Debian Codegraph @

rfc5039.html

raw
1 <pre>Network Working Group J. Rosenberg
2 Request for Comments: 5039 C. Jennings
3 Category: Informational Cisco
4 January 2008
5
6
7 <span class="h1">The Session Initiation Protocol (SIP) and Spam</span>
8
9 Status of This Memo
10
11 This memo provides information for the Internet community. It does
12 not specify an Internet standard of any kind. Distribution of this
13 memo is unlimited.
14
15 Abstract
16
17 Spam, defined as the transmission of bulk unsolicited messages, has
18 plagued Internet email. Unfortunately, spam is not limited to email.
19 It can affect any system that enables user-to-user communications.
20 The Session Initiation Protocol (SIP) defines a system for user-to-
21 user multimedia communications. Therefore, it is susceptible to
22 spam, just as email is. In this document, we analyze the problem of
23 spam in SIP. We first identify the ways in which the problem is the
24 same and the ways in which it is different from email. We then
25 examine the various possible solutions that have been discussed for
26 email and consider their applicability to SIP.
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52 <span class="grey">Rosenberg &amp; Jennings Informational [Page 1]</span></pre>
53 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-2" ></span>
54 <span class="grey"><a href="./rfc5039">RFC 5039</a> SIP Spam January 2008</span>
55
56
57 Table of Contents
58
59 <a href="#section-1">1</a>. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . <a href="#page-3">3</a>
60 <a href="#section-2">2</a>. Problem Definition . . . . . . . . . . . . . . . . . . . . . . <a href="#page-3">3</a>
61 <a href="#section-2.1">2.1</a>. Call Spam . . . . . . . . . . . . . . . . . . . . . . . . <a href="#page-4">4</a>
62 <a href="#section-2.2">2.2</a>. IM Spam . . . . . . . . . . . . . . . . . . . . . . . . . <a href="#page-7">7</a>
63 <a href="#section-2.3">2.3</a>. Presence Spam . . . . . . . . . . . . . . . . . . . . . . <a href="#page-7">7</a>
64 <a href="#section-3">3</a>. Solution Space . . . . . . . . . . . . . . . . . . . . . . . . <a href="#page-8">8</a>
65 <a href="#section-3.1">3.1</a>. Content Filtering . . . . . . . . . . . . . . . . . . . . <a href="#page-8">8</a>
66 <a href="#section-3.2">3.2</a>. Black Lists . . . . . . . . . . . . . . . . . . . . . . . <a href="#page-9">9</a>
67 <a href="#section-3.3">3.3</a>. White Lists . . . . . . . . . . . . . . . . . . . . . . . <a href="#page-9">9</a>
68 <a href="#section-3.4">3.4</a>. Consent-Based Communications . . . . . . . . . . . . . . . <a href="#page-10">10</a>
69 <a href="#section-3.5">3.5</a>. Reputation Systems . . . . . . . . . . . . . . . . . . . . <a href="#page-12">12</a>
70 <a href="#section-3.6">3.6</a>. Address Obfuscation . . . . . . . . . . . . . . . . . . . <a href="#page-14">14</a>
71 <a href="#section-3.7">3.7</a>. Limited-Use Addresses . . . . . . . . . . . . . . . . . . <a href="#page-14">14</a>
72 <a href="#section-3.8">3.8</a>. Turing Tests . . . . . . . . . . . . . . . . . . . . . . . <a href="#page-15">15</a>
73 <a href="#section-3.9">3.9</a>. Computational Puzzles . . . . . . . . . . . . . . . . . . <a href="#page-17">17</a>
74 <a href="#section-3.10">3.10</a>. Payments at Risk . . . . . . . . . . . . . . . . . . . . . <a href="#page-17">17</a>
75 <a href="#section-3.11">3.11</a>. Legal Action . . . . . . . . . . . . . . . . . . . . . . . <a href="#page-18">18</a>
76 <a href="#section-3.12">3.12</a>. Circles of Trust . . . . . . . . . . . . . . . . . . . . . <a href="#page-19">19</a>
77 <a href="#section-3.13">3.13</a>. Centralized SIP Providers . . . . . . . . . . . . . . . . <a href="#page-19">19</a>
78 <a href="#section-4">4</a>. Authenticated Identity in Email . . . . . . . . . . . . . . . <a href="#page-20">20</a>
79 <a href="#section-4.1">4.1</a>. Sender Checks . . . . . . . . . . . . . . . . . . . . . . <a href="#page-21">21</a>
80 <a href="#section-4.2">4.2</a>. Signature-Based Techniques . . . . . . . . . . . . . . . . <a href="#page-21">21</a>
81 <a href="#section-5">5</a>. Authenticated Identity in SIP . . . . . . . . . . . . . . . . <a href="#page-22">22</a>
82 <a href="#section-6">6</a>. Framework for Anti-Spam in SIP . . . . . . . . . . . . . . . . <a href="#page-23">23</a>
83 <a href="#section-7">7</a>. Additional Work . . . . . . . . . . . . . . . . . . . . . . . <a href="#page-24">24</a>
84 <a href="#section-8">8</a>. Security Considerations . . . . . . . . . . . . . . . . . . . <a href="#page-24">24</a>
85 <a href="#section-9">9</a>. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . <a href="#page-24">24</a>
86 <a href="#section-10">10</a>. Informative References . . . . . . . . . . . . . . . . . . . . <a href="#page-25">25</a>
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108 <span class="grey">Rosenberg &amp; Jennings Informational [Page 2]</span></pre>
109 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-3" ></span>
110 <span class="grey"><a href="./rfc5039">RFC 5039</a> SIP Spam January 2008</span>
111
112
113 <span class="h2"><a class="selflink" id="section-1" href="#section-1">1</a>. Introduction</span>
114
115 Spam, defined as the transmission of bulk unsolicited email, has been
116 a plague on the Internet email system. Many solutions have been
117 documented and deployed to counter the problem. None of these
118 solutions is ideal. However, one thing is clear: the spam problem
119 would be much less significant had solutions been deployed
120 ubiquitously before the problem became widespread.
