new URL obtained 924 from the RI. 926 4. The end-user does a DNS lookup using the host name of the URL 927 just provided (node1.op-b.net). B's DNS resolver returns the IP 928 address of the corresponding delivery node. Note that, since the 929 name of the delivery node was already obtained from B using the 930 RI, there should not be any further redirection here (in contrast 931 to the iterative method described above.) 933 5. The end-user requests the content from B's delivery node, 934 potentially resulting in a cache miss. In the case of a cache 935 miss, the content needs to be acquired from uCDN (not the CSP.) 936 The distinguished CDN-Domain op-b.net indicates to dCDN that this 937 content is to be acquired from another CDN; stripping the CDN- 938 Domain reveals the original CDN-Domain cdn.csp.com, dCDN may 939 verify that this CDN-Domain belongs to a known peer (so as to 940 avoid being tricked into serving as an open proxy). It then does 941 a DNS request for the inter-CDN Acquisition "distinguished" CDN- 942 Domain as agreed above (in this case, op-b-acq.op-a.net). 944 6. Operator A DNS resolver processes the DNS request and returns the 945 IP address of a request router in operator A. 947 7. The request router for Operator A processes the HTTP request from 948 Operator B delivery node. Operator A request router recognizes 949 that the request is from a peer CDN rather than an end-user 950 because of the dedicated inter-CDN acquisition domain 951 (op-b-acq.op-a.net). (Note that without this specially defined 952 inter-CDN acquisition domain, operator A would be at risk of 953 redirecting the request back to operator B, resulting in an 954 infinite loop). The request router for Operator A selects a 955 suitable delivery node in uCDN to serve the inter-CDN acquisition 956 request and returns a 302 redirect message for a new URL 957 constructed by replacing the hostname with a subdomain of the 958 Operator A's distinguished inter-CDN acquisition domain that 959 points to the selected delivery node. 961 8. Operator A recognizes that the DNS request is from a peer CDN 962 rather than an end-user (due to the internal CDN-Domain) and so 963 returns the address of a delivery node. (Note that without this 964 specially defined internal domain, Operator A would be at risk of 965 redirecting the request back to Operator B, resulting in an 966 infinite loop.) 968 9. Operator B requests (acquires) the content from Operator A. 969 Operator A serves content for the requested CDN-Domain to dCDN. 970 Although not shown, it is at this point that Operator A processes 971 the rest of the URL: it extracts information identifying the 972 origin server, validates that this server has been registered, 973 and determines the content provider that owns the origin server. 974 It may also perform its own content acquisition steps if needed 975 before returning the content to dCDN. 977 Recursive redirection has the advantage over iterative of being more 978 transparent from the end-user's perspective, but the disadvantage of 979 each CDN exposing more of its internal structure (in particular, the 980 addresses of edge caches) to peer CDNs. By contrast, iterative 981 redirection does not require dCDN to expose the addresses of its edge 982 caches to uCDN. 984 This example happens to use HTTP-based redirection in both CDN A and 985 CDN B, but a similar example could be constructed using DNS-based 986 redirection in either CDN. Hence, the key point to take away here is 987 simply that the end user only sees a single redirection of some type, 988 as opposed to the pair of redirections in the prior (iterative) 989 example. 991 The use of the RI requires that the request routing mechanism be 992 appropriately configured and bootstrapped, which is not shown here. 993 More discussion on the bootstrapping of interfaces is provided in 994 Section 4 996 3.4. Iterative DNS-based Redirection Example 998 In this section we walk through a simple example using DNS-based 999 redirection for request redirection from uCDN to dCDN (as well as for 1000 request routing inside dCDN and uCDN). As noted in Section 2.1, DNS- 1001 based redirection has certain advantages over HTTP-based redirection 1002 (notably, it is transparent to the end-user) as well as some 1003 drawbacks (notably the client IP address is not visible to the 1004 request router). 1006 As before, Operator A has to learn the set of requests that dCDN is 1007 willing or able to serve (e.g. which client IP address prefixes or 1008 geographic regions are part of the dCDN footprint). We assume 1009 Operator has and makes known to operator A some unique identifier 1010 that can be used for the construction of a distinguished CDN-Domain, 1011 as shown in more detail below. (This identifier strictly needs only 1012 to be unique within the scope of Operator A, but a globally unique 1013 identifier, such as an AS number assigned to B, is one easy way to 1014 achieve that.) Also, Operator A obtains the NS records for Operator 1015 B's externally visible redirection servers. Also, as before, a 1016 distinguished CDN-Domain, such as op-b-acq.op-a.net, must be assigned 1017 for inter-CDN acquisition. 1019 We assume DNS is configured in the following way: 1021 o The CSP is configured to make Operator A the authoritative DNS 1022 server for cdn.csp.com (or to return a CNAME for cdn.csp.com for 1023 which operator A is the authoritative DNS server). 1025 o When uCDN sees a request best served by dCDN, it returns CNAME and 1026 NS records for "b.cdn.csp.com", where "b" is the unique identifier 1027 assigned to Operator B. (It may, for example, be an AS number 1028 assigned to Operator B.) 1030 o dCDN is configured so that a request for "b.cdn.csp.com" returns a 1031 delivery node in dCDN. 1033 o uCDN is configured so that a request for "op-b-acq.op-a.net" 1034 returns a delivery node in uCDN. 1036 Figure 5 depicts the exchange of DNS and HTTP requests. The main 1037 differences from Figure 3 are the lack of HTTP redirection and 1038 transparency to the end-user. 1040 End-User Operator B Operator A 1041 |DNS cdn.csp.com | | 1042 |-------------------------------------------------->| 1043 | | |(1) 1044 |CNAME b.cdn.csp.com | | 1045 |NS records for b.cdn.csp.com | 1046 |<--------------------------------------------------| 1047 |DNS b.cdn.csp.com | | 1048 |------------------------>| | 1049 | |(2) | 1050 |IPaddr of B's Delivery Node | 1051 |<------------------------| | 1052 |HTTP cdn.csp.com | | 1053 |------------------------>| | 1054 | |(3) | 1055 | |DNS op-b-acq.op-a.net | 1056 | |------------------------>| 1057 | | |(4) 1058 | |IPaddr of A's Delivery Node 1059 | |<------------------------| 1060 | |HTTP op-b-acq.op-a.net | 1061 | |------------------------>| 1062 | | |(5) 1063 | |Data | 1064 | |<------------------------| 1065 |Data | | 1066 |<------------------------| | 1068 Figure 5: Message Flow for DNS-based Redirection 1070 The steps illustrated in the figure are as follows: 1072 1. Request Router for Operator A processes the DNS request for CDN- 1073 Domain cdn.csp.com and recognizes that the end-user is best 1074 served by another CDN. (This may depend on the IP address of the 1075 user's local DNS resolver, or other information discussed below.) 1076 The Request Router returns a DNS CNAME response by "stacking" the 1077 distinguished identifier for Operator B onto the original CDN- 1078 Domain (e.g., b.cdn.csp.com), plus an NS record that maps 1079 b.cdn.csp.com to B's Request Router. 1081 2. The end-user does a DNS lookup using the modified CDN-Domain 1082 (i.e., b.cdn.csp.com). This causes B's Request Router to respond 1083 with a suitable delivery node. 1085 3. The end-user requests the content from B's delivery node. The 1086 requested URL contains the name cdn.csp.com. (Note that the 1087 returned CNAME does not affect the URL.) At this point the 1088 delivery node has the correct IP address of the end-user and can 1089 do an HTTP 302 redirect if the redirections in steps 2 and 3 were 1090 incorrect. Otherwise B verifies that this CDN-Domain belongs to 1091 a known peer (so as to avoid being tricked into serving as an 1092 open proxy). It then does a DNS request for an "internal" CDN- 1093 Domain as agreed above (op-b-acq.op-a.net). 1095 4. Operator A recognizes that the DNS request is from a peer CDN 1096 rather than an end-user (due to the internal CDN-Domain) and so 1097 returns the address of a delivery node in uCDN. 1099 5. Operator A serves content to dCDN. Although not shown, it is at 1100 this point that Operator A processes the rest of the URL: it 1101 extracts information identifying the origin server, validates 1102 that this server has been registered, and determines the content 1103 provider that owns the origin server. 