idnits 2.17.1 draft-noel-soc-overload-rate-control-02.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- ** The abstract seems to contain references ([RFC3261]), which it shouldn't. Please replace those with straight textual mentions of the documents in question. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (December 12, 2011) is 4519 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) == Missing Reference: 'RFC2119' is mentioned on line 118, but not defined -- Looks like a reference, but probably isn't: '0' on line 296 == Missing Reference: 'T' is mentioned on line 296, but not defined Summary: 1 error (**), 0 flaws (~~), 3 warnings (==), 2 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 1 SOC Working Group Eric Noel 2 Internet-Draft AT&T Labs 3 Intended status: Standards Track Philip M Williams 4 Expires: June 12 2012 BT Innovate & Design 6 December 12, 2011 8 Session Initiation Protocol (SIP) Rate Control 9 draft-noel-soc-overload-rate-control-02.txt 11 Abstract 13 The prevalent use of Session Initiation Protocol (SIP) [RFC3261] in 14 Next Generation Networks necessitates that SIP networks provide 15 adequate control mechanisms to maintain transaction throughput by 16 preventing congestion collapse during traffic overloads. Already 17 [draft-ietf-soc-overload-control-05] proposes a loss-based solution 18 to remedy known vulnerabilities of the [RFC3261] SIP 503 (service 19 unavailable) overload control mechanism. This document proposes a 20 rate-based control solution to complement the loss-based control 21 defined in [draft-ietf-soc-overload-control-05]. 23 Status of this Memo 25 This Internet-Draft is submitted in full conformance with the 26 provisions of BCP 78 and BCP 79. 28 Internet-Drafts are working documents of the Internet Engineering 29 Task Force (IETF). Note that other groups may also distribute 30 working documents as Internet-Drafts. The list of current Internet- 31 Drafts is at http://datatracker.ietf.org/drafts/current/. 33 Internet-Drafts are draft documents valid for a maximum of six 34 months and may be updated, replaced, or obsoleted by other documents 35 at any time. It is inappropriate to use Internet-Drafts as 36 reference material or to cite them other than as "work in progress." 38 This Internet-Draft will expire on June 12, 2012. 40 Copyright Notice 42 Copyright (c) 2011 IETF Trust and the persons identified as the 43 document authors. All rights reserved. 45 This document is subject to BCP 78 and the IETF Trust's Legal 46 Provisions Relating to IETF Documents 47 (http://trustee.ietf.org/license-info) in effect on the date of 48 publication of this document. Please review these documents 49 carefully, as they describe your rights and restrictions with 50 respect to this document. Code Components extracted from this 51 document must include Simplified BSD License text as described in 52 Section 4.e of the Trust Legal Provisions and are provided without 53 warranty as described in the Simplified BSD License. 55 Table of Contents 57 1. Introduction...................................................2 58 2. Terminology....................................................3 59 3. Rate-based algorithm scheme....................................4 60 3.1. Overview..................................................4 61 3.2. Summary of via headers parameters for overload control....4 62 3.3. Client and server rate-control algorithm selection........4 63 3.4. Server operation..........................................5 64 3.5. Client operation..........................................5 65 3.5.1. Default algorithm....................................5 66 3.5.2. Optional enhancement: avoidance of resonance.........7 67 3.5.3. Optional enhancement: Priority.......................8 68 4. Example........................................................8 69 5. Syntax........................................................10 70 6. Security Considerations.......................................10 71 7. IANA Considerations...........................................10 72 8. References....................................................10 73 8.1. Normative References.....................................10 74 8.2. Informative References...................................10 75 Appendix A. Acknowledgments......................................12 77 1. Introduction 79 The use of SIP in large scale Next Generation Networks requires that 80 SIP based networks provide adequate control mechanisms for handling 81 traffic growth. In particular, SIP networks must be able to handle 82 traffic overloads gracefully, maintaining transaction throughput by 83 preventing congestion collapse. 