121
122 The Session Initiation Protocol (SIP) [<a href="#ref-2" title="&quot;SIP: Session Initiation Protocol&quot;">2</a>] is used for multimedia
123 communications between users, including voice, video, instant
124 messaging, and presence. Consequently, it can be just as much of a
125 target for spam as email. To deal with this, solutions need to be
126 defined and recommendations put into place for dealing with spam as
127 soon as possible.
128
129 This document serves to meet those goals by defining the problem
130 space more concretely, analyzing the applicability of solutions used
131 in the email space, identifying protocol mechanisms that have been
132 defined for SIP that can help the problem, and making recommendations
133 for implementors.
134
135 <span class="h2"><a class="selflink" id="section-2" href="#section-2">2</a>. Problem Definition</span>
136
137 The spam problem in email is well understood, and we make no attempt
138 to further elaborate on it here. The question, however, is what is
139 the meaning of spam when applied to SIP? Since SIP covers a broad
140 range of functionality, there appear to be three related but
141 different manifestations:
142
143 Call Spam: This type of spam is defined as a bulk unsolicited set of
144 session initiation attempts (i.e., INVITE requests), attempting to
145 establish a voice, video, instant messaging [<a href="#ref-1" title="&quot;The Message Session Relay Protocol (MSRP)&quot;">1</a>], or other type of
146 communications session. If the user should answer, the spammer
147 proceeds to relay their message over the real-time media. This is
148 the classic telemarketer spam, applied to SIP. This is often
149 called SPam over Ip Telephony, or SPIT.
150
151 IM Spam: This type of spam is similar to email. It is defined as a
152 bulk unsolicited set of instant messages, whose content contains
153 the message that the spammer is seeking to convey. IM spam is
154 most naturally sent using the SIP MESSAGE [<a href="#ref-3" title="&quot;Session Initiation Protocol (SIP) Extension for Instant Messaging&quot;">3</a>] request. However,
155 any other request that causes content to automatically appear on
156 the user's display will also suffice. That might include INVITE
157 requests with large Subject headers (since the Subject is
158 sometimes rendered to the user), or INVITE requests with text or
159 HTML bodies. This is often called SPam over Instant Messaging, or
160 SPIM.
161
162
163
164 <span class="grey">Rosenberg &amp; Jennings Informational [Page 3]</span></pre>
165 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-4" ></span>
166 <span class="grey"><a href="./rfc5039">RFC 5039</a> SIP Spam January 2008</span>
167
168
169 Presence Spam: This type of spam is similar to IM spam. It is
170 defined as a bulk unsolicited set of presence requests (i.e.,
171 SUBSCRIBE requests [<a href="#ref-4" title="&quot;Session Initiation Protocol (SIP)-Specific Event Notification&quot;">4</a>] for the presence event package [<a href="#ref-6" title="&quot;A Presence Event Package for the Session Initiation Protocol (SIP)&quot;">6</a>]), in an
172 attempt to get on the "buddy list" or "white list" of a user in
173 order to send them IM or initiate other forms of communications.
174 This is occasionally called SPam over Presence Protocol, or SPPP.
175
176 There are many other SIP messages that a spammer might send.
177 However, most of the other ones do not result in content being
178 delivered to a user, nor do they seek input from a user. Rather,
179 they are answered by automata. OPTIONS is a good example of this.
180 There is little value for a spammer in sending an OPTIONS request,
181 since it is answered automatically by the User Agent Server (UAS).