1105 The advantages of this approach are that it is more transparent to 1106 the end-user and requires fewer round trips than HTTP-based 1107 redirection (in its worst case, i.e., when none of the needed DNS 1108 information is cached). A potential problem is that the upstream CDN 1109 depends on being able to learn the correct downstream CDN that serves 1110 the end-user from the client address in the DNS request. In standard 1111 DNS operation, uCDN will only obtain the address of the client's 1112 local DNS resolver (LDNS), which is not guaranteed to be in the same 1113 network (or geographic region) as the client. If not-\u002De.g., the 1114 end-user uses a global DNS service-\u002Dthen the upstream CDN cannot 1115 determine the appropriate downstream CDN to serve the end-user. In 1116 this case, and assuming the uCDN is capable of detecting that 1117 situation, one option is for the upstream CDN to treat the end-user 1118 as it would any user not connected to a peer CDN. Another option is 1119 for the upstream CDN to "fall back" to a pure HTTP-based redirection 1120 strategy in this case (i.e., use the first method). Note that this 1121 problem affects existing CDNs that rely on DNS to determine where to 1122 redirect client requests, but the consequences are arguably less 1123 serious for CDNI since the LDNS is likely in the same network as the 1124 dCDN serves. One approach to ensuring that the client's IP address 1125 prefix is correctly determined in such situations is described in 1126 [I-D.vandergaast-edns-client-subnet]. 1128 As with the prior example, this example partially illustrates the 1129 various interfaces involved in CDNI. Operator A could learn 1130 dynamically from Operator B the set of prefixes or regions that B is 1131 willing and able to serve via the CDNI Footprint & Capabilities 1132 Advertisement interface. The distinguished name used for acquisition 1133 and the identifier for Operator B that is prepended to the CDN-Domain 1134 on redirection are examples of information elements that might also 1135 be conveyed by CDNI interfaces (or, alternatively, statically 1136 configured). As before, minimal metadata sufficient to obtain the 1137 content is carried "in-band" as part of the redirection process, and 1138 standard HTTP is used for inter-CDN acquisition. There is no 1139 explicit CDNI Logging interface discussed in this example. 1141 3.5. Dynamic Footprint Discovery Example 1143 There could be situations where being able to dynamically discover 1144 the set of requests that a given dCDN is willing and able to serve is 1145 beneficial. For example, a CDN might at one time be able to serve a 1146 certain set of client IP prefixes, but that set might change over 1147 time due to changes in the topology and routing policies of the IP 1148 network. The following example illustrates this capability. We have 1149 chosen the example of DNS-based redirection, but HTTP-based 1150 redirection could equally well use this approach. 1152 End-User Operator B Operator A 1153 |DNS cdn.csp.com | | 1154 |-------------------------------------------------->| 1155 | | |(1) 1156 | | RI REQ op-b.net | 1157 | |<------------------------| 1158 | | |(2) 1159 | | RI REPLY | 1160 | |------------------------>| 1161 | | |(3) 1162 |CNAME b.cdn.csp.com | | 1163 |NS records for b.cdn.csp.com | 1164 |<--------------------------------------------------| 1165 |DNS b.cdn.csp.com | | 1166 |------------------------>| | 1167 | |(2) | 1168 |IPaddr of B's Delivery Node | 1169 |<------------------------| | 1170 |HTTP cdn.csp.com | | 1171 |------------------------>| | 1172 | |(3) | 1173 | |DNS op-b-acq.op-a.net | 1174 | |------------------------>| 1175 | | |(4) 1176 | |IPaddr of A's Delivery Node 1177 | |<------------------------| 1178 | |HTTP op-b-acq.op-a.net | 1179 | |------------------------>| 1180 | | |(5) 1181 | |Data | 1182 | |<------------------------| 1183 |Data | | 1184 |<------------------------| | 1186 Figure 6: Message Flow for Dynamic Footprint Discovery 1188 This example differs from the one in Figure 5 only in the addition of 1189 a FCI request (step 2) and corresponding response (step 3). The RI 1190 REQ could be a message such as "Can you serve clients from this IP 1191 Prefix?" or it could be "Provide the list of client IP prefixes you 1192 can currently serve". In either case the response might be cached by 1193 operator A to avoid repeatedly asking the same question. 1194 Alternatively, or in addition, Operator B may spontaneously advertise 1195 to Operator A information (or changes) on the set of requests it is 1196 willing and able to serve on behalf of operator A; in that case, 1197 Operator B may spontaneously issue RR/RI REPLY messages that are not 1198 in direct response to a corresponding RR/RI REQ message. (Note that 1199 the issues of determining the client's subnet from DNS requests, as 1200 described above, are exactly the same here as in Section 3.4.) 1202 Once Operator A obtains the RI response, it is now able to determine 1203 that Operator B's CDN is an appropriate dCDN for this request and 1204 therefore a valid candidate dCDN to consider in its Redirection 1205 decision. If that dCDN is selected, the redirection and serving of 1206 the request proceeds as before (i.e. in the absence of dynamic 1207 footprint discovery). 1209 3.6. Content Removal Example 1211 The following example illustrates how the CDNI Control interface may 1212 be used to achieve pre-positioning of an item of content in the dCDN. 1213 In this example, user requests for a particular content, and 1214 corresponding redirection of such requests from Operator A to 1215 Operator B CDN, may (or may not) have taken place earlier. Then, at 1216 some point in time, the uCDN (for example, in response to a 1217 corresponding Trigger from the Content Provider) uses the CI to 1218 request that content identified by a particular URL be removed from 1219 dCDN. The following diagram illustrates the operation. 1221 End-User Operator B Operator A 1222 | |CI purge cdn.csp.com/... | 1223 | |<------------------------| 1224 | | | 1225 | |CI OK | 1226 | |------------------------>| 1227 | | | 1229 Figure 7: Message Flow for Content Removal 1231 The CI is used to convey the request from uCDN to dCDN that some 1232 previously acquired content should be deleted. The URL in the 1233 request specifies which content to remove. This example corresponds 1234 to a DNS-based redirection scenario such as Section 3.4. If HTTP- 1235 based redirection had been used, the URL for removal would be of the 1236 form peer-a.op-b.net/cdn.csp.com/... 1238 The dCDN is expected to confirm to the uCDN, as illustrated by the CI 1239 OK message, the completion of the removal of the targeted content 1240 from all the caches in dCDN. 1242 3.7. Pre-Positioned Content Acquisition Example 1244 The following example illustrates how the CI may be used to pre- 1245 position an item of content in the dCDN. In this example, Operator A 1246 uses the CDNI Metadata interface to request that content identified 1247 by a particular URL be pre-positioned into Operator B CDN. 1249 End-User Operator B Operator A 1251 | |CI pre-position cdn.csp.com/... 1252 | |<------------------------| 1253 | | |(1) 1254 | |CI OK | 1255 | |------------------------>| 1256 | | | 1257 | |DNS op-b-acq.op-a.net | 1258 | |------------------------>| 1259 | | |(2) 1260 | |IPaddr of A's Delivery Node 1261 | |<------------------------| 1262 | |HTTP op-b-acq.op-a.net | 1263 | |------------------------>| 1264 | | |(3) 1265 | |Data | 1266 | |<------------------------| 1267 |DNS cdn.csp.com | | 1268 |-------------------------------------------------->| 1269 | | |(4) 1270 |IPaddr of A's Request Router | 1271 |<--------------------------------------------------| 1272 |HTTP cdn.csp.com | | 1273 |-------------------------------------------------->| 1274 | | |(5) 1275 |302 peer-a.op-b.net/cdn.csp.com | 1276 |<--------------------------------------------------| 1277 |DNS peer-a.op-b.net | | 1278 |------------------------>| | 1279 | |(6) | 1280 |IPaddr of B's Delivery Node | 1281 |<------------------------| | 1282 |HTTP peer-a.op-b.net/cdn.csp.com | 1283 |------------------------>| | 1284 | |(7) | 1285 |Data | | 1286 |<------------------------| | 1288 Figure 8: Message Flow for Content Pre-Positioning 1290 The steps illustrated in the figure are as follows: 1292 1. Operator A uses the CI to request that Operator B pre-positions a 1293 particular content item identified by its URL. Operator B 1294 responds by confirming that it is willing to perform this 1295 operation. 