idnits 2.17.1 draft-ietf-dime-ovli-10.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- No issues found here. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (August 19, 2015) is 3171 days in the past. Is this intentional? Checking references for intended status: Proposed Standard ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) == Unused Reference: 'I-D.ietf-dime-e2e-sec-req' is defined on line 1566, but no explicit reference was found in the text ** Obsolete normative reference: RFC 5226 (Obsoleted by RFC 8126) == Outdated reference: A later version (-05) exists of draft-ietf-dime-e2e-sec-req-01 -- Obsolete informational reference (is this intentional?): RFC 4006 (Obsoleted by RFC 8506) Summary: 1 error (**), 0 flaws (~~), 3 warnings (==), 2 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Diameter Maintenance and Extensions (DIME) J. Korhonen, Ed. 3 Internet-Draft Broadcom 4 Intended status: Standards Track S. Donovan, Ed. 5 Expires: February 20, 2016 B. Campbell 6 Oracle 7 L. Morand 8 Orange Labs 9 August 19, 2015 11 Diameter Overload Indication Conveyance 12 draft-ietf-dime-ovli-10.txt 14 Abstract 16 This specification defines a base solution for Diameter overload 17 control, referred to as Diameter Overload Indication Conveyance 18 (DOIC). 20 Status of This Memo 22 This Internet-Draft is submitted in full conformance with the 23 provisions of BCP 78 and BCP 79. 25 Internet-Drafts are working documents of the Internet Engineering 26 Task Force (IETF). Note that other groups may also distribute 27 working documents as Internet-Drafts. The list of current Internet- 28 Drafts is at http://datatracker.ietf.org/drafts/current/. 30 Internet-Drafts are draft documents valid for a maximum of six months 31 and may be updated, replaced, or obsoleted by other documents at any 32 time. It is inappropriate to use Internet-Drafts as reference 33 material or to cite them other than as "work in progress." 35 This Internet-Draft will expire on February 20, 2016. 37 Copyright Notice 39 Copyright (c) 2015 IETF Trust and the persons identified as the 40 document authors. All rights reserved. 42 This document is subject to BCP 78 and the IETF Trust's Legal 43 Provisions Relating to IETF Documents 44 (http://trustee.ietf.org/license-info) in effect on the date of 45 publication of this document. Please review these documents 46 carefully, as they describe your rights and restrictions with respect 47 to this document. Code Components extracted from this document must 48 include Simplified BSD License text as described in Section 4.e of 49 the Trust Legal Provisions and are provided without warranty as 50 described in the Simplified BSD License. 52 Table of Contents 54 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 55 2. Terminology and Abbreviations . . . . . . . . . . . . . . . . 3 56 3. Conventions Used in This Document . . . . . . . . . . . . . . 5 57 4. Solution Overview . . . . . . . . . . . . . . . . . . . . . . 5 58 4.1. Piggybacking . . . . . . . . . . . . . . . . . . . . . . 6 59 4.2. DOIC Capability Announcement . . . . . . . . . . . . . . 7 60 4.3. DOIC Overload Condition Reporting . . . . . . . . . . . . 9 61 4.4. DOIC Extensibility . . . . . . . . . . . . . . . . . . . 11 62 4.5. Simplified Example Architecture . . . . . . . . . . . . . 11 63 5. Solution Procedures . . . . . . . . . . . . . . . . . . . . . 12 64 5.1. Capability Announcement . . . . . . . . . . . . . . . . . 12 65 5.1.1. Reacting Node Behavior . . . . . . . . . . . . . . . 13 66 5.1.2. Reporting Node Behavior . . . . . . . . . . . . . . . 13 67 5.1.3. Agent Behavior . . . . . . . . . . . . . . . . . . . 14 68 5.2. Overload Report Processing . . . . . . . . . . . . . . . 15 69 5.2.1. Overload Control State . . . . . . . . . . . . . . . 15 70 5.2.2. Reacting Node Behavior . . . . . . . . . . . . . . . 19 71 5.2.3. Reporting Node Behavior . . . . . . . . . . . . . . . 20 72 5.3. Protocol Extensibility . . . . . . . . . . . . . . . . . 22 73 6. Loss Algorithm . . . . . . . . . . . . . . . . . . . . . . . 22 74 6.1. Overview . . . . . . . . . . . . . . . . . . . . . . . . 23 75 6.2. Reporting Node Behavior . . . . . . . . . . . . . . . . . 23 76 6.3. Reacting Node Behavior . . . . . . . . . . . . . . . . . 24 77 7. Attribute Value Pairs . . . . . . . . . . . . . . . . . . . . 24 78 7.1. OC-Supported-Features AVP . . . . . . . . . . . . . . . . 25 79 7.2. OC-Feature-Vector AVP . . . . . . . . . . . . . . . . . . 25 80 7.3. OC-OLR AVP . . . . . . . . . . . . . . . . . . . . . . . 25 81 7.4. OC-Sequence-Number AVP . . . . . . . . . . . . . . . . . 26 82 7.5. OC-Validity-Duration AVP . . . . . . . . . . . . . . . . 26 83 7.6. OC-Report-Type AVP . . . . . . . . . . . . . . . . . . . 26 84 7.7. OC-Reduction-Percentage AVP . . . . . . . . . . . . . . . 27 85 7.8. Attribute Value Pair flag rules . . . . . . . . . . . . . 27 86 8. Error Response Codes . . . . . . . . . . . . . . . . . . . . 28 87 9. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 28 88 9.1. AVP codes . . . . . . . . . . . . . . . . . . . . . . . . 28 89 9.2. New registries . . . . . . . . . . . . . . . . . . . . . 29 90 10. Security Considerations . . . . . . . . . . . . . . . . . . . 29 91 10.1. Potential Threat Modes . . . . . . . . . . . . . . . . . 30 92 10.2. Denial of Service Attacks . . . . . . . . . . . . . . . 31 93 10.3. Non-Compliant Nodes . . . . . . . . . . . . . . . . . . 31 94 10.4. End-to End-Security Issues . . . . . . . . . . . . . . . 32 95 11. Contributors . . . . . . . . . . . . . . . . . . . . . . . . 33 96 12. References . . . . . . . . . . . . . . . . . . . . . . . . . 33 97 12.1. Normative References . . . . . . . . . . . . . . . . . . 33 98 12.2. Informative References . . . . . . . . . . . . . . . . . 34 99 Appendix A. Issues left for future specifications . . . . . . . 34 100 A.1. Additional traffic abatement algorithms . . . . . . . . . 34 101 A.2. Agent Overload . . . . . . . . . . . . . . . . . . . . . 34 102 A.3. New Error Diagnostic AVP . . . . . . . . . . . . . . . . 35 103 Appendix B. Deployment Considerations . . . . . . . . . . . . . 35 104 Appendix C. Considerations for Applications Integrating the DOIC 105 Solution . . . . . . . . . . . . . . . . . . . . . . 35 106 C.1. Application Classification . . . . . . . . . . . . . . . 35 107 C.2. Application Type Overload Implications . . . . . . . . . 36 108 C.3. Request Transaction Classification . . . . . . . . . . . 38 109 C.4. Request Type Overload Implications . . . . . . . . . . . 38 110 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 40 112 1. Introduction 114 This specification defines a base solution for Diameter overload 115 control, referred to as Diameter Overload Indication Conveyance 116 (DOIC), based on the requirements identified in [RFC7068]. 118 This specification addresses Diameter overload control between 119 Diameter nodes that support the DOIC solution. The solution, which 120 is designed to apply to existing and future Diameter applications, 121 requires no changes to the Diameter base protocol [RFC6733] and is 122 deployable in environments where some Diameter nodes do not implement 123 the Diameter overload control solution defined in this specification. 125 A new application specification can incorporate the overload control 126 mechanism specified in this document by making it mandatory to 127 implement for the application and referencing this specification 128 normatively. It is the responsibility of the Diameter application 129 designers to define how overload control mechanisms works on that 130 application. 132 Note that the overload control solution defined in this specification 133 does not address all the requirements listed in [RFC7068]. A number 134 of overload control related features are left for future 135 specifications. See Appendix A for a list of extensions that are 136 currently being considered. 138 2. Terminology and Abbreviations 140 Abatement 142 Reaction to receipt of an overload report resulting in a reduction 143 in traffic sent to the reporting node. Abatement actions include 144 diversion and throttling. 146 Abatement Algorithm 148 An extensible method requested by reporting nodes and used by 149 reacting nodes to reduce the amount of traffic sent during an 150 occurrence of overload control. 152 Diversion 154 An overload abatement treatment where the reacting node selects 155 alternate destinations or paths for requests. 157 Host-Routed Requests 159 Requests that a reacting node knows will be served by a particular 160 host, either due to the presence of a Destination-Host Attribute 161 Value Pair (AVP), or by some other local knowledge on the part of 162 the reacting node. 164 Overload Control State (OCS) 166 Internal state maintained by a reporting or reacting node 167 describing occurrences of overload control. 169 Overload Report (OLR) 171 Overload control information for a particular overload occurrence 172 sent by a reporting node. 174 Reacting Node 176 A Diameter node that acts upon an overload report. 178 Realm-Routed Requests 180 Requests that a reacting node does not know which host will 181 service the request. 183 Reporting Node 185 A Diameter node that generates an overload report. (This may or 186 may not be the overloaded node.) 188 Throttling 190 An abatement treatment that limits the number of requests sent by 191 the reacting node. Throttling can include a Diameter Client 192 choosing to not send requests, or a Diameter Agent or Server 193 rejecting requests with appropriate error responses. In both 194 cases the result of the throttling is a permanent rejection of the 195 transaction. 197 3. Conventions Used in This Document 199 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 200 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 201 document are to be interpreted as described in RFC 2119 [RFC2119]. 203 RFC 2119 [RFC2119] interpretation does not apply for the above listed 204 words when they are not used in all-caps format. 206 4. Solution Overview 208 The Diameter Overload Information Conveyance (DOIC) solution allows 209 Diameter nodes to request other Diameter nodes to perform overload 210 abatement actions, that is, actions to reduce the load offered to the 211 overloaded node or realm. 213 A Diameter node that supports DOIC is known as a "DOIC node". Any 214 Diameter node can act as a DOIC node, including Diameter Clients, 215 Diameter Servers, and Diameter Agents. DOIC nodes are further 216 divided into "Reporting Nodes" and "Reacting Nodes." A reporting 217 node requests overload abatement by sending Overload Reports (OLR). 219 A reacting node acts upon OLRs, and performs whatever actions are 220 needed to fulfill the abatement requests included in the OLRs. A 221 Reporting node may report overload on its own behalf, or on behalf of 222 other nodes. Likewise, a reacting node may perform overload 223 abatement on its own behalf, or on behalf of other nodes. 