182 No content is delivered to the user, and they are not consulted.
183
184 In the sections below, we consider the likelihood of these various
185 forms of SIP spam. This is done in some cases by a rough cost
186 analysis. It should be noted that all of these analyses are
187 approximate, and serve only to give a rough sense of the order of
188 magnitude of the problem.
189
190 <span class="h3"><a class="selflink" id="section-2.1" href="#section-2.1">2.1</a>. Call Spam</span>
191
192 Will call spam occur? That is an important question to answer.
193 Clearly, it does occur in the existing telephone network, in the form
194 of telemarketer calls. Although these calls are annoying, they do
195 not arrive in the same kind of volume as email spam. The difference
196 is cost; it costs more for the spammer to make a phone call than it
197 does to send email. This cost manifests itself in terms of the cost
198 for systems that can perform telemarketer call, and in cost per call.
199
200 Both of these costs are substantially reduced by SIP. A SIP call
201 spam application is easy to write. It is just a SIP User Agent that
202 initiates, in parallel, a large number of calls. If a call connects,
203 the spam application generates an ACK and proceeds to play out a
204 recorded announcement, and then it terminates the call. This kind of
205 application can be built entirely in software, using readily
206 available (and indeed, free) off-the-shelf software components. It
207 can run on a low-end PC and requires no special expertise to execute.
208
209 The cost per call is also substantially reduced. A normal
210 residential phone line allows only one call to be placed at a time.
211 If additional lines are required, a user must purchase more expensive
212 connectivity. Typically, a T1 or T3 would be required for a large-
213 volume telemarketing service. That kind of access is very expensive
214 and well beyond the reach of an average user. A T1 line is
215 approximately US $250 per month, and about 1.5 cents per minute for
216 calls. T1 lines used only for outbound calls (such as in this case)
217
218
219
220 <span class="grey">Rosenberg &amp; Jennings Informational [Page 4]</span></pre>
221 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-5" ></span>
222 <span class="grey"><a href="./rfc5039">RFC 5039</a> SIP Spam January 2008</span>
223
224
225 are even more expensive than inbound trunks due to the reciprocal
226 termination charges that a provider pays and receives.
227
228 There are two aspects to the capacity: the call attempt rate, and the
229 number of simultaneous successful calls that can be in progress. A
230 T1 would allow a spammer, at most, 24 simultaneous calls, and
231 assuming about 10 seconds for each call attempt, about 2.4 call
232 attempts per second. At high-volume calling, the per-minute rates
233 far exceed the flat monthly fee for the T1. The result is a cost of
234 250,000 microcents for each successful spam delivery, assuming 10
235 seconds of content.
236
237 With SIP, this cost is much reduced. Consider a spammer using a
238 typical broadband Internet connection that provides 500 Kbps of
239 upstream bandwidth. Initiating a call requires just a single INVITE
240 message. Assuming, for simplicity's sake, that this is 1 KB, a 500
241 Kbps upstream DSL or cable modem connection will allow about 62 call
242 attempts per second. A successful call requires enough bandwidth to
243 transmit a message to the receiver. Assuming a low compression codec
244 (say, G.723.1 at 5.3 Kbps), this requires approximately 16 Kbps after
245 RTP, UDP, and IP overheads. With 500 Kbps upstream bandwidth, this
246 means as many as 31 simultaneous calls can be in progress. With 10
247 seconds of content per call, that allows for 3.1 successful call
248 attempts per second. If broadband access is around $50/month, the
249 cost per successful voice spam is about 6.22 microcents each. This
250 assumes that calls can be made 24 hours a day, 30 days a month, which
251 may or may not be the case.
252
253 These figures indicate that SIP call spam is roughly four orders of
254 magnitude cheaper to send than traditional circuit-based telemarketer
255 calls. This low cost is certainly going to be very attractive to
256 spammers. Indeed, many spammers utilize computational and bandwidth
257 resources provided by others, by infecting their machines with
258 viruses that turn them into "zombies" that can be used to generate
259 spam. This can reduce the cost of call spam to nearly zero.
260
261 Even ignoring the zombie issue, this reduction in cost is even more
262 amplified for international calls. Currently, there are few
263 telemarketing calls across international borders, largely due to the
264 large cost of making international calls. This is one of the reasons
265 why the "do not call list", a United States national list of numbers
266 that telemarketers cannot call -- has been effective. The law only
267 affects U.S. companies, but since most telemarketing calls are
268 domestic, it has been effective. Unfortunately (and fortunately),
269 the IP network provides no boundaries of these sorts, and calls to
270 any SIP URI are possible from anywhere in the world. This will allow
271 for international spam at a significantly reduced cost.
272
273
274
275
276 <span class="grey">Rosenberg &amp; Jennings Informational [Page 5]</span></pre>
277 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-6" ></span>
278 <span class="grey"><a href="./rfc5039">RFC 5039</a> SIP Spam January 2008</span>
279
280
281 International spam is likely to be even more annoying than national
282 spam, since it may arrive in languages that the recipient doesn't
283 even speak.