1297 Steps 2 and 3 are exactly the same as steps 5 and 6 of Figure 3, only 1298 this time those steps happen as the result of the Pre-positioning 1299 request instead of as the result of a cache miss. 1301 Steps 4, 5, 6, 7 are exactly the same as steps 1, 2, 3, 4 of Figure 1302 3, only this time Operator B CDN can serve the end-user request 1303 without triggering dynamic content acquisition, since the content has 1304 been pre-positioned in dCDN. Note that, depending on dCDN operations 1305 and policies, the content pre-positioned in the dCDN may be pre- 1306 positioned to all, or a subset of, dCDN caches. In the latter case, 1307 intra-CDN dynamic content acquisition may take place inside the dCDN 1308 serving requests from caches on which the content has not been pre- 1309 positioning; however, such intra-CDN dynamic acquisition would not 1310 involve the uCDN. 1312 3.8. Asynchronous CDNI Metadata Example 1314 In this section we walk through a simple example illustrating a 1315 scenario of asynchronously exchanging CDNI metadata, where the 1316 downstream CDN obtains CDNI metadata for content ahead of a 1317 corresponding content request. The example that follows assumes that 1318 HTTP-based inter-CDN redirection and recursive CDNI request-routing 1319 are used, as in Section 3.3. However, Asynchronous exchange of CDNI 1320 Metadata is similarly applicable to DNS-based inter-CDN redirection 1321 and iterative request routing (in which cases the CDNI metadata may 1322 be used at slightly different processing stages of the message 1323 flows). 1325 End-User Operator B Operator A 1326 | | | 1327 | |CI pre-position (Trigger)| 1328 | |<------------------------|(1) 1329 | | | 1330 | |CI OK | 1331 | |------------------------>|(2) 1332 | | | 1333 | |MI pull REQ | 1334 | |------------------------>|(3) 1335 | | | 1336 | |MI metadata REP |(4) 1337 | | | 1338 | | | 1339 | CONTENT REQUEST | | 1340 |-------------------------------------------------->|(5) 1341 | | | 1342 | | RI REQ | 1343 | |<------------------------|(6) 1344 | | | 1345 | | RI RESP | 1346 | |------------------------>|(7) 1347 | | | 1348 | CONTENT REDIRECTION | | 1349 |<--------------------------------------------------| (8) 1350 | | | 1351 | CONTENT REQUEST | | 1352 |------------------------>|(9) | 1353 | | | 1354 : : : 1355 | CONTENT DATA | | 1356 |<------------------------| |(10) 1358 Figure 9: Message Flow for Asynchronous CDNI Metadata 1360 The steps illustrated in the figure are as follows: 1362 1. Operator A uses the CI to Trigger to signal the availability of 1363 CDNI metadata to Operator B. 1365 2. Operator B acknowledges the receipt of this Trigger. 1367 3. Operator B requests the latest metadata from Operator A using 1368 the MI. 1370 4. Operator A replies with the requested metadata. This document 1371 does not constrain how the CDNI metadata information is actually 1372 represented. For the purposes of this example, we assume that 1373 Operator A provides CDNI metadata to Operator B indicating that: 1375 * this CDNI Metadata is applicable to any content referenced by 1376 some CDN-Domain. 1378 * this CDNI metadata consists of a distribution policy 1379 requiring enforcement by the delivery node of a specific per- 1380 request authorization mechanism (e.g. URI signature or token 1381 validation). 1383 5. A Content Request occurs as usual. 1385 6. A CDNI Request Routing Redirection request (RI REQ) is issued by 1386 operator A CDN, as discussed in Section 3.3. Operator B's 1387 request router can access the CDNI Metadata that are relevant to 1388 the requested content and that have been pre-positioned as per 1389 Steps 1-4, which may or may not affect the response. 1391 7. Operator B's request router issues a CDNI Request Routing 1392 Redirection response (RI RESP) as in Section 3.3. 1394 8. Operator B performs content redirection as discussed in 1395 Section 3.3. 1397 9. On receipt of the Content Request by the end user, the delivery 1398 node detects that previously acquired CDNI metadata is 1399 applicable to the requested content. In accordance with the 1400 specific CDNI metadata of this example, the delivery node will 1401 invoke the appropriate per-request authorization mechanism, 1402 before serving the content. (Details of this authorization are 1403 not shown.) 1405 10. Assuming successful per-request authorization, serving of 1406 Content Data (possibly preceded by inter-CDN acquisition) 1407 proceeds as in Section 3.3. 1409 3.9. Synchronous CDNI Metadata Acquisition Example 1411 In this section we walk through a simple example illustrating a 1412 scenario of Synchronous CDNI metadata acquisition, in which the 1413 downstream CDN obtains CDNI metadata for content at the time of 1414 handling a first request for the corresponding content. As in the 1415 preceding section, this example assumes that HTTP-based inter-CDN 1416 redirection and recursive CDNI request-routing are used (as in 1417 Section 3.3), but dynamic CDNI metadata acquisition is applicable to 1418 other variations of request routing. 1420 End-User Operator B Operator A 1421 | | | 1422 | CONTENT REQUEST | | 1423 |-------------------------------------------------->|(1) 1424 | | | 1425 | | RI REQ | 1426 | (2)|<------------------------| 1427 | | | 1428 | | MI REQ | 1429 | (3)|------------------------>| 1430 | | MI RESP | 1431 | |<------------------------|(4) 1432 | | | 1433 | | RI RESP | 1434 | |------------------------>|(5) 1435 | | | 1436 | | | 1437 | CONTENT REDIRECTION | | 1438 |<--------------------------------------------------|(6) 1439 | | | 1440 | CONTENT REQUEST | | 1441 |------------------------>| (7) | 1442 | | | 1443 | | MI REQ | 1444 | (8)|------------------------>| 1445 | | MI RESP | 1446 | |<------------------------|(9) 1447 | | | 1448 : : : 1449 | CONTENT DATA | | 1450 |<------------------------| | (10) 1452 Figure 10: Message Flow for Synchronous CDNI Metadata Acquisition 1454 The steps illustrated in the figure are as follows: 1456 1. A Content Request arrives as normal. 1458 2. An RI request occurs as in the prior example. 1460 3. On receipt of the CDNI Request Routing Request, Operator B's CDN 1461 initiates Synchronous acquisition of CDNI Metadata that are 1462 needed for routing of the end-user request. We assume the URI 1463 for the a Metadata server is known ahead of time through some 1464 out-of-band means. 1466 4. On receipt of a CDNI Metadata Request, Operator A's CDN 1467 responds, making the corresponding CDNI metadata information 1468 available to Operator B's CDN. This metadata is considered by 1469 operator B's CDN before responding to the Request Routing 1470 request. (In a simple case, the metadata could simply be an 1471 allow or deny response for this particular request.) 1473 5. Response to the RI request as normal. 1475 6. Redirection message is sent to the end user. 1477 7. A delivery node of Operator B receives the end user request. 1479 8. The delivery node Triggers dynamic acquisition of additional 1480 CDNI metadata that are needed to process the end-user content 1481 request. Note that there may exist cases where this step need 1482 not happen, for example because the metadata were already 1483 acquired previously. 1485 9. Operator A's CDN responds to the CDNI Metadata Request and makes 1486 the corresponding CDNI metadata available to Operator B. This 1487 metadata influence how Operator B's CDN processes the end-user 1488 request. 1490 10. Content is served (possibly preceded by inter-CDN acquisition) 1491 as in Section 3.3. 1493 3.10. Content and Metadata Acquisition with Multiple Upstream CDNs 1495 A single dCDN may receive end-user requests from multiple uCDNs. 1496 When a dCDN receives an end-user request, it must determine the 1497 identity of the uCDN from which it should acquire the requested 1498 content. 1500 Ideally, the acquisition path of an end-user request will follow the 1501 redirection path of the request. The dCDN should acquire the content 1502 from the same uCDN which redirected the request. 1504 Determining the acquisition path requires the dCDN to reconstruct the 1505 redirection path based on information in the end-user request. The 1506 method for reconstructing the redirection path differs based on the 1507 redirection approach: HTTP or DNS. 1509 With HTTP-redirection, the rewritten URI should include sufficient 1510 information for the dCDN to directly or indirectly determine the uCDN 1511 when the end-user request is received. The HTTP-redirection approach 1512 can be further broken-down based on the how the URL is rewritten 1513 during redirection: HTTP-redirection with or without Site 1514 Aggregation. HTTP-redirection with Site Aggregation hides the 1515 identity of the original CSP. HTTP-redirection without Site 1516 Aggregation does not attempt to hide the identity of the original 1517 CSP. With both approaches, the rewritten URI includes enough 1518 information to identify the immediate neighbor uCDN. 1520 With DNS-redirection, the dCDN receives the published URI (instead of 1521 a rewritten URI) and does not have sufficient information for the 1522 dCDN to identify the appropriate uCDN. The dCDN may narrow the set 1523 of viable uCDNs by examining the CDNI metadata from each to determine 1524 which uCDNs are hosting metadata for the requested content. If there 1525 is a single uCDN hosting metadata for the requested content, the dCDN 1526 can assume that the request redirection is coming from this uCDN and 1527 can acquire content from that uCDN. If there are multiple uCDNs 1528 hosting metadata for the requested content, the dCDN may be ready to 1529 trust any of these uCDNs to acquire the content (provided the uCDN is 1530 in a position to serve it). If the dCDN is not ready to trust any of 1531 these uCDNs, it needs to ensure via out of band arrangements that, 1532 for a given content, only a single uCDN will ever redirect requests 1533 to the dCDN. 1535 Content acquisition may be preceded by content metadata acquisition. 