225 A Diameter node's role as a DOIC node is independent of its Diameter 226 role. For example, Diameter Agents may act as DOIC nodes, even 227 though they are not endpoints in the Diameter sense. Since Diameter 228 enables bi-directional applications, where Diameter Servers can send 229 requests towards Diameter Clients, a given Diameter node can 230 simultaneously act as both a reporting node and a reacting node. 232 Likewise, a Diameter Agent may act as a reacting node from the 233 perspective of upstream nodes, and a reporting node from the 234 perspective of downstream nodes. 236 DOIC nodes do not generate new messages to carry DOIC related 237 information. Rather, they "piggyback" DOIC information over existing 238 Diameter messages by inserting new AVPs into existing Diameter 239 requests and responses. Nodes indicate support for DOIC, and any 240 needed DOIC parameters, by inserting an OC-Supported-Features AVP 241 (Section 7.2) into existing requests and responses. Reporting nodes 242 send OLRs by inserting OC-OLR AVPs (Section 7.3). 244 A given OLR applies to the Diameter realm and application of the 245 Diameter message that carries it. If a reporting node supports more 246 than one realm and/or application, it reports independently for each 247 combination of realm and application. Similarly, the OC-Supported- 248 Features AVP applies to the realm and application of the enclosing 249 message. This implies that a node may support DOIC for one 250 application and/or realm, but not another, and may indicate different 251 DOIC parameters for each application and realm for which it supports 252 DOIC. 254 Reacting nodes perform overload abatement according to an agreed-upon 255 abatement algorithm. An abatement algorithm defines the meaning of 256 some of the parameters of an OLR and the procedures required for 257 overload abatement. An overload abatement algorithm separates 258 Diameter requests into two sets. The first set contains the requests 259 that are to undergo overload abatement treatment of either throttling 260 or diversion. The second set contains the requests that are to be 261 given normal routing treatment. This document specifies a single 262 must-support algorithm, namely the "loss" algorithm (Section 6). 263 Future specifications may introduce new algorithms. 265 Overload conditions may vary in scope. For example, a single 266 Diameter node may be overloaded, in which case reacting nodes may 267 attempt to send requests to other destinations. On the other hand, 268 an entire Diameter realm may be overloaded, in which case such 269 attempts would do harm. DOIC OLRs have a concept of "report type" 270 (Section 7.6), where the type defines such behaviors. Report types 271 are extensible. This document defines report types for overload of a 272 specific host, and for overload of an entire realm. 274 DOIC works through non supporting Diameter Agents that properly pass 275 unknown AVPs unchanged. 277 4.1. Piggybacking 279 There is no new Diameter application defined to carry overload 280 related AVPs. The overload control AVPs defined in this 281 specification have been designed to be piggybacked on top of existing 282 application messages. This is made possible by adding the optional 283 overload control AVPs OC-OLR and OC-Supported-Features into existing 284 commands. 286 Reacting nodes indicate support for DOIC by including the OC- 287 Supported-Features AVP in all request messages originated or relayed 288 by the reacting node. 290 Reporting nodes indicate support for DOIC by including the OC- 291 Supported-Features AVP in all answer messages originated or relayed 292 by the reporting node that are in response to a request that 293 contained the OC-Supported-Features AVP. Reporting nodes may include 294 overload reports using the OC-OLR AVP in answer messages. 296 Note that the overload control solution does not have fixed server 297 and client roles. The DOIC node role is determined based on the 298 message type: whether the message is a request (i.e., sent by a 299 "reacting node") or an answer (i.e., sent by a "reporting node"). 300 Therefore, in a typical "client-server" deployment, the Diameter 301 Client may report its overload condition to the Diameter Server for 302 any Diameter Server initiated message exchange. An example of such 303 is the Diameter Server requesting a re-authentication from a Diameter 304 Client. 306 4.2. DOIC Capability Announcement 308 The DOIC solution supports the ability for Diameter nodes to 309 determine if other nodes in the path of a request support the 310 solution. This capability is referred to as DOIC Capability 311 Announcement (DCA) and is separate from Diameter Capability Exchange. 313 The DCA mechanism uses the OC-Supported-Features AVPs to indicate the 314 Diameter overload features supported. 316 The first node in the path of a Diameter request that supports the 317 DOIC solution inserts the OC-Supported-Features AVP in the request 318 message. 320 The individual features supported by the DOIC nodes are indicated in 321 the OC-Feature-Vector AVP. Any semantics associated with the 322 features will be defined in extension specifications that introduce 323 the features. 325 Note: As discussed elsewhere in the document, agents in the path 326 of the request can modify the OC-Supported-Features AVP. 328 Note: The DOIC solution must support deployments where Diameter 329 Clients and/or Diameter Servers do not support the DOIC solution. 330 In this scenario, Diameter Agents that support the DOIC solution 331 may handle overload abatement for the non-supporting Diameter 332 nodes. In this case the DOIC agent will insert the OC-Supported- 333 Features AVP in requests that do not already contain one, telling 334 the reporting node that there is a DOIC node that will handle 335 overload abatement. For transactions where there was an OC- 336 Supporting-Features AVP in the request, the agent will insert the 337 OC-Supported-Features AVP in answers, telling the reacting node 338 that there is a reporting node. 340 The OC-Feature-Vector AVP will always contain an indication of 341 support for the loss overload abatement algorithm defined in this 342 specification (see Section 6). This ensures that a reporting node 343 always supports at least one of the advertized abatement algorithms 344 received in a request messages. 346 The reporting node inserts the OC-Supported-Features AVP in all 347 answer messages to requests that contained the OC-Supported-Features 348 AVP. The contents of the reporting node's OC-Supported-Features AVP 349 indicate the set of Diameter overload features supported by the 350 reporting node. This specification defines one exception - the 351 reporting node only includes an indication of support for one 352 overload abatement algorithm, independent of the number of overload 353 abatement algorithms actually supported by the reacting node. The 354 overload abatement algorithm indicated is the algorithm that the 355 reporting node intends to use should it enter an overload condition. 356 Reacting nodes can use the indicated overload abatement algorithm to 357 prepare for possible overload reports and must use the indicated 358 overload abatement algorithm if traffic reduction is actually 359 requested. 361 Note that the loss algorithm defined in this document is a 362 stateless abatement algorithm. As a result it does not require 363 any actions by reacting nodes prior to the receipt of an overload 364 report. Stateful abatement algorithms that base the abatement 365 logic on a history of request messages sent might require reacting 366 nodes to maintain state in advance of receiving an overload report 367 to ensure that the overload reports can be properly handled. 369 While it should only be done in exceptional circumstances and not 370 during an active occurrence of overload, a reacting node that wishes 371 to transition to a different abatement algorithm can stop advertising 372 support for the algorithm indicated by the reporting node, as long as 373 support for the loss algorithm is always advertised. 375 The DCA mechanism must also allow the scenario where the set of 376 features supported by the sender of a request and by agents in the 377 path of a request differ. In this case, the agent can update the OC- 378 Supported-Features AVP to reflect the mixture of the two sets of 379 supported features. 381 Note: The logic to determine if the content of the OC-Supported- 382 Features AVP should be changed is out-of-scope for this document, 383 as is the logic to determine the content of a modified OC- 384 Supported-Features AVP. These are left to implementation 385 decisions. Care must be taken not to introduce interoperability 386 issues for downstream or upstream DOIC nodes. As such, the agent 387 must act as a fully compliant reporting node to the downstream 388 reacting node and as a fully compliant reacting node to the 389 upstream reporting node. 391 4.3. DOIC Overload Condition Reporting 393 As with DOIC capability announcement, overload condition reporting 394 uses new AVPs (Section 7.3) to indicate an overload condition. 396 The OC-OLR AVP is referred to as an overload report. The OC-OLR AVP 397 includes the type of report, a sequence number, the length of time 398 that the report is valid and abatement algorithm specific AVPs. 400 Two types of overload reports are defined in this document: host 401 reports and realm reports. 403 A report of type "HOST_REPORT" is sent to indicate the overload of a 404 specific host, identified by the Origin-Host AVP of the message 405 containing the OLR, for the application-id indicated in the 406 transaction. When receiving an OLR of type "HOST_REPORT", a reacting 407 node applies overload abatement treatment to the host-routed requests 408 identified by the overload abatement algorithm (see definition in 409 Section 2) sent for this application to the overloaded host. 411 A report of type "REALM_REPORT" is sent to indicate the overload of a 412 realm for the application-id indicated in the transaction. The 413 overloaded realm is identified by the Destination-Realm AVP of the 414 message containing the OLR. When receiving an OLR of type 415 "REALM_REPORT", a reacting node applies overload abatement treatment 416 to realm-routed requests identified by the overload abatement 417 algorithm (see definition in Section 2) sent for this application to 418 the overloaded realm. 420 This document assumes that there is a single source for realm-reports 421 for a given realm, or that if multiple nodes can send realm reports, 422 that each such node has full knowledge of the overload state of the 423 entire realm. A reacting node cannot distinguish between receiving 424 realm-reports from a single node, or from multiple nodes. 