85 A promising SIP based overload control solution has been proposed in 86 [draft-ietf-soc-overload-control-05]. That solution provides a 87 communication scheme for overload control algorithms. It also 88 includes a default loss-based overload control algorithm that makes 89 it possible for a set of clients to limit offered load towards an 90 overloaded server. 92 However, such loss control algorithm is sensitive to variations in 93 load so that any increase in load would be directly reflected by the 94 clients in the offered load presented to the overloaded servers. 95 More importantly, a loss-based control cannot guarantee clients to 96 produce a bounded offered load towards an overloaded server and 97 requires frequent updates which may have implications for stability. 99 This document proposes extensions to [draft-ietf-soc-overload- 100 control-05] so as to support a rate-based control that guarantees 101 clients produce a limited upper bound to the Invite rate towards an 102 overloaded server, which is constant between server updates. The 103 penalty for such a benefit is in terms of algorithmic complexity, 104 since the overloaded server must estimate a target offered load and 105 allocate a portion to each conversing client. 107 The proposed rate-based overload control algorithm mitigates 108 congestion in SIP networks while adhering to the overload signaling 109 scheme in [draft-ietf-soc-overload-control-05] and proposing a rate 110 control in addition to the default loss-based control in [draft- 111 ietf-soc-overload-control-05]. 113 2. Terminology 115 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 116 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 117 document are to be interpreted as described in RFC 2119 [RFC2119]. 119 The normative statements in this specification as they apply to SIP 120 clients and SIP servers assume that both the SIP clients and SIP 121 servers support this specification. If, for instance, only a SIP 122 client supports this specification and not the SIP server, then 123 follows that the normative statements in this specification 124 pertinent to the behavior of a SIP server do not apply to the server 125 that does not support this specification. 127 3. Rate-based algorithm scheme 129 3.1. Overview 131 The server is what the overload control algorithm defined here 132 protects and the client is what throttles traffic towards the 133 server. 135 Following the procedures defined in [draft-ietf-soc-overload- 136 control-05], the server and clients signal one another support for 137 rate-based overload control. 139 Then periodically, the server relies on internal measurements (e.g. 140 CPU utilization, queueing delay...) to evaluate its overload state 141 and estimate a target SIP request rate (as opposed to target percent 142 loss in the case of loss-based control). 144 When in overload, the server uses [draft-ietf-soc-overload-control- 145 05] via header oc parameters of SIP responses to inform the clients 146 of its overload state and of the target SIP request rate. 148 Upon receiving the oc parameters with a target SIP request rate, 149 each client throttles new SIP requests towards the overloaded 150 server. 152 3.2. Summary of via headers parameters for overload control 154 To do: Repeat draft-ietf-soc-overload-control-05 definitions of the 155 oc parameters here. 157 The use of the via header oc parameter(s) inform of the desired 158 rate, but they don't explicitly "inform clients of the overload 159 state". 161 3.3. Client and server rate-control algorithm selection 163 Per [draft-ietf-soc-overload-control-05], new clients indicate 164 supported overload control algorithms to servers by inserting oc and 165 oc-algo in Via header of SIP requests destined to servers. While 166 servers notify clients of selected overload control algorithm 167 through the oc-algo parameter in the Via header of SIP responses to 168 clients. 170 Support of rate-based control MUST be indicated by clients and 171 servers by setting oc-algo to "rate". 173 3.4. Server operation 175 The actual algorithm used by the server to determine its overload 176 state and estimate a target SIP request rate is beyond the scope of 177 this document. 