1536 If possible, the acquisition path for metadata should also follow the 1537 redirection path. Additionally, we assume metadata is indexed based 1538 on rewritten URIs in the case of HTTP-redirection and is indexed 1539 based on published URIs in the case of DNS-redirection. Thus, the RI 1540 and the MI are tightly coupled in that the result of request routing 1541 (a rewritten URI pointing to the dCDN) serves as an input to metadata 1542 lookup. If the content metadata includes information for acquiring 1543 the content, then the MI is also tightly coupled with the acquisition 1544 interface in that the result of the metadata lookup (an acquisition 1545 URL likely hosted by the uCDN) should serve as input to the content 1546 acquisition. 1548 4. Main Interfaces 1550 Figure 1 illustrates the main interfaces that are in scope for the 1551 CDNI WG, along with several others. The detailed specifications of 1552 these interfaces are left to other documents, but see RFC 6707 and 1553 [I-D.ietf-cdni-requirements] for some discussion of the interfaces. 1555 One interface that is not shown in Figure 1 is the interface between 1556 the user and the CSP. While for the purposes of CDNI that interface 1557 is out of scope, it is worth noting that it does exist and can 1558 provide useful functions, such as end-to-end performance monitoring 1559 and some forms of authentication and authorization. 1561 There is also an important interface between the user and the Request 1562 Routing function of both uCDN and dCDN (shown as the "Request" 1563 Interface in Figure 1). As we saw in some of the preceding examples, 1564 that interface can be used as a way of passing information a subset 1565 of metadata such as the minimum information that is required for dCDN 1566 to obtain the content from uCDN. 1568 In this section we will provide an overview of the functions 1569 performed by each of the CDNI interfaces and discuss how they fit 1570 into the overall solution. We also examine some of the design 1571 tradeoffs, and explore several cross-interface concerns. We begin 1572 with an examination of one such tradeoff that affects all the 1573 interfaces - the use of in-band or out-of-band communication. 1575 4.1. In-Band versus Out-of-Band Interfaces 1577 Before getting to the individual interfaces, we observe that there is 1578 a high-level design choice for each, involving the use of existing 1579 in-band communication channels versus defining new out-of-band 1580 interfaces. 1582 It is possible that the information needed to carry out various 1583 interconnection functions can be communicated between peer CDNs using 1584 existing in-band protocols. The use of HTTP 302 redirect is an 1585 example of how certain aspects of request routing can be implemented 1586 in-band (embedded in URIs). Note that using existing in-band 1587 protocols does not imply that the CDNI interfaces are null; it is 1588 still necessary to establish the rules (conventions) by which such 1589 protocols are used to implement the various interface functions. 1591 There are other opportunities for in-band communication beyond HTTP 1592 redirects. For example, many of the HTTP directives used by proxy 1593 servers can also be used by peer CDNs to inform each other of caching 1594 activity. Of these, one that is particularly relevant is the If- 1595 Modified-Since directive, which is used with the GET method to make 1596 it conditional: if the requested object has not been modified since 1597 the time specified in this field, a copy of the object will not be 1598 returned, and instead, a 304 (not modified) response will be 1599 returned. 1601 4.2. Cross Interface Concerns 1603 Although the CDNI interfaces are largely independent, there are a set 1604 of conventions practiced consistently across all interfaces. Most 1605 important among these is how resources are named, for exampmle, how 1606 the CDNI Metadata and Control interfaces identify the set of 1607 resources to which a given directive applies, or the CDNI Logging 1608 interface identifies the set of resources for which a summary record 1609 applies. 1611 While in the limit the CDNI interfaces could explicitly identify 1612 every individual resource, in practice, they name resource aggregates 1613 (sets of URIs) that are to be treated in a similar way. For example, 1614 URI aggregates can be identified by a CDN-Domain (i.e., the FQDN at 1615 the beginning of a URI) or by a URI-Filter (i.e., a regular 1616 expression that matches a subset of URIs contained in some CDN- 1617 Doman). In other words, CDN-Domains and URI-Filters provide a 1618 uniform means to aggregate sets (and subsets) of URIs for the purpose 1619 of defining the scope for some operation in one of the CDNI 1620 interfaces. 1622 4.3. Request Routing Interfaces 1624 The Request Routing interface comprises two parts: the Asynchronous 1625 interface used by a dCDN to advertize footprint and capabilities 1626 (denoted FCI) to a uCDN, allowing the uCDN to decide whether to 1627 redirect particular user requests to that dCDN; and the Synchronous 1628 interface used by the uCDN to redirect a user request to the dCDN 1629 (denoted RI). (These are somewhat analogous to the operations of 1630 routing and forwarding in IP.) 1632 As illustrated in Section 3, the RI part of request routing may be 1633 implemented in part by DNS and HTTP. Naming conventions may be 1634 established by which CDN peers communicate whether a request should 1635 be routed or content served. 1637 We also note that RI plays a key role in enabling recursive 1638 redirection, as illustrated in Section 3.3. It enables the user to 1639 be redirected to the correct delivery node in dCDN with only a single 1640 redirection step (as seen by the user). This may be particularly 1641 valuable as the chain of interconnected CDNs increases beyond two 1642 CDNs. 1644 In support of these redirection requests, it is necessary for CDN 1645 peers to exchange additional information with each other, and this is 1646 the role of the FCI part of request routing. Depending on the 1647 method(s) supported, this might includes 1649 o The operator's unique id (operator-id) or distinguished CDN-Domain 1650 (operator-domain); 1652 o NS records for the operator's set of externally visible request 1653 routers; 1655 o The set of requests the dCDN operator is prepared to serve (e.g. a 1656 set of client IP prefixes or geographic regions that may be served 1657 by dCDN). 1659 o Additional capabilities of the dCDN, such as its ability to 1660 support different CDNI Metadata requests. 1662 Note that the set of requests that dCDN is willing to serve could in 1663 some cases be relatively static (e.g., a set of IP prefixes) which 1664 could be exchanged off-line, or might even be negotiated as part of a 1665 peering agreement. However, it may also be more dynamic, in which 1666 case the exchange supported by FCI would be be helpful. A further 1667 discussion of the Footprint & Capability Advertisement interface can 1668 be found in [I-D.spp-cdni-rr-foot-cap-semantics]. 1670 4.4. CDNI Logging Interface 1672 It is necessary for the upstream CDN to have visibility into the 1673 delivery of content that it redirected to a downstream CDN. This 1674 allows the upstream CDN to properly bill its customers for multiple 1675 deliveries of content cached by the downstream CDN, as well as to 1676 report accurate traffic statistics to those content providers. This 1677 is one role of the LI. 1679 Other operational data that may be relevant to CDNI can also be 1680 exchanged by the LI. For example, dCDN may report the amount of 1681 content it has acquired from uCDN, and how much cache storage has 1682 been consumed by content cached on behalf of uCDN. 1684 Traffic logs are easily exchanged off-line. For example, the 1685 following traffic log is a small deviation from the Apache log file 1686 format, where entries include the following fields: 1688 o Domain - the full domain name of the origin server 1690 o IP address - the IP address of the client making the request 1692 o End time - the ending time of the transfer 1694 o Time zone - any time zone modifier for the end time 1696 o Method - the transfer command itself (e.g., GET, POST, HEAD) 1698 o URL - the requested URL 1700 o Version - the protocol version, such as HTTP/1.0 1702 o Response - a numeric response code indicating transfer result 1704 o Bytes Sent - the number of bytes in the body sent to the client 1706 o Request ID - a unique identifier for this transfer 1708 o User agent - the user agent, if supplied 1710 o Duration - the duration of the transfer in milliseconds 1711 o Cached Bytes - the number of body bytes served from the cache 1713 o Referrer - the referrer string from the client, if supplied 1715 Of these, only the Domain field is indirect in the downstream 1716 CDN-\u002Dit is set to the CDN-Domain used by the upstream CDN rather 1717 than the actual origin server. This field could then used to filter 1718 traffic log entries so only those entries matching the upstream CDN 1719 are reported to the corresponding operator. Further discussion of 1720 the LI can be found in [I-D.bertrand-cdni-logging]. 