426 Note: Known issues exist if multiple sources for overload reports 427 which apply to the same Diameter entity exist. Reacting nodes 428 have no way of determining the source and, as such, will treat 429 them as coming from a single source. Variance in sequence numbers 430 between the two sources can then cause incorrect overload 431 abatement treatment to be applied for indeterminate periods of 432 time. 434 Reporting nodes are responsible for determining the need for a 435 reduction of traffic. The method for making this determination is 436 implementation specific and depends on the type of overload report 437 being generated. A host-report might be generated by tracking use of 438 resources required by the host to handle transactions for the 439 Diameter application. A realm-report generally impacts the traffic 440 sent to multiple hosts and, as such, requires tracking the capacity 441 of all servers able to handle realm-routed requests for the 442 application and realm. 444 Once a reporting node determines the need for a reduction in traffic, 445 it uses the DOIC defined AVPs to report on the condition. These AVPs 446 are included in answer messages sent or relayed by the reporting 447 node. The reporting node indicates the overload abatement algorithm 448 that is to be used to handle the traffic reduction in the OC- 449 Supported-Features AVP. The OC-OLR AVP is used to communicate 450 information about the requested reduction. 452 Reacting nodes, upon receipt of an overload report, apply the 453 overload abatement algorithm to traffic impacted by the overload 454 report. The method used to determine the requests that are to 455 receive overload abatement treatment is dependent on the abatement 456 algorithm. The loss abatement algorithm is defined in this document 457 (Section 6). Other abatement algorithms can be defined in extensions 458 to the DOIC solution. 460 Two types of overload abatement treatment are defined, diversion and 461 throttling. Reacting nodes are responsible for determining which 462 treatment is appropriate for individual requests. 464 As the conditions that lead to the generation of the overload report 465 change the reporting node can send new overload reports requesting 466 greater reduction if the condition gets worse or less reduction if 467 the condition improves. The reporting node sends an overload report 468 with a duration of zero to indicate that the overload condition has 469 ended and abatement is no longer needed. 471 The reacting node also determines when the overload report expires 472 based on the OC-Validity-Duration AVP in the overload report and 473 stops applying the abatement algorithm when the report expires. 475 Note that erroneous overload reports can be used for DoS attacks. 476 This includes the ability to indicate that a significant reduction in 477 traffic, up to and including a request for no traffic, should be sent 478 to a reporting node. As such, care should be taken to verify the 479 sender of overload reports. 481 4.4. DOIC Extensibility 483 The DOIC solution is designed to be extensible. This extensibility 484 is based on existing Diameter based extensibility mechanisms, along 485 with the DOIC capability announcement mechanism. 487 There are multiple categories of extensions that are expected. This 488 includes the definition of new overload abatement algorithms, the 489 definition of new report types and the definition of new scopes of 490 messages impacted by an overload report. 492 A DOIC node communicates supported features by including them in the 493 OC-Feature-Vector AVP, as a sub-AVP of OC-Supported-Features. Any 494 non-backwards compatible DOIC extensions define new values for the 495 OC-Feature-Vector AVP. DOIC extensions also have the ability to add 496 new AVPs to the OC-Supported-Features AVP, if additional information 497 about the new feature is required. 499 Overload reports can also be extended by adding new sub-AVPs to the 500 OC-OLR AVP, allowing reporting nodes to communicate additional 501 information about handling an overload condition. 503 If necessary, new extensions can also define new AVPs that are not 504 part of the OC-Supported-Features and OC-OLR group AVPs. It is, 505 however, recommended that DOIC extensions use the OC-Supported- 506 Features AVP and OC-OLR AVP to carry all DOIC related AVPs. 508 4.5. Simplified Example Architecture 510 Figure 1 illustrates the simplified architecture for Diameter 511 overload information conveyance. 513 Realm X Same or other Realms 514 <--------------------------------------> <----------------------> 516 +--------+ : (optional) : 517 |Diameter| : : 518 |Server A|--+ .--. : +--------+ : .--. 519 +--------+ | _( `. : |Diameter| : _( `. +--------+ 520 +--( )--:-| Agent |-:--( )--|Diameter| 521 +--------+ | ( ` . ) ) : +--------+ : ( ` . ) ) | Client | 522 |Diameter|--+ `--(___.-' : : `--(___.-' +--------+ 523 |Server B| : : 524 +--------+ : : 526 End-to-end Overload Indication 527 1) <-----------------------------------------------> 528 Diameter Application Y 530 Overload Indication A Overload Indication A' 531 2) <----------------------> <----------------------> 532 Diameter Application Y Diameter Application Y 534 Figure 1: Simplified architecture choices for overload indication 535 delivery 537 In Figure 1, the Diameter overload indication can be conveyed (1) 538 end-to-end between servers and clients or (2) between servers and 539 Diameter agent inside the realm and then between the Diameter agent 540 and the clients. 542 5. Solution Procedures 544 This section outlines the normative behavior for the DOIC solution. 546 5.1. Capability Announcement 548 This section defines DOIC Capability Announcement (DCA) behavior. 550 Note: This specification assumes that changes in DOIC node 551 capabilities are relatively rare events that occur as a result of 552 administrative action. Reacting nodes ought to minimize changes 553 that force the reporting node to change the features being used, 554 especially during active overload conditions. But even if 555 reacting nodes avoid such changes, reporting nodes still have to 556 be prepared for them to occur. For example, differing 557 capabilities between multiple reacting nodes may still force a 558 reporting node to select different features on a per-transaction 559 basis. 561 5.1.1. Reacting Node Behavior 563 A reacting node MUST include the OC-Supported-Features AVP in all 564 requests. It MAY include the OC-Feature-Vector AVP, as a sub-avp of 565 OC-Supported-Features. If it does so, it MUST indicate support for 566 the "loss" algorithm. If the reacting node is configured to support 567 features (including other algorithms) in addition to the loss 568 algorithm, it MUST indicate such support in an OC-Feature-Vector AVP. 570 An OC-Supported-Features AVP in answer messages indicates there is a 571 reporting node for the transaction. The reacting node MAY take 572 action, for example creating state for some stateful abatement 573 algorithm, based on the features indicated in the OC-Feature-Vector 574 AVP. 576 Note: The loss abatement algorithm does not require stateful 577 behavior when there is no active overload report. 579 Reacting nodes need to be prepared for the reporting node to change 580 selected algorithms. This can happen at any time, including when the 581 reporting node has sent an active overload report. The reacting node 582 can minimize the potential for changes by modifying the advertised 583 abatement algorithms sent to an overloaded reporting node to the 584 currently selected algorithm and loss (or just loss if it is the 585 currently selected algorithm). This has the effect of limiting the 586 potential change in abatement algorithm from the currently selected 587 algorithm to loss, avoiding changes to more complex abatement 588 algorithms that require state to operate properly. 590 5.1.2. Reporting Node Behavior 592 Upon receipt of a request message, a reporting node determines if 593 there is a reacting node for the transaction based on the presence of 594 the OC-Supported-Features AVP in the request message. 596 If the request message contains an OC-Supported-Features AVP then a 597 reporting node MUST include the OC-Supported-Features AVP in the 598 answer message for that transaction. 600 Note: Capability announcement is done on a per transaction basis. 601 The reporting node cannot assume that the capabilities announced 602 by a reacting node will be the same between transactions. 604 A reporting node MUST NOT include the OC-Supported-Features AVP, OC- 605 OLR AVP or any other overload control AVPs defined in extension 606 drafts in response messages for transactions where the request 607 message does not include the OC-Supported-Features AVP. Lack of the 608 OC-Supported-Features AVP in the request message indicates that there 609 is no reacting node for the transaction. 611 A reporting node knows what overload control functionality is 612 supported by the reacting node based on the content or absence of the 613 OC-Feature-Vector AVP within the OC-Supported-Features AVP in the 614 request message. 616 A reporting node MUST select a single abatement algorithm in the OC- 617 Feature-Vector AVP. The abatement algorithm selected MUST indicate 618 the abatement algorithm the reporting node wants the reacting node to 619 use when the reporting node enters an overload condition. 621 The abatement algorithm selected MUST be from the set of abatement 622 algorithms contained in the request message's OC-Feature-Vector AVP. 624 A reporting node that selects the loss algorithm may do so by 625 including the OC-Feature-Vector AVP with an explicit indication of 626 the loss algorithm, or it MAY omit OC-Feature-Vector. If it selects 627 a different algorithm, it MUST include the OC-Feature-Vector AVP with 628 an explicit indication of the selected algorithm. 630 The reporting node SHOULD indicate support for other DOIC features 631 defined in extension drafts that it supports and that apply to the 632 transaction. It does so using the OC-Feature-Vector AVP. 634 Note: Not all DOIC features will apply to all Diameter 635 applications or deployment scenarios. The features included in 636 the OC-Feature-Vector AVP are based on local reporting node 637 policy. 639 5.1.3. Agent Behavior 641 Diameter Agents that support DOIC can ensure that all messages 642 relayed by the agent contain the OC-Supported-Features AVP. 644 A Diameter Agent MAY take on reacting node behavior for Diameter 645 endpoints that do not support the DOIC solution. A Diameter Agent 646 detects that a Diameter endpoint does not support DOIC reacting node 647 behavior when there is no OC-Supported-Features AVP in a request 648 message. 650 For a Diameter Agent to be a reacting node for a non-supporting 651 Diameter endpoint, the Diameter Agent MUST include the OC-Supported- 652 Features AVP in request messages it relays that do not contain the 653 OC-Supported-Features AVP. 655 A Diameter Agent MAY take on reporting node behavior for Diameter 656 endpoints that do not support the DOIC solution. The Diameter Agent 657 MUST have visibility to all traffic destined for the non-supporting 658 host in order to become the reporting node for the Diameter endpoint. 659 A Diameter Agent detects that a Diameter endpoint does not support 660 DOIC reporting node behavior when there is no OC-Supported-Features 661 AVP in an answer message for a transaction that contained the OC- 662 Supported-Features AVP in the request message. 