179 However, the server MUST be able to evaluate periodically its 180 overload state and estimate a target SIP request rate beyond which 181 it would become overloaded. The server must allocate a portion of 182 the target SIP request rate to each of its client. Note that the 183 target SIP request rate is a max rate that may not be attained by 184 the arrival rate at the client, and the server cannot assume that it 185 will. 187 Upon detection of overload, the server MUST follow the 188 specifications in [draft-ietf-soc-overload-control-05] to notify its 189 clients of its overload state and of the allocated target SIP 190 request rate. 192 The server MUST use [draft-ietf-soc-overload-control-05] oc 193 parameter to send a target SIP request rate to each of its client. 195 3.5. Client operation 197 3.5.1. Default algorithm 199 To throttle new SIP requests at the rate specified in the oc value 200 sent by the server to its clients, the client MAY use the proposed 201 default algorithm for rate-based control or any other equivalent 202 algorithm. 204 The default Leaky Bucket algorithm presented here is based on [ITU-T 205 Rec. I.371] Appendix A.2. The algorithm makes it possible for 206 clients to deliver SIP requests at a rate specified in the oc value 207 with tolerance parameter TAU (preferably configurable). 209 Conceptually, the Leaky Bucket algorithm can be viewed as a finite 210 capacity bucket whose real-valued content drains out at a continuous 211 rate of 1 unit of content per time unit and whose content increases 212 by the increment T for each forwarded SIP request. T is computed as 213 the inverse of the rate specified in the oc value, namely T = 1 / 214 oc-value. 216 Note that when the oc-value is 0 with a non zero oc-validity, then 217 the client should reject 100% of SIP requests destined to the 218 overload server. However, when both oc-value and oc-validity are 0, 219 the client should immediately stop throttling. 221 If at a new SIP request arrival the content of the bucket is less 222 than or equal to the limit value TAU, then the SIP request is 223 forwarded to the server; otherwise, the SIP request is rejected. 225 Note that the capacity of the bucket (the upper bound of the 226 counter) is (T + TAU). 228 At the arrival time of the k-th new SIP request ta(k) after control 229 has been activated, the content of the bucket is provisionally 230 updated to the value 232 X' = X - ([ta(k) - LCT]) 234 where X is the content of the bucket after arrival of the last 235 forwarded SIP request, and LCT is the time at which the last SIP 236 request was forwarded. 238 If X' is less than or equal to the limit value TAU, then the new SIP 239 request is forwarded and the bucket content X is set to X' (or to 0 240 if X' is negative) plus the increment T, and LCT is set to the 241 current time ta(k). If X' is greater than the limit value tau, then 242 the new SIP request is rejected and the values of X and LCT are 243 unchanged. 245 When the first response from the server has been received indicating 246 control activation (oc-validity>0), LCT is set to the time of 247 activation, and the occupancy of the bucket is initialized to the 248 parameter TAU0 (preferably configurable) which is 0 or larger but 249 less than or equal to TAU. 251 Note that specification of a value for TAU is beyond the scope of 252 this document. 254 3.5.2. Optional enhancement: avoidance of resonance 256 As the number of client sources of traffic increases and the 257 throughput of the server decreases, the maximum rate admitted by 258 each client needs to decrease, and therefore the value of T becomes 259 larger. Under some circumstances, e.g. if the traffic arises very 260 quickly simultaneously at many sources, the occupancies of each 261 bucket can become synchronized, resulting in the admissions from 262 each source being close in time and batched or very 'peaky' arrivals 263 at the server, which not only gives rise to control instability, but 264 also very poor delays and even lost messages. An appropriate term 265 for this is 'resonance' [Erramilli]. 267 If the network topology is such that this can occur, then a simple 268 way to avoid this is to randomize the bucket occupancy at two 269 appropriate points: At the activation of control, and whenever the 270 bucket empties, as follows. 272 After updating the bucket occupancy to X', generate a value u as 273 follows: 275 if X' > 0, then u=0 277 else if X' <= 0 then uniformly distributed between -1/2 and +1/2 279 Then (only) if the arrival is admitted, increase the bucket by an 280 amount T + uT, which will therefore be just T if the bucket hadn't 281 emptied, or lie between T/2 and 3T/2 if it had. 283 This randomization should also be done when control is activated, 284 i.e. instead of simply initializing the bucket fill to TAU0, 285 initialize it to TAU0 + uT, where u is uniformly distributed as 286 above. Since activation would have been a result of response to a 287 request sent by the client, the second term in this expression can 288 be interpreted as being the bucket increment following that 289 admission. 