1722 One open question is who does the filtering. One option is that the 1723 downstream CDN filters its own logs, and passes the relevant records 1724 directly to each upstream peer. This requires that the downstream 1725 CDN knows the set of CDN-Domains that belong to each upstream peer. 1726 If this information is already exchanged between peers as part of 1727 another interface, then direct peer-to-peer reporting is 1728 straightforward. If it is not available, and operators do not wish 1729 to advertise the set of CDN-Domains they serve to their peers, then 1730 the second option is for each CDN to send both its non-local traffic 1731 records and the set of CDN-Domains it serves to an independent third- 1732 party (i.e., a CDN Exchange), which subsequently filters, merges, and 1733 distributes traffic records on behalf of each participating CDN 1734 operator. 1736 A second open question is how timely traffic information should be. 1737 For example, in addition to offline traffic logs, accurate real-time 1738 traffic monitoring might also be useful, but such information 1739 requires that the downstream CDN inform the upstream CDN each time it 1740 serves upstream content from its cache. The downstream CDN can do 1741 this, for example, by sending a conditional HTTP GET request (If- 1742 Modified-Since) to the upstream CDN each time it receives an HTTP GET 1743 request from one of its end-users. This allows the upstream CDN to 1744 record that a request has been issued for the purpose of real-time 1745 traffic monitoring. The upstream CDN can also use this information 1746 to validate the traffic logs received later from the downstream CDN. 1748 There is obviously a tradeoff between accuracy of such monitoring and 1749 the overhead of the downstream CDN having to go back to the upstream 1750 CDN for every request. 1752 Another design tradeoff in the LI is the degree of aggregation or 1753 summarization of data. One situation that lends itself to 1754 summarization is the delivery of HTTP adaptive streaming (HAS), since 1755 the large number of individual chunk requests potentially results in 1756 large volumes of logging information. This case is discussed below, 1757 but other forms of aggregation may also be useful. For example, 1758 there may be situations where bulk metrics such as bytes delivered 1759 per hour may suffice rather than the detailed per-request logs 1760 outlined above. It seems likely that a range of granularities of 1761 logging will be needed along with ways to specify the type and degree 1762 of aggregation required. 1764 4.5. CDNI Control Interface 1766 The CDNI Control interface is initially used to bootstrap the other 1767 interfaces. As a simple example, it could be used to provide the 1768 address of the logging server in dCDN to uCDN in order to bootstrap 1769 the CDNI Logging interface. It may also be used, for example, to 1770 establish security associations for the other interfaces. 1772 The other role the CI plays is to allow the uCDN to pre-position, 1773 revalidate, or purge metadata and content on a dCDN. These 1774 operations, sometimes collectively called the Trigger interface, are 1775 discussed further in [I-D.murray-cdni-triggers]. 1777 4.6. CDNI Metadata Interface 1779 The role of the CDNI Metadata interface is to enable CDNI 1780 distribution metadata to be conveyed to the downstream CDN by the 1781 upstream CDN. For example, see [I-D.ietf-cdni-metadata]. Such 1782 metadata includes geo-blocking restrictions, availability windows, 1783 access control policies, and so on. It may also include information 1784 to facilitate acquisition of content by dCDN (e.g., alternate sources 1785 for the content, authorization information needed to acquire the 1786 content from the source). 1788 Some distribution metadata may be partially emulated using in-band 1789 mechanisms. For example, in case of any geo-blocking restrictions or 1790 availability windows, the upstream CDN can elect to redirect a 1791 request to the downstream CDN only if that CDN's advertised delivery 1792 footprint is acceptable for the requested URL. Similarly, the 1793 request could be forwarded only if the current time is within the 1794 availability window. However, such approaches typically come with 1795 shortcomings such as inability to prevent from replay outside the 1796 time window or inability to make use of a downstream CDN that covers 1797 a broader footprint than the geo-blocking restrictions. 1799 Similarly, some forms of access control may also be performed on a 1800 per-request basis using HTTP directives. For example, being able to 1801 respond to a conditional GET request gives the upstream CDN an 1802 opportunity to influence how the downstream CDN delivers its content. 1803 Minimally, the upstream CDN can invalidate (purge) content previously 1804 cached by the downstream CDN. 1806 Fine-grain control over how the downstream CDN delivers content on 1807 behalf of the upstream CDN is also possible. For example, by 1808 including the Forwarded HTTP header [I-D.ietf-appsawg-http-forwarded] 1809 with the conditional GET request, the downstream CDN can report the 1810 end-user's IP address to the upstream CDN, giving it an opportunity 1811 to control whether the downstream CDN should serve the content to 1812 this particular end-user. The upstream CDN would communicate its 1813 directive through its response to the conditional GET. The 1814 downstream CDN can cache information for a period of time specified 1815 by the upstream CDN, thereby reducing control overhead, but then 1816 preventing per-request control during the corresponding caching 1817 period. 1819 All of these in-band techniques serve to illustrate that uCDNs have 1820 the option of enforcing some of their access control policies 1821 themselves (at the expense of increased inter-CDN signaling load), 1822 rather than delegating enforcement to dCDNs using the MI. As a 1823 consequence, the MI could provide a means for the uCDN to express its 1824 desire to retain enforcement for itself. For example, this might be 1825 done by including a "check with me" flag in the metadata associated 1826 with certain content. The realization of such in-band techniques 1827 over the various inter-CDN acquisition protocols (e.g., HTTP) 1828 requires further investigation and may require small extensions or 1829 semantic changes to the acquisition protocol. 1831 4.7. HTTP Adaptive Streaming Concerns 1833 We consider HTTP Adaptive Streaming (HAS) and the impact it has on 1834 the CDNI interfaces because large objects (e.g., videos) are broken 1835 into a sequence of small, independent chunks. For each of the 1836 following, a more thorough discussion, including an overview of the 1837 tradeoffs involved in alternative designs, can be found in RFC 6983. 1839 First, with respect to Content Acquisition and File Management, which 1840 are out-of-scope for the CDNI interfaces but nontheless relevant to 1841 the overall operation, we assume no additional measures are required 1842 to deal with large numbers of chunks. This means that the dCDN is 1843 not explicitly made aware of any relationship between different 1844 chunks and the dCDN handles each chunk as if it were an individual 1845 and independent content item. The result is that content acquisition 1846 between uCDN and dCDN also happens on a per-chunk basis. This 1847 approach is in line with the recommendations made in RFC 6983, which 1848 also identifies potential improvements in this area that might be 1849 considered in the future. 