664 If a request already has the OC-Supported-Features AVP, a Diameter 665 agent MAY modify it to reflect the features appropriate for the 666 transaction. Otherwise, the agent relays the OC-Supported-Features 667 AVP without change. 669 For instance, if the agent supports a superset of the features 670 reported by the reacting node then the agent might choose, based 671 on local policy, to advertise that superset of features to the 672 reporting node. 674 If the Diameter Agent changes the OC-Supported-Features AVP in a 675 request message then it is likely it will also need to modify the OC- 676 Supported-Features AVP in the answer message for the transaction. A 677 Diameter Agent MAY modify the OC-Supported-Features AVP carried in 678 answer messages. 680 When making changes to the OC-Supported-Features or OC-OLR AVPs, the 681 Diameter Agent needs to ensure consistency in its behavior with both 682 upstream and downstream DOIC nodes. 684 5.2. Overload Report Processing 686 5.2.1. Overload Control State 688 Both reacting and reporting nodes maintain Overload Control State 689 (OCS) for active overload conditions. The following sections define 690 behavior associated with that OCS. 692 The contents of the OCS in the reporting node and in the reacting 693 node represent logical constructs. The actual internal physical 694 structure of the state included in the OCS is an implementation 695 decision. 697 5.2.1.1. Overload Control State for Reacting Nodes 699 A reacting node maintains the following OCS per supported Diameter 700 application: 702 o A host-type OCS entry for each Destination-Host to which it sends 703 host-type requests and 705 o A realm-type OCS entry for each Destination-Realm to which it 706 sends realm-type requests. 708 A host-type OCS entry is identified by the pair of application-id and 709 the node's DiameterIdentity. 711 A realm-type OCS entry is identified by the pair of application-id 712 and realm. 714 The host-type and realm-type OCS entries include the following 715 information (the actual information stored is an implementation 716 decision): 718 o Sequence number (as received in OC-OLR, see Section 7.3) 720 o Time of expiry (derived from OC-Validity-Duration AVP received in 721 the OC-OLR AVP and time of reception of the message carrying OC- 722 OLR AVP) 724 o Selected Abatement Algorithm (as received in the OC-Supported- 725 Features AVP) 727 o Abatement Algorithm specific input data (as received in the OC-OLR 728 AVP, for example, OC-Reduction-Percentage for the Loss abatement 729 algorithm) 731 5.2.1.2. Overload Control State for Reporting Nodes 733 A reporting node maintains OCS entries per supported Diameter 734 application, per supported (and eventually selected) Abatement 735 Algorithm and per report-type. 737 An OCS entry is identified by the tuple of Application-Id, Report- 738 Type and Abatement Algorithm and includes the following information 739 (the actual information stored is an implementation decision): 741 o Sequence number 743 o Validity Duration 745 o Expiration Time 747 o Algorithm specific input data (for example, the Reduction 748 Percentage for the Loss Abatement Algorithm) 750 5.2.1.3. Reacting Node Maintenance of Overload Control State 752 When a reacting node receives an OC-OLR AVP, it MUST determine if it 753 is for an existing or new overload condition. 755 Note: For the remainder of this section the term OLR refers to the 756 combination of the contents of the received OC-OLR AVP and the 757 abatement algorithm indicated in the received OC-Supported- 758 Features AVP. 760 When receiving an answer message with multiple OLRs of different 761 supported report types, a reacting node MUST process each received 762 OLR. 764 The OLR is for an existing overload condition if a reacting node has 765 an OCS that matches the received OLR. 767 For a host-report this means it matches the application-id and the 768 host's DiameterIdentity in an existing host OCS entry. 770 For a realm-report this means it matches the application-id and the 771 realm in an existing realm OCS entry. 773 If the OLR is for an existing overload condition then a reacting node 774 MUST determine if the OLR is a retransmission or an update to the 775 existing OLR. 777 If the sequence number for the received OLR is greater than the 778 sequence number stored in the matching OCS entry then a reacting node 779 MUST update the matching OCS entry. 781 If the sequence number for the received OLR is less than or equal to 782 the sequence number in the matching OCS entry then a reacting node 783 MUST silently ignore the received OLR. The matching OCS MUST NOT be 784 updated in this case. 786 If the reacting node determines that the sequence number has rolled 787 over then the reacting node MUST update the matching OCS entry. This 788 can be determined by recognizing that the number has changed from 789 something close to the maximum value in the OC-Sequence-Number AVP to 790 something close to the minimum value in the OC-Sequence-Number AVP. 792 If the received OLR is for a new overload condition then a reacting 793 node MUST generate a new OCS entry for the overload condition. 795 For a host-report this means a reacting node creates on OCS entry 796 with the application-id in the received message and DiameterIdentity 797 of the Origin-Host in the received message. 799 Note: This solution assumes that the Origin-Host AVP in the answer 800 message included by the reporting node is not changed along the 801 path to the reacting node. 803 For a realm-report this means a reacting node creates on OCS entry 804 with the application-id in the received message and realm of the 805 Origin-Realm in the received message. 807 If the received OLR contains a validity duration of zero ("0") then a 808 reacting node MUST update the OCS entry as being expired. 810 Note: It is not necessarily appropriate to delete the OCS entry, 811 as there is recommended behavior that the reacting node slowly 812 returns to full traffic when ending an overload abatement period. 814 The reacting node does not delete an OCS when receiving an answer 815 message that does not contain an OC-OLR AVP (i.e., absence of OLR 816 means "no change"). 818 5.2.1.4. Reporting Node Maintenance of Overload Control State 820 A reporting node SHOULD create a new OCS entry when entering an 821 overload condition. 823 Note: If a reporting node knows through absence of the OC- 824 Supported-Features AVP in received messages that there are no 825 reacting nodes supporting DOIC then the reporting node can choose 826 to not create OCS entries. 828 When generating a new OCS entry the sequence number SHOULD be set to 829 zero ("0"). 831 When generating sequence numbers for new overload conditions, the new 832 sequence number MUST be greater than any sequence number in an active 833 (unexpired) overload report for the same application and report-type 834 previously sent by the reporting node. This property MUST hold over 835 a reboot of the reporting node. 837 Note: One way of addressing this over a reboot of a reporting node 838 is to use a time stamp for the first overload condition that 839 occurs after the report and to start using sequences beginning 840 with zero for subsequent overload conditions. 842 A reporting node MUST update an OCS entry when it needs to adjust the 843 validity duration of the overload condition at reacting nodes. 845 For instance, if a reporting node wishes to instruct reacting 846 nodes to continue overload abatement for a longer period of time 847 than originally communicated. This also applies if the reporting 848 node wishes to shorten the period of time that overload abatement 849 is to continue. 851 A reporting node MUST update an OCS entry when it wishes to adjust 852 any abatement algorithm specific parameters, including, for example, 853 the reduction percentage used for the Loss abatement algorithm. 855 For instance, if a reporting node wishes to change the reduction 856 percentage either higher, if the overload condition has worsened, 857 or lower, if the overload condition has improved, then the 858 reporting node would update the appropriate OCS entry. 860 A reporting node MUST increment the sequence number associated with 861 the OCS entry anytime the contents of the OCS entry are changed. 862 This will result in a new sequence number being sent to reacting 863 nodes, instructing reacting nodes to process the OC-OLR AVP. 865 A reporting node SHOULD update an OCS entry with a validity duration 866 of zero ("0") when the overload condition ends. 868 Note: If a reporting node knows that the OCS entries in the 869 reacting nodes are near expiration then the reporting node might 870 decide not to send an OLR with a validity duration of zero. 872 A reporting node MUST keep an OCS entry with a validity duration of 873 zero ("0") for a period of time long enough to ensure that any non- 874 expired reacting node's OCS entry created as a result of the overload 875 condition in the reporting node is deleted. 877 5.2.2. Reacting Node Behavior 879 When a reacting node sends a request it MUST determine if that 880 request matches an active OCS. 882 If the request matches an active OCS then the reacting node MUST use 883 the overload abatement algorithm indicated in the OCS to determine if 884 the request is to receive overload abatement treatment. 886 For the Loss abatement algorithm defined in this specification, see 887 Section 6 for the overload abatement algorithm logic applied. 889 If the overload abatement algorithm selects the request for overload 890 abatement treatment then the reacting node MUST apply overload 891 abatement treatment on the request. The abatement treatment applied 892 depends on the context of the request. 894 If diversion abatement treatment is possible (i.e., a different path 895 for the request can be selected where the overloaded node is not part 896 of the different path), then the reacting node SHOULD apply diversion 897 abatement treatment to the request. The reacting node MUST apply 898 throttling abatement treatment to requests identified for abatement 899 treatment when diversion treatment is not possible or was not 900 applied. 902 Note: This only addresses the case where there are two defined 903 abatement treatments, diversion and throttling. Any extension 904 that defines a new abatement treatment must also define the 905 interaction of the new abatement treatment with existing 906 treatments. 908 If the overload abatement treatment results in throttling of the 909 request and if the reacting node is an agent then the agent MUST send 910 an appropriate error as defined in Section 8. 912 Diameter endpoints that throttle requests need to do so according to 913 the rules of the client application. Those rules will vary by 914 application, and are beyond the scope of this document. 916 In the case that the OCS entry indicated no traffic was to be sent to 917 the overloaded entity and the validity duration expires then overload 918 abatement associated with the overload report MUST be ended in a 919 controlled fashion. 