291 This method has the following characteristics: 293 . If TAU0 is chosen to be equal to TAU and all sources were to 294 activate control at the same time due to an extremely high 295 request rate, then the time until the first request admitted by 296 each client would be uniformly distributed over [0,T]; 298 . The maximum occupancy is TAU + (3/2)T, rather than TAU + T 299 without randomization; 301 . For the special case of 'classic gapping' where TAU=0, then the 302 minimum time between admissions is uniformly distributed over 303 [T/2, 3T/2], and the mean time between admissions is the same, 304 i.e. T+1/R where R is the request arrival rate; 306 . At high load randomization rarely occurs, so there is no loss 307 of precision of the admitted rate, even though the randomized 308 'phasing' of the buckets remains. 310 3.5.3. Optional enhancement: priority 312 The proposed Leaky bucket implementation could be modified to 313 support priority using multiple thresholds. 315 For instance, with two priorities it requires two thresholds TAU1 < 316 TAU2: 318 . All new requests would be admitted when the bucket fill is at 319 or below TAU1, 321 . Only higher priority requests would be admitted when the bucket 322 fill is between TAU1 and TAU2, 324 . All requests would be rejected when the bucket fill is above 325 TAU2. 327 This can be generalized to n priorities using n thresholds for n>2 328 in the obvious way. 330 4. Example 332 Adapting [draft-ietf-soc-overload-control-05] example in section 6.2 333 where SIP client P1 sends requests to a downstream server P2: 335 INVITE sips:user@example.com SIP/2.0 337 Via: SIP/2.0/TLS p1.example.net; 339 branch=z9hG4bK2d4790.1;received=192.0.2.111; 341 oc;oc-algo="loss,rate" 343 ... 345 SIP/2.0 100 Trying 347 Via: SIP/2.0/TLS p1.example.net; 349 branch=z9hG4bK2d4790.1;received=192.0.2.111; 351 oc-algo="rate";oc-validity=0; 353 oc-seq=1282321615.781 355 ... 357 In the messages above, the first line is sent by P1 to P2. This 358 line is a SIP request; because P1 supports overload control, it 359 inserts the "oc" parameter in the topmost Via header that it 360 created. P1 supports two overload control algorithms: loss and rate. 362 The second line --- a SIP response --- shows the topmost Via header 363 amended by P2 according to this specification and sent to P1. 364 Because P2 also supports overload control, it chooses the "rate" 365 based scheme and sends that back to P1 in the "oc-algo" parameter. 366 It uses oc-validity=0 to indicate no overload. 368 At some later time, P2 starts to experience overload. It sends the 369 following SIP message indicating P1 should send SIP requests at a 370 rate no greater than or equal to 150 SIP requests per seconds. 372 SIP/2.0 180 Ringing 374 Via: SIP/2.0/TLS p1.example.net; 376 branch=z9hG4bK2d4790.1;received=192.0.2.111; 377 oc=150;oc-algo="rate";oc-validity=1000; 379 oc-seq=1282321615.782 381 ... 383 5. Syntax 385 This specification extends the existing definition of the Via header 386 field parameters of [RFC3261] as follows: 388 oc = "oc" EQUAL oc-value 390 oc-value = "NaN" / oc-num 392 oc-num = 1*DIGIT 394 6. Security Considerations 396 To do: Use draft-ietf-soc-overload-control-05 section here. 398 7. IANA Considerations 400 None. 402 8. References 404 8.1. Normative References 406 [RFC3261] Rosenberg, J., Schulzrinne, H., Camarillo, G., Johnston, 407 A., Peterson, J., Sparks, R., Handley, M., and E. 408 Schooler, "SIP: Session Initiation Protocol", RFC 3261, 409 June 2002. 411 8.2. Informative References 413 [draft-ietf-soc-overload-control-05] 414 Gurbani, V., Hilt, V., Schulzrinne, H., "Session 415 Initiation Protocol (SIP) Overload Control", draft-ietf- 416 soc-overload-control-05. 418 [ITU-T Rec. I.371] 419 "Traffic control and congestion control in B-ISDN", ITU-T 420 Recommendation I.371. 422 [Erramilli] 423 A. Erramilli and L. J. Forys, "Traffic Synchronization 424 Effects In Teletraffic Systems", ITC-13, 1991. 426 Appendix A. Acknowledgments 428 Many thanks for the contributions, comments and feedback on this 429 document to: Janet Gunn. 431 This document was prepared using 2-Word-v2.0.template.dot. 433 Authors' Addresses 435 Eric Noel 436 AT&T Labs 437 200s Laurel Avenue 438 Middletown, NJ, 07747 439 USA 441 Philip M Williams 442 BT Innovate & Design 443 UK