1851 Second, with respect to Request Routing, we note that HAS manifest 1852 files have the potential to interfere with request routing since 1853 manifest files contain URLs pointing to the location of content 1854 chunks. To make sure that a manifest file does not hinder CDNI 1855 request routing and does not place excessive load on CDNI resources, 1856 the use of manifest files could either be limited to those containing 1857 relative URLs or the uCDN could modify the URLs in the manifest. Our 1858 approach for dealing with these issues is twofold. As a mandatory 1859 requirement, CDNs should be able to handle unmodified manifest files 1860 containing either relative or absolute URLs. To limit the number of 1861 redirects, and thus the load placed on the CDNI interfaces, as an 1862 optional feature uCDNs can use the information obtained through the 1863 CNDI Request Routing Redirection interface to modify the URLs in the 1864 manifest file. Since the modification of the manifest file is an 1865 optional uCDN-internal process, this does not require any 1866 standardization effort beyond being able to communicate chunk 1867 locations in the CDNI Request Routing Redirection interface. 1869 Third, with respect to the CDNI Logging interface, there are several 1870 potential issues, including the large number of individual chunk 1871 requests potentially resulting in large volumes of logging 1872 information, and the desire to correlate logging information for 1873 chunk requests that correspond to the same HAS session. For the 1874 initial CDNI specification, our approach is to expect participating 1875 CDNs to support per-chunk logging (e.g. logging each chunk request as 1876 if it were an independent content request) over the CDNI Logging 1877 interface. Optionally, the LI may include a Content Collection 1878 IDentifier (CCID) and/or a Session IDentifier (SID) as part of the 1879 logging fields, thereby facilitating correlation of per-chunk logs 1880 into per-session logs for applications benefiting from such session 1881 level information (e.g. session-based analytics). This approach is 1882 in line with the recommendations made in RFC 6983, which also 1883 identifies potential improvements in this area that might be 1884 considered in the future. 1886 Fourth, with respect to the CDNI Control interface, and in particular 1887 purging HAS chunks from a given CDN, our approach is to expect each 1888 CDN supports per-chunk content purge (e.g. purging of chunks as if 1889 they were individual content items). Optionally, a CDN may support 1890 content purge on the basis of a "Purge IDentifier (Purge-ID)" 1891 allowing the removal of all chunks related to a given Content 1892 Collection with a single reference. It is possible that this Purge- 1893 ID could be merged with the CCID discussed above for HAS Logging, or 1894 alternatively, they may remain distinct. 1896 4.8. URI Rewriting 1898 When using HTTP redirection, content URIs may be rewritten when 1899 redirection takes place within an uCDN, from an uCDN to a dCDN, and 1900 within the dCDN. In the case of cascaded CDNs, content URIs may be 1901 rewritten at every CDN hop (e.g., between the uCDN and the dCDN 1902 acting as the transit CDN, and between the transit CDN and the dCDN 1903 serving the request. The content URI used between any uCDN/dCDN pair 1904 becomes a common handle that can be referred without ambiguity by 1905 both CDNs in all their inter-CDN communications. This handle allows 1906 the uCDN and dCDN to correlate information exchanged using other CDNI 1907 interfaces in both the downstream direction (e.g., when using the MI) 1908 and the upstream direction (e.g., when using the LI). 1910 Consider the simple case of a single uCDN/dCDN pair using HTTP 1911 redirection. We introduce the following terminology for content URIs 1912 to simplify the discussion: 1914 "u-URI" represents a content URI in a request presented to the 1915 uCDN; 1917 "ud-URI" is a content URI acting as the common handle across uCDN 1918 and dCDN for requests redirected by the uCDN to a specific dCDN; 1920 "d-URI" represents a content URI in a request made within the 1921 delegate dCDN. 1923 In our simple pair-wise example, the "ud-URI" effectively becomes the 1924 handle that the uCDN/dCDN pair use to correlate all CDNI information. 1925 In particular, for a given pair of CDNs executing the HTTP 1926 redirection, the uCDN needs to map the u-URI to the ud-URI handle for 1927 all MI message exchanges, while the dCDN needs to map the d-URI to 1928 the ud-URI handle for all LI message exchanges. 1930 In the case of a cascaded CDNs, the transit CDN will rewrite the 1931 content URI when redirecting to the dCDN, thereby establishing a new 1932 handle between the transit CDN and the dCDN, that is different from 1933 the handle between the uCDN and transit CDN. It is the 1934 responsibility of the transit CDN to manage its mapping across 1935 handles so the right handle for all pairs of CDNs is always used in 1936 its CDNI communication. 1938 In summary, all CDNI interfaces between a given pair of CDNs need to 1939 always use the "ud-URI" handle for that specific CDN pair as their 1940 content URI reference. 1942 5. Deployment Models 1944 In this section we describe a number of possible deployment models 1945 that may be achieved using the CDNI interfaces described above. We 1946 note that these models are by no means exhaustive, and that many 1947 other models may be possible. 1949 Although the reference model of Figure 1 shows all CDN functions on 1950 each side of the CDNI interface, deployments can rely on entities 1951 that are involved in any subset of these functions, and therefore 1952 only support the relevant subset of CDNI interfaces. As already 1953 noted in Section 3, effective CDNI deployments can be built without 1954 necessarily implementing all the interfaces. Some examples of such 1955 deployments are shown below. 1957 Note that, while we refer to upstream and downstream CDNs, this 1958 distinction applies to specific content items and transactions. That 1959 is, a given CDN may be upstream for some transactions and downstream 1960 for others, depending on many factors such as location of the 1961 requesting client and the particular piece of content requested. 1963 5.1. Meshed CDNs 1965 Although the reference model illustrated in Figure 1 shows a 1966 unidirectional CDN interconnection with a single uCDN and a single 1967 dCDN, any arbitrary CDNI meshing can be built from this, such as the 1968 example meshing illustrated in Figure 11. (Support for arbitrary 1969 meshing may or may not be in the initial scope for the working group, 1970 but the model allows for it.) 1972 ------------- ----------- 1973 / CDN A \<==CDNI===>/ CDN B \ 1974 \ / \ / 1975 ------------- ----------- 1976 /\ \\ /\ 1977 || \\ || 1978 CDNI \==CDNI===\\ CDNI 1979 || \\ || 1980 \/ \/ \/ 1981 ------------- ----------- 1982 / CDN C \===CDNI===>/ CDN D \ 1983 \ / \ / 1984 ------------- ----------- 1985 /\ 1986 || 1987 CDNI 1988 || 1989 \/ 1990 ------------- 1991 / CDN E \ 1992 \ / 1993 ------------- 1995 ===> CDNI interfaces, with right-hand side CDN acting as dCDN 1996 to left-hand side CDN 1998 <==> CDNI interfaces, with right-hand side CDN acting as dCDN 1999 to left-hand side CDN and with left-hand side CDN acting 2000 as dCDN to right-hand side CDN 2002 Figure 11: CDNI Deployment Model: CDN Meshing Example 2004 5.2. CSP combined with CDN 2006 Note that our terminology refers to functional roles and not economic 2007 or business roles. That is, a given organization may be operating as 2008 both a CSP and a fully fledged uCDN when we consider the functions 2009 performed, as illustrated in Figure 12. 2011 ##################################### ################## 2012 # # # # 2013 # Organization A # # Organization B # 2014 # # # # 2015 # -------- ------------- # # ----------- # 2016 # / CSP \ / uCDN \ # # / dCDN \ # 2017 # | | | +----+ | # # | +----+ | # 2018 # | | | | C | | # # | | C | | # 2019 # | | | +----+ | # # | +----+ | # 2020 # | | | +----+ | # # | +----+ | # 2021 # | | | | L | | # # | | L | | # 2022 # | |*****| +----+ |===CDNI===>| +----+ | # 2023 # | | | +----+ | # # | +----+ | # 2024 # | | | | RR | | # # | | RR | | # 2025 # | | | +----+ | # # | +----+ | # 2026 # | | | +----+ | # # | +----+ | # 2027 # | | | | D | | # # | | D | | # 2028 # | | | +----+ | # # | +----+ | # 2029 # \ / \ / # # \ / # 2030 # -------- ------------- # # ----------- # 2031 # # # # 2032 ##################################### ################## 2034 ===> CDNI interfaces, with right-hand side CDN acting as dCDN 2035 to left-hand side CDN 2036 **** interfaces outside the scope of CDNI 2037 C Control component of the CDN 2038 L Logging component of the CDN 2039 RR Request Routing component of the CDN 2040 D Distribution component of the CDN 2042 Figure 12: CDNI Deployment Model: Organization combining CSP & uCDN 2044 5.3. CSP using CDNI Request Routing Interface 2045 As another example, a content provider organization may choose to run 2046 its own request routing function as a way to select among multiple 2047 candidate CDN providers; In this case the content provider may be 2048 modeled as the combination of a CSP and of a special, restricted case 2049 of a CDN. In that case, as illustrated in Figure 13, the CDNI 2050 Request Routing interfaces can be used between the restricted CDN 2051 operated by the content provider Organization and the CDN operated by 2052 the full CDN organization acting as a dCDN in the request routing 2053 control plane. Interfaces outside the scope of the CDNI work can be 2054 used between the CSP functional entities of the content provider 2055 organization and the CDN operated by the full CDN organization acting 2056 as a uCDN) in the CDNI control planes other than the request routing 2057 plane (i.e. Control, Distribution, Logging). 