921 5.2.3. Reporting Node Behavior 923 If there is an active OCS entry then a reporting node SHOULD include 924 the OC-OLR AVP in all answers to requests that contain the OC- 925 Supported-Features AVP and that match the active OCS entry. 927 Note: A request matches if the application-id in the request 928 matches the application-id in any active OCS entry and if the 929 report-type in the OCS entry matches a report-type supported by 930 the reporting node as indicated in the OC-Supported-Features AVP. 932 The contents of the OC-OLR AVP depend on the selected algorithm. 934 A reporting node MAY choose to not resend an overload report to a 935 reacting node if it can guarantee that this overload report is 936 already active in the reacting node. 938 Note: In some cases (e.g., when there are one or more agents in 939 the path between reporting and reacting nodes, or when overload 940 reports are discarded by reacting nodes) a reporting node may not 941 be able to guarantee that the reacting node has received the 942 report. 944 A reporting node MUST NOT send overload reports of a type that has 945 not been advertised as supported by the reacting node. 947 Note: A reacting node implicitly advertises support for the host 948 and realm report types by including the OC-Supported-Features AVP 949 in the request. Support for other report types will be explicitly 950 indicated by new feature bits in the OC-Feature-Vector AVP. 952 A reporting node SHOULD explicitly indicate the end of an overload 953 occurrence by sending a new OLR with OC-Validity-Duration set to a 954 value of zero ("0"). The reporting node SHOULD ensure that all 955 reacting nodes receive the updated overload report. 957 A reporting node MAY rely on the OC-Validity-Duration AVP values for 958 the implicit overload control state cleanup on the reacting node. 960 Note: All OLRs sent have an expiration time calculated by adding 961 the validity-duration contained in the OLR to the time the message 962 was sent. Transit time for the OLR can be safely ignored. The 963 reporting node can ensure that all reacting nodes have received 964 the OLR by continuing to send it in answer messages until the 965 expiration time for all OLRs sent for that overload condition have 966 expired. 968 When a reporting node sends an OLR, it effectively delegates any 969 necessary throttling to downstream nodes. If the reporting node also 970 locally throttles the same set of messages, the overall number of 971 throttled requests may be higher than intended. Therefore, before 972 applying local message throttling, a reporting node needs to check if 973 these messages match existing OCS entries, indicating that these 974 messages have survived throttling applied by downstream nodes that 975 have received the related OLR. 977 However, even if the set of messages match existing OCS entries, the 978 reporting node can still apply other abatement methods such as 979 diversion. The reporting node might also need to throttle requests 980 for reasons other than overload. For example, an agent or server 981 might have a configured rate limit for each client, and throttle 982 requests that exceed that limit, even if such requests had already 983 been candidates for throttling by downstream nodes. The reporting 984 node also has the option to send new OLRs requesting greater 985 reductions in traffic, reducing the need for local throttling. 987 A reporting node SHOULD decrease requested overload abatement 988 treatment in a controlled fashion to avoid oscillations in traffic. 990 For example, it might wait some period of time after overload ends 991 before terminating the OLR, or it might send a series of OLRs 992 indicating progressively less overload severity. 994 5.3. Protocol Extensibility 996 The DOIC solution can be extended. Types of potential extensions 997 include new traffic abatement algorithms, new report types or other 998 new functionality. 1000 When defining a new extension that requires new normative behavior, 1001 the specification must define a new feature for the OC-Feature- 1002 Vector. This feature bit is used to communicate support for the new 1003 feature. 1005 The extension may define new AVPs for use in DOIC Capability 1006 Announcement and for use in DOIC Overload reporting. These new AVPs 1007 SHOULD be defined to be extensions to the OC-Supported-Features or 1008 OC-OLR AVPs defined in this document. 1010 [RFC6733] defined Grouped AVP extension mechanisms apply. This 1011 allows, for example, defining a new feature that is mandatory to be 1012 understood even when piggybacked on an existing application. 1014 When defining new report type values, the corresponding specification 1015 must define the semantics of the new report types and how they affect 1016 the OC-OLR AVP handling. 1018 The OC-Supported-Feature and OC-OLR AVPs can be expanded with 1019 optional sub-AVPs only if a legacy DOIC implementation can safely 1020 ignore them without breaking backward compatibility for the given OC- 1021 Report-Type AVP value. Any new sub-AVPs must not require that the 1022 M-bit be set. 1024 Documents that introduce new report types must describe any 1025 limitations on their use across non-supporting agents. 1027 As with any Diameter specification, RFC6733 requires all new AVPs to 1028 be registered with IANA. See Section 9 for the required procedures. 1029 New features (feature bits in the OC-Feature-Vector AVP) and report 1030 types (in the OC-Report-Type AVP) MUST be registered with IANA. 1032 6. Loss Algorithm 1034 This section documents the Diameter overload loss abatement 1035 algorithm. 1037 6.1. Overview 1039 The DOIC specification supports the ability for multiple overload 1040 abatement algorithms to be specified. The abatement algorithm used 1041 for any instance of overload is determined by the Diameter Overload 1042 Capability Announcement process documented in Section 5.1. 1044 The loss algorithm described in this section is the default algorithm 1045 that must be supported by all Diameter nodes that support DOIC. 1047 The loss algorithm is designed to be a straightforward and stateless 1048 overload abatement algorithm. It is used by reporting nodes to 1049 request a percentage reduction in the amount of traffic sent. The 1050 traffic impacted by the requested reduction depends on the type of 1051 overload report. 1053 Reporting nodes request the stateless reduction of the number of 1054 requests by an indicated percentage. This percentage reduction is in 1055 comparison to the number of messages the node otherwise would send, 1056 regardless of how many requests the node might have sent in the past. 1058 From a conceptual level, the logic at the reacting node could be 1059 outlined as follows. 1061 1. An overload report is received and the associated OCS is either 1062 saved or updated (if required) by the reacting node. 1064 2. A new Diameter request is generated by the application running on 1065 the reacting node. 1067 3. The reacting node determines that an active overload report 1068 applies to the request, as indicated by the corresponding OCS 1069 entry. 1071 4. The reacting node determines if overload abatement treatment 1072 should be applied to the request. One approach that could be 1073 taken for each request is to select a uniformly selected random 1074 number between 1 and 100. If the random number is less than or 1075 equal to the indicated reduction percentage then the request is 1076 given abatement treatment, otherwise the request is given normal 1077 routing treatment. 1079 6.2. Reporting Node Behavior 1081 The method a reporting node uses to determine the amount of traffic 1082 reduction required to address an overload condition is an 1083 implementation decision. 1085 When a reporting node that has selected the loss abatement algorithm 1086 determines the need to request a reduction in traffic, it includes an 1087 OC-OLR AVP in answer messages as described in Section 5.2.3. 1089 When sending the OC-OLR AVP, the reporting node MUST indicate a 1090 percentage reduction in the OC-Reduction-Percentage AVP. 1092 The reporting node MAY change the reduction percentage in subsequent 1093 overload reports. When doing so the reporting node must conform to 1094 overload report handing specified in Section 5.2.3. 1096 6.3. Reacting Node Behavior 1098 The method a reacting node uses to determine which request messages 1099 are given abatement treatment is an implementation decision. 1101 When receiving an OC-OLR in an answer message where the algorithm 1102 indicated in the OC-Supported-Features AVP is the loss algorithm, the 1103 reacting node MUST apply abatement treatment to the requested 1104 percentage of request messages sent. 1106 Note: The loss algorithm is a stateless algorithm. As a result, 1107 the reacting node does not guarantee that there will be an 1108 absolute reduction in traffic sent. Rather, it guarantees that 1109 the requested percentage of new requests will be given abatement 1110 treatment. 1112 If reacting node comes out of the 100 percent traffic reduction, 1113 meaning it has received an OLR indicating that no traffic should be 1114 sent, as a result of the overload report timing out the reacting node 1115 sending the traffic SHOULD be conservative and, for example, first 1116 send "probe" messages to learn the overload condition of the 1117 overloaded node before converging to any traffic amount/rate decided 1118 by the sender. Similar concerns apply in all cases when the overload 1119 report times out unless the previous overload report stated 0 percent 1120 reduction. 1122 The goal of this behavior is to reduce the probability of overload 1123 condition thrashing where an immediate transition from 100% 1124 reduction to 0% reduction results in the reporting node moving 1125 quickly back into an overload condition. 1127 7. Attribute Value Pairs 1129 This section describes the encoding and semantics of the Diameter 1130 Overload Indication Attribute Value Pairs (AVPs) defined in this 1131 document. 1133 Refer to section 4 of [RFC6733] for more information on AVPs and AVP 1134 data types. 1136 7.1. OC-Supported-Features AVP 1138 The OC-Supported-Features AVP (AVP code TBD1) is of type Grouped and 1139 serves two purposes. First, it announces a node's support for the 1140 DOIC solution in general. Second, it contains the description of the 1141 supported DOIC features of the sending node. The OC-Supported- 1142 Features AVP MUST be included in every Diameter request message a 1143 DOIC supporting node sends. 1145 OC-Supported-Features ::= < AVP Header: TBD1 > 1146 [ OC-Feature-Vector ] 1147 * [ AVP ] 1149 7.2. OC-Feature-Vector AVP 1151 The OC-Feature-Vector AVP (AVP code TBD2) is of type Unsigned64 and 1152 contains a 64 bit flags field of announced capabilities of a DOIC 1153 node. The value of zero (0) is reserved. 1155 The OC-Feature-Vector sub-AVP is used to announce the DOIC features 1156 supported by the DOIC node, in the form of a flag-bits field in which 1157 each bit announces one feature or capability supported by the node. 1158 The absence of the OC-Feature-Vector AVP in request messages 1159 indicates that only the default traffic abatement algorithm described 1160 in this specification is supported. The absence of the OC- Feature- 1161 Vector AVP in answer messages indicates that the default traffic 1162 abatement algorithm described in this specification is selected 1163 (while other traffic abatement algorithms may be supported), and no 1164 features other than abatement algorithms are supported. 1166 The following capabilities are defined in this document: 1168 OLR_DEFAULT_ALGO (0x0000000000000001) 1170 When this flag is set by the a DOIC reacting node it means that 1171 the default traffic abatement (loss) algorithm is supported. When 1172 this flag is set by a DOIC reporting node it means that the loss 1173 algorithm will be used for requested overload abatement. 1175 7.3. OC-OLR AVP 1177 The OC-OLR AVP (AVP code TBD3) is of type Grouped and contains the 1178 information necessary to convey an overload report on an overload 1179 condition at the reporting node. The application the OC-OLR AVP 1180 applies to is the same as the Application-Id found in the Diameter 1181 message header. The host or realm the OC-OLR AVP concerns is 1182 determined from the Origin-Host AVP and/or Origin-Realm AVP found in 1183 the encapsulating Diameter command. The OC-OLR AVP is intended to be 1184 sent only by a reporting node. 1186 OC-OLR ::= < AVP Header: TBD2 > 1187 < OC-Sequence-Number > 1188 < OC-Report-Type > 1189 [ OC-Reduction-Percentage ] 1190 [ OC-Validity-Duration ] 1191 * [ AVP ] 1193 7.4. OC-Sequence-Number AVP 1195 The OC-Sequence-Number AVP (AVP code TBD4) is of type Unsigned64. 1196 Its usage in the context of overload control is described in 1197 Section 5.2. 1199 From the functionality point of view, the OC-Sequence-Number AVP is 1200 used as a non-volatile increasing counter for a sequence of overload 1201 reports between two DOIC nodes for the same overload occurrence. 1202 Sequence numbers are treated in a uni-directional manner, i.e., two 1203 sequence numbers on each direction between two DOIC nodes are not 1204 related or correlated. 1206 7.5. OC-Validity-Duration AVP 1208 The OC-Validity-Duration AVP (AVP code TBD5) is of type Unsigned32 1209 and indicates in seconds the validity time of the overload report. 1210 The number of seconds is measured after reception of the first OC-OLR 1211 AVP with a given value of OC-Sequence-Number AVP. The default value 1212 for the OC-Validity-Duration AVP is 30 seconds. When the OC- 1213 Validity-Duration AVP is not present in the OC-OLR AVP, the default 1214 value applies. The maximum value for the OC-Validity-Duration AVP is 1215 86,400 seconds (24 hours). If the value received in the OC-Validity- 1216 Duration is greater than the maximum value then the default value 1217 applies. 1219 7.6. OC-Report-Type AVP 1221 The OC-Report-Type AVP (AVP code TBD6) is of type Enumerated. The 1222 value of the AVP describes what the overload report concerns. The 1223 following values are initially defined: 1225 HOST_REPORT 0 The overload report is for a host. Overload abatement 1226 treatment applies to host-routed requests. 1228 REALM_REPORT 1 The overload report is for a realm. Overload 1229 abatement treatment applies to realm-routed requests. 1231 7.7. OC-Reduction-Percentage AVP 1233 The OC-Reduction-Percentage AVP (AVP code TBD7) is of type Unsigned32 1234 and describes the percentage of the traffic that the sender is 1235 requested to reduce, compared to what it otherwise would send. The 1236 OC-Reduction-Percentage AVP applies to the default (loss) algorithm 1237 specified in this specification. However, the AVP can be reused for 1238 future abatement algorithms, if its semantics fit into the new 1239 algorithm. 1241 The value of the Reduction-Percentage AVP is between zero (0) and one 1242 hundred (100). Values greater than 100 are ignored. The value of 1243 100 means that all traffic is to be throttled, i.e., the reporting 1244 node is under a severe load and ceases to process any new messages. 1245 The value of 0 means that the reporting node is in a stable state and 1246 has no need for the reacting node to apply any traffic abatement. 1248 7.8. Attribute Value Pair flag rules 1250 +---------+ 1251 |AVP flag | 1252 |rules | 1253 +----+----+ 1254 AVP Section | |MUST| 1255 Attribute Name Code Defined Value Type |MUST| NOT| 1256 +--------------------------------------------------+----+----+ 1257 |OC-Supported-Features TBD1 7.1 Grouped | | V | 1258 +--------------------------------------------------+----+----+ 1259 |OC-Feature-Vector TBD2 7.2 Unsigned64 | | V | 1260 +--------------------------------------------------+----+----+ 1261 |OC-OLR TBD3 7.3 Grouped | | V | 1262 +--------------------------------------------------+----+----+ 1263 |OC-Sequence-Number TBD4 7.4 Unsigned64 | | V | 1264 +--------------------------------------------------+----+----+ 1265 |OC-Validity-Duration TBD5 7.5 Unsigned32 | | V | 1266 +--------------------------------------------------+----+----+ 1267 |OC-Report-Type TBD6 7.6 Enumerated | | V | 1268 +--------------------------------------------------+----+----+ 1269 |OC-Reduction | | | 1270 | -Percentage TBD7 7.7 Unsigned32 | | V | 1271 +--------------------------------------------------+----+----+ 1273 As described in the Diameter base protocol [RFC6733], the M-bit usage 1274 for a given AVP in a given command may be defined by the application. 1276 8. Error Response Codes 1278 When a DOIC node rejects a Diameter request due to overload, the DOIC 1279 node MUST select an appropriate error response code. This 1280 determination is made based on the probability of the request 1281 succeeding if retried on a different path. 1283 Note: This only applies for DOIC nodes that are not the originator 1284 of the request. 1286 A reporting node rejecting a Diameter request due to an overload 1287 condition SHOULD send a DIAMETER_TOO_BUSY error response, if it can 1288 assume that the same request may succeed on a different path. 1290 If a reporting node knows or assumes that the same request will not 1291 succeed on a different path, DIAMETER_UNABLE_TO_COMPLY error response 1292 SHOULD be used. Retrying would consume valuable resources during an 1293 occurrence of overload. 1295 For instance, if the request arrived at the reporting node without 1296 a Destination-Host AVP then the reporting node might determine 1297 that there is an alternative Diameter node that could successfully 1298 process the request and that retrying the transaction would not 1299 negatively impact the reporting node. DIAMETER_TOO_BUSY would be 1300 sent in this case. 1302 If the request arrived at the reporting node with a Destination- 1303 Host AVP populated with its own Diameter identity then the 1304 reporting node can assume that retrying the request would result 1305 in it coming to the same reporting node. 1306 DIAMETER_UNABLE_TO_COMPLY would be sent in this case. 1308 A second example is when an agent that supports the DOIC solution 1309 is performing the role of a reacting node for a non-supporting 1310 client. Requests that are rejected as a result of DOIC throttling 1311 by the agent in this scenario would generally be rejected with a 1312 DIAMETER_UNABLE_TO_COMPLY response code. 1314 9. IANA Considerations 1316 9.1. AVP codes 1318 New AVPs defined by this specification are listed in Section 7. All 1319 AVP codes are allocated from the 'Authentication, Authorization, and 1320 Accounting (AAA) Parameters' AVP Codes registry. 1322 9.2. New registries 1324 Two new registries are needed under the 'Authentication, 1325 Authorization, and Accounting (AAA) Parameters' registry. 1327 A new "Overload Control Feature Vector" registry is required. The 1328 registry must contain the following: 1330 Feature Vector Value Name 1332 Feature Vector Value 1334 Specification - the specification that defines the new value. 1336 See Section 7.2 for the initial Feature Vector Value in the registry. 1337 This specification is the specification defining the value. New 1338 values can be added into the registry using the Specification 1339 Required policy. [RFC5226]. 1341 A new "Overload Report Type" registry is required. The registry must 1342 contain the following: 1344 Report Type Value Name 1346 Report Type Value 1348 Specification - the specification that defines the new value. 1350 See Section 7.6 for the initial assignment in the registry. New 1351 types can be added using the Specification Required policy [RFC5226]. 1353 10. Security Considerations 1355 DOIC gives Diameter nodes the ability to request that downstream 1356 nodes send fewer Diameter requests. Nodes do this by exchanging 1357 overload reports that directly effect this reduction. This exchange 1358 is potentially subject to multiple methods of attack, and has the 1359 potential to be used as a Denial-of-Service (DoS) attack vector. For 1360 instance, a series of injected realm OLRs with a requested reduction 1361 percentage of 100% could be used to completely eliminate any traffic 1362 from being sent to that realm. 1364 Overload reports may contain information about the topology and 1365 current status of a Diameter network. This information is 1366 potentially sensitive. Network operators may wish to control 1367 disclosure of overload reports to unauthorized parties to avoid its 1368 use for competitive intelligence or to target attacks. 1370 Diameter does not include features to provide end-to-end 1371 authentication, integrity protection, or confidentiality. This may 1372 cause complications when sending overload reports between non- 1373 adjacent nodes. 1375 10.1. Potential Threat Modes 1377 The Diameter protocol involves transactions in the form of requests 1378 and answers exchanged between clients and servers. These clients and 1379 servers may be peers, that is, they may share a direct transport 1380 (e.g., TCP or SCTP) connection, or the messages may traverse one or 1381 more intermediaries, known as Diameter Agents. Diameter nodes use 1382 TLS, DTLS, or IPsec to authenticate peers, and to provide 1383 confidentiality and integrity protection of traffic between peers. 1384 Nodes can make authorization decisions based on the peer identities 1385 authenticated at the transport layer. 1387 When agents are involved, this presents an effectively transitive 1388 trust model. That is, a Diameter client or server can authorize an 1389 agent for certain actions, but it must trust that agent to make 1390 appropriate authorization decisions about its peers, and so on. 1391 Since confidentiality and integrity protection occurs at the 1392 transport layer, agents can read, and perhaps modify, any part of a 1393 Diameter message, including an overload report. 