2059 ##################################### ################## 2060 # # # # 2061 # Organization A # # Organization B # 2062 # # # # 2063 # -------- ------------- # # ----------- # 2064 # / CSP \ / uCDN(RR) \ # # / dCDN(RR) \ # 2065 # | | | +----+ | # # | +----+ | # 2066 # | |*****| | RR |==========CDNI=====>| RR | | # 2067 # | | | +----+ | # RR # | +----+ | # 2068 # | | \ / # # | | # 2069 # | | ------------- # # |uCDN(C,L,D)| # 2070 # | | # # | +----+ | # 2071 # | | # # | | C | | # 2072 # | |*******************************| +----+ | # 2073 # | | # # | +----+ | # 2074 # | | # # | | L | | # 2075 # | | # # | +----+ | # 2076 # | | # # | +----+ | # 2077 # | | # # | | D | | # 2078 # | | # # | +----+ | # 2079 # \ / # # \ / # 2080 # -------- # # ----------- # 2081 # # # # 2082 ##################################### ################## 2084 ===> CDNI Request Routing Interface 2085 **** interfaces outside the scope of CDNI 2087 Figure 13: CDNI Deployment Model: Organization combining CSP and 2088 partial CDN 2090 5.4. CDN Federations and CDN Exchanges 2091 There are two additional concepts related to, but distinct from CDN 2092 Interconnection. The first is CDN Federation. Our view is that CDNI 2093 is the more general concept, involving two or more CDNs serving 2094 content to each other's users, while federation implies a multi- 2095 lateral interconnection arrangement, but other CDN interconnection 2096 agreements are also possible (e.g., symmetric bilateral, asymmetric 2097 bilateral). An important conclusion is that CDNI technology should 2098 not presume (or bake in) a particular interconnection agreement, but 2099 should instead be general enough to permit alternative 2100 interconnection arrangements to evolve. 2102 The second concept often used in the context of CDN Federation is CDN 2103 Exchange-\u002Da third party broker or exchange that is used to 2104 facilitate a CDN federation. Our view is that a CDN exchange offers 2105 valuable machinery to scale the number of CDN operators involved in a 2106 multi-lateral (federated) agreement, but that this machinery is built 2107 on top of the core CDNI interconnection mechanisms. For example, as 2108 illustrated in Figure 14, the exchange might aggregate and 2109 redistribute information about each CDN footprint and capacity, as 2110 well as collect, filter, and redistribute traffic logs that each 2111 participant needs for interconnection settlement, but inter-CDN 2112 request routing, inter-CDN content distribution (including inter-CDN 2113 acquisition) and inter-CDN control which fundamentally involve a 2114 direct interaction between an upstream CDN and a downstream 2115 CDN-\u002Doperate exactly as in a pair-wise peering arrangement. 2116 Turning to Figure 14, we observe that in this example: 2118 o each CDN supports a direct CDNI Control interface to every other 2119 CDN 2121 o each CDN supports a direct CDNI Metadata interface to every other 2122 CDN 2124 o each CDN supports a CDNI Logging interface with the CDN Exchange 2126 o each CDN supports both a CDNI Request Routing interfaces with the 2127 CDN Exchange (for aggregation and redistribution of dynamic CDN 2128 footprint discovery information) and a direct RI to every other 2129 CDN (for actual request redirection). 2131 ---------- --------- 2132 / CDN A \ / CDN B \ 2133 | +----+ | | +----+ | 2134 //========>| C |<==============CDNI============>| C |<==========\\ 2135 || | +----+ | C | +----+ | || 2136 || | +----+ | | +----+ | || 2137 || //=====>| D |<==============CDNI============>| D |<=======\\ || 2138 || || | +----+ | M | +----+ | || || 2139 || || | | /------------\ | | || || 2140 || || | +----+ | | +--+ CDN Ex| | +----+ | || || 2141 || || //==>| RR |<===CDNI==>|RR|<=======CDNI====>| RR |<====\\ || || 2142 || || || | +----+ | RR | +--+ | RR | +----+ | || || || 2143 || || || | | | /\ | | | || || || 2144 || || || | +----+ | | || +---+ | | +----+ | || || || 2145 || || || | | L |<===CDNI=======>| L |<=CDNI====>| L | | || || || 2146 || || || | +----+ | L | || +---+ | L | +----+ | || || || 2147 || || || \ / \ || /\ / \ / || || || 2148 || || || ----------- --||----||-- ----------- || || || 2149 || || || || || || || || 2150 || || || CDNI RR || || || || 2151 || || || || CDNI L || || || 2152 || || || || || || || || 2153 || || || ---||----||---- || || || 2154 || || || / \/ || \ || || || 2155 || || || | +----+ || | || || || 2156 || || \\=====CDNI==========>| RR |<=============CDNI========// || || 2157 || || RR | +----+ \/ | RR || || 2158 || || | +----+ | || || 2159 || || | | L | | || || 2160 || || | +----+ | || || 2161 || || | +----+ | || || 2162 || \\=======CDNI===========>| D |<=============CDNI===========// || 2163 || M | +----+ | M || 2164 || | +----+ | || 2165 \\==========CDNI===========>| C |<=============CDNI==============// 2166 C | +----+ | C 2167 \ CDN C / 2168 -------------- 2170 <=CDNI RR=> CDNI Request Routing Interface 2171 <=CDNI M==> CDNI Metadata Interface 2172 <=CDNI C==> CDNI Control Interface 2173 <=CDNI L==> CDNI Logging Interface 2175 Figure 14: CDNI Deployment Model: CDN Exchange 2177 Note that a CDN exchange may alternatively support a different set of 2178 functionality (e.g. Logging only, or Logging and full request 2179 routing, or all the functionality of a CDN including content 2180 distribution). All these options are expected to be allowed by the 2181 IETF CDNI specifications. 2183 6. Trust Model 2185 There are a number of trust issues that need to be addressed by a 2186 CDNI solution. Many of them are in fact similar or identical to 2187 those in a simple CDN without interconnection. In a standard CDN 2188 environment (without CDNI), the CSP places a degree of trust in a 2189 single CDN operator to perform many functions. The CDN is trusted to 2190 deliver content with appropriate quality of experience for the end 2191 user. The CSP trusts the CDN operator not to corrupt or modify the 2192 content. The CSP often relies on the CDN operator to provide 2193 reliable accounting information regarding the volume of delivered 2194 content. The CSP may also trust the CDN operator to perform actions 2195 such as timely invalidation of content and restriction of access to 2196 content based on certain criteria such as location of the user and 2197 time of day, and to enforce per-request authorization performed by 2198 the CSP using techniques such as URI signing. 2200 A CSP also places trust in the CDN not to distribute any information 2201 that is confidential to the CSP (e.g., how popular a given piece of 2202 content is) or confidential to the end user (e.g., which content has 2203 been watched by which user). 2205 A CSP does not necessarily have to place complete trust in a CDN. A 2206 CSP will in some cases take steps to protect its content from 2207 improper distribution by a CDN, e.g. by encrypting it and 2208 distributing keys in some out of band way. A CSP also depends on 2209 monitoring (possibly by third parties) and reporting to verify that 2210 the CDN has performed adequately. A CSP may use techniques such as 2211 client-based metering to verify that accounting information provided 2212 by the CDN is reliable. HTTP conditional requests may be used to 2213 provide the CSP with some checks on CDN operation. In other words, 2214 while a CSP may trust a CDN to perform some functions in the short 2215 term, the CSP is able in most cases to verify whether these actions 2216 have been performed correctly and to take action (such as moving the 2217 content to a different CDN) if the CDN does not live up to 2218 expectations. 2220 The main trust issue raised by CDNI is that it introduces transitive 2221 trust. A CDN that has a direct relationship with a CSP can now 2222 "outsource" the delivery of content to another (downstream) CDN. 2223 That CDN may in term outsource delivery to yet another downstream 2224 CDN, and so on. 2226 The top level CDN in such a chain of delegation is responsible for 2227 ensuring that the requirements of the CSP are met. Failure to do so 2228 is presumably just as serious as in the traditional single CDN case. 2229 Hence, an upstream CDN is essentially trusting a downstream CDN to 2230 perform functions on its behalf in just the same way as a CSP trusts 2231 a single CDN. Monitoring and reporting can similarly be used to 2232 verify that the downstream CDN has performed appropriately. However, 2233 the introduction of multiple CDNs in the path between CSP and end 2234 user complicates the picture. For example, third party monitoring of 2235 CDN performance (or other aspects of operation, such as timely 2236 invalidation) might be able to identify the fact that a problem 2237 occurred somewhere in the chain but not point to the particular CDN 2238 at fault. 