1395 There are several ways an attacker might attempt to exploit the 1396 overload control mechanism. An unauthorized third party might inject 1397 an overload report into the network. If this third party is upstream 1398 of an agent, and that agent fails to apply proper authorization 1399 policies, downstream nodes may mistakenly trust the report. This 1400 attack is at least partially mitigated by the assumption that nodes 1401 include overload reports in Diameter answers but not in requests. 1402 This requires an attacker to have knowledge of the original request 1403 in order to construct an answer. Such an answer would also need to 1404 arrive at a Diameter node via a protected transport connection. 1405 Therefore, implementations MUST validate that an answer containing an 1406 overload report is a properly constructed response to a pending 1407 request prior to acting on the overload report, and that the answer 1408 was received via an appropriate transport connection. 1410 A similar attack involves a compromised but otherwise authorized node 1411 that sends an inappropriate overload report. For example, a server 1412 for the realm "example.com" might send an overload report indicating 1413 that a competitor's realm "example.net" is overloaded. If other 1414 nodes act on the report, they may falsely believe that "example.net" 1415 is overloaded, effectively reducing that realm's capacity. 1416 Therefore, it's critical that nodes validate that an overload report 1417 received from a peer actually falls within that peer's responsibility 1418 before acting on the report or forwarding the report to other peers. 1419 For example, an overload report from a peer that applies to a realm 1420 not handled by that peer is suspect. This may require out-of-band, 1421 non Diameter agreements and/or mechanisms. 1423 This attack is partially mitigated by the fact that the 1424 application, as well as host and realm, for a given OLR is 1425 determined implicitly by respective AVPs in the enclosing answer. 1426 If a reporting node modifies any of those AVPs, the enclosing 1427 transaction will also be affected. 1429 10.2. Denial of Service Attacks 1431 Diameter overload reports, especially realm-reports, can cause a node 1432 to cease sending some or all Diameter requests for an extended 1433 period. This makes them a tempting vector for DoS attacks. 1434 Furthermore, since Diameter is almost always used in support of other 1435 protocols, a DoS attack on Diameter is likely to impact those 1436 protocols as well. In the worst case, where the Diameter application 1437 is being used for access control into an IP network, a coordinated 1438 DOS attack could result in the blockage of all traffic into that 1439 network. Therefore, Diameter nodes MUST NOT honor or forward OLRs 1440 received from peers that are not trusted to send them. 1442 An attacker might use the information in an OLR to assist in DoS 1443 attacks. For example, an attacker could use information about 1444 current overload conditions to time an attack for maximum effect, or 1445 use subsequent overload reports as a feedback mechanism to learn the 1446 results of a previous or ongoing attack. Operators need the ability 1447 to ensure that OLRs are not leaked to untrusted parties. 1449 10.3. Non-Compliant Nodes 1451 In the absence of an overload control mechanism, Diameter nodes need 1452 to implement strategies to protect themselves from floods of 1453 requests, and to make sure that a disproportionate load from one 1454 source does not prevent other sources from receiving service. For 1455 example, a Diameter server might throttle a certain percentage of 1456 requests from sources that exceed certain limits. Overload control 1457 can be thought of as an optimization for such strategies, where 1458 downstream nodes never send the excess requests in the first place. 1459 However, the presence of an overload control mechanism does not 1460 remove the need for these other protection strategies. 1462 When a Diameter node sends an overload report, it cannot assume that 1463 all nodes will comply, even if they indicate support for DOIC. A 1464 non-compliant node might continue to send requests with no reduction 1465 in load. Such non-compliance could be done accidentally, or 1466 maliciously to gain an unfair advantage over compliant nodes. 1467 Requirement 28 [RFC7068] indicates that the overload control solution 1468 cannot assume that all Diameter nodes in a network are trusted. It 1469 also requires that malicious nodes not be allowed to take advantage 1470 of the overload control mechanism to get more than their fair share 1471 of service. 1473 10.4. End-to End-Security Issues 1475 The lack of end-to-end integrity features makes it difficult to 1476 establish trust in overload reports received from non-adjacent nodes. 1477 Any agents in the message path may insert or modify overload reports. 1478 Nodes must trust that their adjacent peers perform proper checks on 1479 overload reports from their peers, and so on, creating a transitive- 1480 trust requirement extending for potentially long chains of nodes. 1481 Network operators must determine if this transitive trust requirement 1482 is acceptable for their deployments. Nodes supporting Diameter 1483 overload control MUST give operators the ability to select which 1484 peers are trusted to deliver overload reports, and whether they are 1485 trusted to forward overload reports from non-adjacent nodes. DOIC 1486 nodes MUST strip DOIC AVPs from messages received from peers that are 1487 not trusted for DOIC purposes. 1489 The lack of end-to-end confidentiality protection means that any 1490 Diameter agent in the path of an overload report can view the 1491 contents of that report. In addition to the requirement to select 1492 which peers are trusted to send overload reports, operators MUST be 1493 able to select which peers are authorized to receive reports. A node 1494 MUST NOT send an overload report to a peer not authorized to receive 1495 it. Furthermore, an agent MUST remove any overload reports that 1496 might have been inserted by other nodes before forwarding a Diameter 1497 message to a peer that is not authorized to receive overload reports. 1499 A DOIC node cannot always automatically detect that a peer also 1500 supports DOIC. For example, a node might have a peer that is a 1501 non-supporting agent. If nodes on the other side of that agent 1502 send OC-Supported-Features AVPs, the agent is likely to forward 1503 them as unknown AVPs. Messages received across the non-supporting 1504 agent may be indistinguishable from messages received across a 1505 DOIC supporting agent, giving the false impression that the non- 1506 supporting agent actually supports DOIC. This complicates the 1507 transitive-trust nature of DOIC. Operators need to be careful to 1508 avoid situations where a non-supporting agent is mistakenly 1509 trusted to enforce DOIC related authorization policies. 1511 It is expected that work on end-to-end Diameter security might make 1512 it easier to establish trust in non-adjacent nodes for overload 1513 control purposes. Readers should be reminded, however, that the 1514 overload control mechanism allows Diameter agents to modify AVPs in, 1515 or insert additional AVPs into, existing messages that are originated 1516 by other nodes. If end-to-end security is enabled, there is a risk 1517 that such modification could violate integrity protection. The 1518 details of using any future Diameter end-to-end security mechanism 1519 with overload control will require careful consideration, and are 1520 beyond the scope of this document. 1522 11. Contributors 1524 The following people contributed substantial ideas, feedback, and 1525 discussion to this document: 1527 o Eric McMurry 1529 o Hannes Tschofenig 1531 o Ulrich Wiehe 1533 o Jean-Jacques Trottin 1535 o Maria Cruz Bartolome 1537 o Martin Dolly 1539 o Nirav Salot 1541 o Susan Shishufeng 1543 12. References 1545 12.1. Normative References 1547 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 1548 Requirement Levels", BCP 14, RFC 2119, 1549 DOI 10.17487/RFC2119, March 1997, 1550 . 1552 [RFC5226] Narten, T. and H. Alvestrand, "Guidelines for Writing an 1553 IANA Considerations Section in RFCs", BCP 26, RFC 5226, 1554 DOI 10.17487/RFC5226, May 2008, 1555 . 1557 [RFC6733] Fajardo, V., Ed., Arkko, J., Loughney, J., and G. Zorn, 1558 Ed., "Diameter Base Protocol", RFC 6733, 1559 DOI 10.17487/RFC6733, October 2012, 1560 . 1562 12.2. Informative References 1564 [Cx] 3GPP, , "ETSI TS 129 229 V11.4.0", August 2013. 1566 [I-D.ietf-dime-e2e-sec-req] 1567 Tschofenig, H., Korhonen, J., Zorn, G., and K. Pillay, 1568 "Diameter AVP Level Security: Scenarios and Requirements", 1569 draft-ietf-dime-e2e-sec-req-01 (work in progress), October 1570 2013. 1572 [PCC] 3GPP, , "ETSI TS 123 203 V11.12.0", December 2013. 1574 [RFC4006] Hakala, H., Mattila, L., Koskinen, J-P., Stura, M., and J. 1575 Loughney, "Diameter Credit-Control Application", RFC 4006, 1576 DOI 10.17487/RFC4006, August 2005, 1577 . 1579 [RFC7068] McMurry, E. and B. Campbell, "Diameter Overload Control 1580 Requirements", RFC 7068, DOI 10.17487/RFC7068, November 1581 2013, . 1583 [S13] 3GPP, , "ETSI TS 129 272 V11.9.0", December 2012. 1585 Appendix A. Issues left for future specifications 1587 The base solution for the overload control does not cover all 1588 possible use cases. A number of solution aspects were intentionally 1589 left for future specification and protocol work. The following sub- 1590 sections define some of the potential extensions to the DOIC 1591 solution. 1593 A.1. Additional traffic abatement algorithms 1595 This specification describes only means for a simple loss based 1596 algorithm. Future algorithms can be added using the designed 1597 solution extension mechanism. The new algorithms need to be 1598 registered with IANA. See Sections 7.1 and 9 for the required IANA 1599 steps. 1601 A.2. Agent Overload 1603 This specification focuses on Diameter endpoint (server or client) 1604 overload. A separate extension will be required to outline the 1605 handling of the case of agent overload. 1607 A.3. New Error Diagnostic AVP 1609 This specification indicates the use of existing error messages when 1610 nodes reject requests due to overload. There is an expectation that 1611 additional error codes or AVPs will be defined in a separate 1612 specification to indicate that overload was the reason for the 1613 rejection of the message. 1615 Appendix B. Deployment Considerations 1617 Non-Supporting Agents 1619 Due to the way that realm-routed requests are handled in Diameter 1620 networks with the server selection for the request done by an 1621 agent, network operators should enable DOIC at agents that perform 1622 server selection first. 1624 Topology Hiding Interactions 1626 There exist proxies that implement what is referred to as Topology 1627 Hiding. This can include cases where the agent modifies the 1628 Origin-Host in answer messages. The behavior of the DOIC solution 1629 is not well understood when this happens. As such, the DOIC 1630 solution does not address this scenario. 1632 Inter Realm/Administrative Domain Considerations 1634 There are likely to be special considerations for handling DOIC 1635 signaling across administrative boundaries. This includes 1636 considerations for whether or not information included in the DOIC 1637 signaling should be sent across those boundaries. In addition 1638 consideration should be taken as to whether or not a reacting node 1639 in one realm can be trusted to implement the requested overload 1640 abatement handling for overload reports received from a separately 1641 administered realm. 1643 Appendix C. Considerations for Applications Integrating the DOIC 1644 Solution 1646 This section outlines considerations to be taken into account when 1647 integrating the DOIC solution into Diameter applications. 1649 C.1. Application Classification 1651 The following is a classification of Diameter applications and 1652 request types. This discussion is meant to document factors that 1653 play into decisions made by the Diameter entity responsible for 1654 handling overload reports. 1656 Section 8.1 of [RFC6733] defines two state machines that imply two 1657 types of applications, session-less and session-based applications. 1658 The primary difference between these types of applications is the 1659 lifetime of Session-Ids. 1661 For session-based applications, the Session-Id is used to tie 1662 multiple requests into a single session. 1664 The Credit-Control application defined in [RFC4006] is an example of 1665 a Diameter session-based application. 1667 In session-less applications, the lifetime of the Session-Id is a 1668 single Diameter transaction, i.e., the session is implicitly 1669 terminated after a single Diameter transaction and a new Session-Id 1670 is generated for each Diameter request. 1672 For the purposes of this discussion, session-less applications are 1673 further divided into two types of applications: 1675 Stateless Applications: 1677 Requests within a stateless application have no relationship to 1678 each other. The 3GPP defined S13 application is an example of a 1679 stateless application [S13], where only a Diameter command is 1680 defined between a client and a server and no state is maintained 1681 between two consecutive transactions. 1683 Pseudo-Session Applications: 1685 Applications that do not rely on the Session-Id AVP for 1686 correlation of application messages related to the same session 1687 but use other session-related information in the Diameter requests 1688 for this purpose. The 3GPP defined Cx application [Cx] is an 1689 example of a pseudo-session application. 1691 The handling of overload reports must take the type of application 1692 into consideration, as discussed in Appendix C.2. 1694 C.2. Application Type Overload Implications 1696 This section discusses considerations for mitigating overload 1697 reported by a Diameter entity. This discussion focuses on the type 1698 of application. Appendix C.3 discusses considerations for handling 1699 various request types when the target server is known to be in an 1700 overloaded state. 1702 These discussions assume that the strategy for mitigating the 1703 reported overload is to reduce the overall workload sent to the 1704 overloaded entity. The concept of applying overload treatment to 1705 requests targeted for an overloaded Diameter entity is inherent to 1706 this discussion. The method used to reduce offered load is not 1707 specified here but could include routing requests to another Diameter 1708 entity known to be able to handle them, or it could mean rejecting 1709 certain requests. For a Diameter agent, rejecting requests will 1710 usually mean generating appropriate Diameter error responses. For a 1711 Diameter client, rejecting requests will depend upon the application. 1712 For example, it could mean giving an indication to the entity 1713 requesting the Diameter service that the network is busy and to try 1714 again later. 1716 Stateless Applications: 1718 By definition there is no relationship between individual requests 1719 in a stateless application. As a result, when a request is sent 1720 or relayed to an overloaded Diameter entity - either a Diameter 1721 Server or a Diameter Agent - the sending or relaying entity can 1722 choose to apply the overload treatment to any request targeted for 1723 the overloaded entity. 1725 Pseudo-Session Applications: 1727 For pseudo-session applications, there is an implied ordering of 1728 requests. As a result, decisions about which requests towards an 1729 overloaded entity to reject could take the command code of the 1730 request into consideration. This generally means that 1731 transactions later in the sequence of transactions should be given 1732 more favorable treatment than messages earlier in the sequence. 1733 This is because more work has already been done by the Diameter 1734 network for those transactions that occur later in the sequence. 1735 Rejecting them could result in increasing the load on the network 1736 as the transactions earlier in the sequence might also need to be 1737 repeated. 1739 Session-Based Applications: 1741 Overload handling for session-based applications must take into 1742 consideration the work load associated with setting up and 1743 maintaining a session. As such, the entity sending requests 1744 towards an overloaded Diameter entity for a session-based 1745 application might tend to reject new session requests prior to 1746 rejecting intra-session requests. In addition, session ending 1747 requests might be given a lower probability of being rejected as 1748 rejecting session ending requests could result in session status 1749 being out of sync between the Diameter clients and servers. 1750 Application designers that would decide to reject mid-session 1751 requests will need to consider whether the rejection invalidates 1752 the session and any resulting session cleanup procedures. 1754 C.3. Request Transaction Classification 1756 Independent Request: 1758 An independent request is not correlated to any other requests 1759 and, as such, the lifetime of the session-id is constrained to an 1760 individual transaction. 1762 Session-Initiating Request: 1764 A session-initiating request is the initial message that 1765 establishes a Diameter session. The ACR message defined in 1766 [RFC6733] is an example of a session-initiating request. 1768 Correlated Session-Initiating Request: 1770 There are cases when multiple session-initiated requests must be 1771 correlated and managed by the same Diameter server. It is notably 1772 the case in the 3GPP PCC architecture [PCC], where multiple 1773 apparently independent Diameter application sessions are actually 1774 correlated and must be handled by the same Diameter server. 1776 Intra-Session Request: 1778 An intra-session request is a request that uses the same Session- 1779 Id than the one used in a previous request. An intra-session 1780 request generally needs to be delivered to the server that handled 1781 the session creating request for the session. The STR message 1782 defined in [RFC6733] is an example of an intra-session request. 1784 Pseudo-Session Requests: 1786 Pseudo-session requests are independent requests and do not use 1787 the same Session-Id but are correlated by other session-related 1788 information contained in the request. There exists Diameter 1789 applications that define an expected ordering of transactions. 1790 This sequencing of independent transactions results in a pseudo 1791 session. The AIR, MAR and SAR requests in the 3GPP defined Cx 1792 [Cx] application are examples of pseudo-session requests. 1794 C.4. Request Type Overload Implications 1796 The request classes identified in Appendix C.3 have implications on 1797 decisions about which requests should be throttled first. The 1798 following list of request treatment regarding throttling is provided 1799 as guidelines for application designers when implementing the 1800 Diameter overload control mechanism described in this document. The 1801 exact behavior regarding throttling is a matter of local policy, 1802 unless specifically defined for the application. 1804 Independent Requests: 1806 Independent requests can generally be given equal treatment when 1807 making throttling decisions, unless otherwise indicated by 1808 application requirements or local policy. 1810 Session-Initiating Requests: 1812 Session-initiating requests often represent more work than 1813 independent or intra-session requests. Moreover, session- 1814 initiating requests are typically followed by other session- 1815 related requests. Since the main objective of the overload 1816 control is to reduce the total number of requests sent to the 1817 overloaded entity, throttling decisions might favor allowing 1818 intra-session requests over session-initiating requests. In the 1819 absence of local policies or application specific requirements to 1820 the contrary, Individual session-initiating requests can be given 1821 equal treatment when making throttling decisions. 1823 Correlated Session-Initiating Requests: 1825 A Request that results in a new binding, where the binding is used 1826 for routing of subsequent session-initiating requests to the same 1827 server, represents more work load than other requests. As such, 1828 these requests might be throttled more frequently than other 1829 request types. 1831 Pseudo-Session Requests: 1833 Throttling decisions for pseudo-session requests can take into 1834 consideration where individual requests fit into the overall 1835 sequence of requests within the pseudo session. Requests that are 1836 earlier in the sequence might be throttled more aggressively than 1837 requests that occur later in the sequence. 1839 Intra-Session Requests: 1841 There are two types of intra-sessions requests, requests that 1842 terminate a session and the remainder of intra-session requests. 1843 Implementers and operators may choose to throttle session- 1844 terminating requests less aggressively in order to gracefully 1845 terminate sessions, allow cleanup of the related resources (e.g., 1846 session state) and avoid the need for additional intra-session 1847 requests. Favoring session-termination requests may reduce the 1848 session management impact on the overloaded entity. The default 1849 handling of other intra-session requests might be to treat them 1850 equally when making throttling decisions. There might also be 1851 application level considerations whether some request types are 1852 favored over others. 1854 Authors' Addresses 1856 Jouni Korhonen (editor) 1857 Broadcom 1858 Porkkalankatu 24 1859 Helsinki FIN-00180 1860 Finland 1862 Email: jouni.nospam@gmail.com 1864 Steve Donovan (editor) 1865 Oracle 1866 7460 Warren Parkway 1867 Frisco, Texas 75034 1868 United States 1870 Email: srdonovan@usdonovans.com 1872 Ben Campbell 1873 Oracle 1874 7460 Warren Parkway 1875 Frisco, Texas 75034 1876 United States 1878 Email: ben@nostrum.com 1880 Lionel Morand 1881 Orange Labs 1882 38/40 rue du General Leclerc 1883 Issy-Les-Moulineaux Cedex 9 92794 1884 France 1886 Phone: +33145296257 1887 Email: lionel.morand@orange.com