2240 In summary, we assume that an upstream CDN will invest a certain 2241 amount of trust in a downstream CDN, but that it will verify that the 2242 downstream CDN is performing correctly, and take corrective action 2243 (including potentially breaking off its relationship with that CDN) 2244 if behavior is not correct. We do not expect that the trust 2245 relationship between a CSP and its "top level" CDN will differ 2246 significantly from that found today in single CDN situations. 2247 However, it does appear that more sophisticated tools and techniques 2248 for monitoring CDN performance and behavior will be required to 2249 enable the identification of the CDN at fault in a particular 2250 delivery chain. 2252 We expect that the detailed designs for the specific interfaces for 2253 CDNI will need to take the transitive trust issues into account. For 2254 example, explicit confirmation that some action (such as content 2255 removal) has taken place in a downstream CDN may help to mitigate 2256 some issues of transitive trust. 2258 7. IANA Considerations 2260 This memo includes no request to IANA. 2262 8. Security Considerations 2264 While there is a variety of security issues introduced by a single 2265 CDN, we are concerned here specifically with the additional issues 2266 that arise when CDNs are interconnected. For example, when a single 2267 CDN has the ability to distribute content on behalf of a CSP, there 2268 may be concerns that such content could be distributed to parties who 2269 are not authorized to receive it, and there are mechanisms to deal 2270 with such concerns. Our focus in this section is on how CDN 2271 interconnection introduces new security issues not found in the 2272 single CDN case. 2274 Many of the security issues that arise in CDNI are related to the 2275 transitivity of trust (or lack thereof) described in Section 6. As 2276 noted above, the design of the various interfaces for CDNI must take 2277 account of the additional risks posed by the fact that a CDN with 2278 whom a CSP has no direct relationship is now potentially distributing 2279 content for that CSP. The mechanisms used to mitigate these risks 2280 may be similar to those used in the single CDN case, but their 2281 suitability in this more complex environment must be validated. 2283 Another concern that arises in any CDN is that information about the 2284 behavior of users (what content they access, how much content they 2285 consume, etc.) may be gathered by the CDN. This risk certainly 2286 exists in inter-connected CDNs, but it should be possible to apply 2287 the same techniques to mitigate it as in the single CDN case. 2289 CDNs today offer a variety of means to control access to content, 2290 such as time-of-day restrictions, geo-blocking, and URI signing. 2291 These mechanisms must continue to function in CDNI environments, and 2292 this consideration is likely to affect the design of certain CDNI 2293 interfaces (e.g. metadata, request routing.) 2295 Just as with a single CDN, each peer CDN must ensure that it is not 2296 used as an "open proxy" to deliver content on behalf of a malicious 2297 CSP. Whereas a single CDN typically addresses this problem by having 2298 CSPs explicitly register content (or origin servers) that is to be 2299 served, simply propagating this information to peer downstream CDNs 2300 may be problematic because it reveals more information than the 2301 upstream CDN is willing to specify. (To this end, the content 2302 acquisition step in the earlier examples force the dCDN to retrieve 2303 content from the uCDN rather than go directly to the origin server.) 2305 There are several approaches to this problem. One is for the uCDN to 2306 encoded a signed token generated from a shared secret in each URL 2307 routed to a dCDN, and for the dCDN to validate the request based on 2308 this token. Another one is to have each upstream CDN advertise the 2309 set of CDN-Domains they serve, where the downstream CDN checks each 2310 request against this set before caching and delivering the associated 2311 object. Although straightforward, this approach requires operators 2312 to reveal additional information, which may or may not be an issue. 2314 8.1. Security of CDNI Interfaces 2316 It is noted in [I-D.ietf-cdni-requirements] that all CDNI interfaces 2317 must be able to operate securely over insecure IP networks. Since it 2318 is expected that the CDNI interfaces will be implemented using 2319 existing application protocols such as HTTP or XMPP, we also expect 2320 that the security mechanisms available to those protocols may be used 2321 by the CDNI interfaces. Details of how these interfaces are secured 2322 will be specified in the relevant interface documents. 2324 8.2. Digital Rights Management 2326 Issues of digital rights management (DRM, also sometimes called 2327 digital restrictions management) is often employed for content 2328 distributed via CDNs. In general, DRM relies on the CDN to 2329 distribute encrypted content, with decryption keys distributed to 2330 users by some other means (e.g. directly from the CSP to the end 2331 user.) For this reason, DRM is considered out of scope for the CDNI 2332 WG RFC 6707 and does not introduce additional security issues for 2333 CDNI. 2335 9. Contributors 2337 The following individuals contributed to this document: 2339 o Ray van Brandenburg 2341 o Matt Caulfield 2343 o Francois le Faucheur 2345 o Aaron Falk 2347 o David Ferguson 2349 o John Hartman 2351 o Ben Niven-Jenkins 2353 o Kent Leung 2355 10. Acknowledgements 2357 We thank Huw Jones for helpful input to the draft. 2359 11. Informative References 2361 [I-D.bertrand-cdni-logging] 2362 Bertrand, G., Emile, S., Peterkofsky, R., Faucheur, F., 2363 and P. Grochocki, "CDNI Logging Interface", draft- 2364 bertrand-cdni-logging-02 (work in progress), October 2012. 2366 [I-D.ietf-appsawg-http-forwarded] 2367 Petersson, A. and M. Nilsson, "Forwarded HTTP Extension", 2368 draft-ietf-appsawg-http-forwarded-10 (work in progress), 2369 October 2012. 2371 [I-D.ietf-cdni-metadata] 2372 Niven-Jenkins, B., Murray, R., Watson, G., Caulfield, M., 2373 Leung, K., and K. Ma, "CDN Interconnect Metadata", draft- 2374 ietf-cdni-metadata-03 (work in progress), October 2013. 2376 [I-D.ietf-cdni-requirements] 2377 Leung, K. and Y. Lee, "Content Distribution Network 2378 Interconnection (CDNI) Requirements", draft-ietf-cdni- 2379 requirements-12 (work in progress), November 2013. 2381 [I-D.lefaucheur-cdni-logging-delivery] 2382 Faucheur, F., Viveganandhan, M., and K. Leung, "CDNI 2383 Logging Formats for HTTP and HTTP Adaptive Streaming 2384 Deliveries", draft-lefaucheur-cdni-logging-delivery-01 2385 (work in progress), July 2012. 2387 [I-D.murray-cdni-triggers] 2388 Murray, R. and B. Niven-Jenkins, "CDNI Control Interface / 2389 Triggers", draft-murray-cdni-triggers-03 (work in 2390 progress), April 2013. 2392 [I-D.seedorf-alto-for-cdni] 2393 Seedorf, J., "ALTO for CDNi Request Routing", draft- 2394 seedorf-alto-for-cdni-00 (work in progress), October 2011. 2396 [I-D.spp-cdni-rr-foot-cap-semantics] 2397 Seedorf, J., Peterson, J., Previdi, S., Brandenburg, R., 2398 and K. Ma, "CDNI Request Routing: Footprint and 2399 Capabilities Semantics", draft-spp-cdni-rr-foot-cap- 2400 semantics-04 (work in progress), February 2013. 2402 [I-D.vandergaast-edns-client-subnet] 2403 Contavalli, C., Gaast, W., Leach, S., and E. Lewis, 2404 "Client Subnet in DNS Requests", draft-vandergaast-edns- 2405 client-subnet-02 (work in progress), July 2013. 2407 [RFC3466] Day, M., Cain, B., Tomlinson, G., and P. Rzewski, "A Model 2408 for Content Internetworking (CDI)", RFC 3466, February 2409 2003. 2411 [RFC5693] Seedorf, J. and E. Burger, "Application-Layer Traffic 2412 Optimization (ALTO) Problem Statement", RFC 5693, October 2413 2009. 2415 [RFC6707] Niven-Jenkins, B., Le Faucheur, F., and N. Bitar, "Content 2416 Distribution Network Interconnection (CDNI) Problem 2417 Statement", RFC 6707, September 2012. 2419 [RFC6770] Bertrand, G., Stephan, E., Burbridge, T., Eardley, P., Ma, 2420 K., and G. Watson, "Use Cases for Content Delivery Network 2421 Interconnection", RFC 6770, November 2012. 2423 [RFC6983] van Brandenburg, R., van Deventer, O., Le Faucheur, F., 2424 and K. Leung, "Models for HTTP-Adaptive-Streaming-Aware 2425 Content Distribution Network Interconnection (CDNI)", RFC 2426 6983, July 2013. 2428 Authors' Addresses 2430 Larry Peterson (editor) 2431 Akamai Technologies, Inc. 2432 8 Cambridge Center 2433 Cambridge, MA 02142 2434 USA 2436 Email: lapeters@akamai.com 2438 Bruce Davie 2439 VMware, Inc. 2440 3401 Hillview Ave. 2441 Palo Alto, CA 94304 2442 USA 2444 Email: bdavie@vmware.com