284
285 These figures assume that the primary limitation is the access
286 bandwidth and not CPU, disk, or termination costs. Termination costs
287 merit further discussion. Currently, most Voice over IP (VoIP) calls
288 terminate on the Public Switched Telephone Network (PSTN), and this
289 termination costs the originator of the call money. These costs are
290 similar to the per-minute rates of a T1. It ranges anywhere from
291 half a cent to three cents per minute, depending on volume and other
292 factors. However, equipment costs, training, and other factors are
293 much lower for SIP-based termination than a T1, making the cost still
294 lower than circuit connectivity. Furthermore, the current trend in
295 VoIP systems is to make termination free for calls that never touch
296 the PSTN, that is, calls to actual SIP endpoints. Thus, as more and
297 more SIP endpoints come online, termination costs will probably drop.
298 Until then, SIP spam can be used in concert with termination services
299 for a lower-cost form of traditional telemarketer calls, made to
300 normal PSTN endpoints.
301
302 It is useful to compare these figures with email. VoIP can deliver
303 approximately 3.1 successful call attempts per second. Email spam
304 can, of course, deliver more. Assuming 1 KB per email, and an
305 upstream link of 500 Kbps, a spammer can generate 62.5 messages per
306 second. This number goes down with larger messages of course.
307 Interestingly, spam filters delete large numbers of these mails, so
308 the cost per viewed message is likely to be much higher. In that
309 sense, call spam is much more attractive, since its content is much
310 more likely to be examined by a user if a call attempt is successful.
311
312 Another part of the cost of spamming is collecting addresses.
313 Spammers have, over time, built up immense lists of email addresses,
314 each of the form user@domain, to which spam is directed. SIP uses
315 the same form of addressing, making it likely that email addresses
316 can easily be turned into valid SIP addresses. Telephone numbers
317 also represent valid SIP addresses; in concert with a termination
318 provider, a spammer can direct SIP calls at traditional PSTN devices.
319 It is not clear whether email spammers have also been collecting
320 phone numbers as they perform their Web sweeps, but it is probably
321 not hard to do so. Furthermore, unlike email addresses, phone
322 numbers are a finite address space and one that is fairly densely
323 packed. As a result, going sequentially through phone numbers is
324 likely to produce a fairly high hit rate. Thus, it seems like the
325 cost is relatively low for a spammer to obtain large numbers of SIP
326 addresses to which spam can be directed.
327
328
329
330
331
332 <span class="grey">Rosenberg &amp; Jennings Informational [Page 6]</span></pre>
333 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-7" ></span>
334 <span class="grey"><a href="./rfc5039">RFC 5039</a> SIP Spam January 2008</span>
335
336
337 <span class="h3"><a class="selflink" id="section-2.2" href="#section-2.2">2.2</a>. IM Spam</span>
338
339 IM spam is very much like email, in terms of the costs for deploying
340 and generating spam. Assuming, for the sake of argument, a 1KB
341 message to be sent and 500 Kbps of upstream bandwidth, that is 62.5
342 messages per second. At $50/month, the result is .31 microcents per
343 message. This is less than voice spam, but not substantially less.
344 The cost is probably on par with email spam. However, IM is much
345 more intrusive than email. In today's systems, IMs automatically pop
346 up and present themselves to the user. Email, of course, must be
347 deliberately selected and displayed. However, most popular IM
348 systems employ white lists, which only allow IM to be delivered if
349 the sender is on the white list. Thus, whether or not IM spam will
350 be useful seems to depend a lot on the nature of the systems as the
351 network is opened up. If they are ubiquitously deployed with white-
352 list access, the value of IM spam is likely to be low.
353
354 It is important to point out that there are two different types of IM
355 systems: page mode and session mode. Page mode IM systems work much
356 like email, with each IM being sent as a separate message. In
357 session mode IM, there is signaling in advance of communication to
358 establish a session, and then IMs are exchanged, perhaps point-to-
359 point, as part of the session. The modality impacts the types of
360 spam techniques that can be applied. Techniques for email can be
361 applied identically to page mode IM, but session mode IM is more like
362 telephony, and many techniques (such as content filtering) are harder
363 to apply.
364
365 <span class="h3"><a class="selflink" id="section-2.3" href="#section-2.3">2.3</a>. Presence Spam</span>
366
367 As defined above, presence spam is the generation of bulk unsolicited
368 SUBSCRIBE messages. The cost of this is within a small constant
369 factor of IM spam so the same cost estimates can be used here. What
370 would be the effect of such spam? Most presence systems provide some
371 kind of consent framework. A watcher that has not been granted
372 permission to see the user's presence will not gain access to their
373 presence. However, the presence request is usually noted and
374 conveyed to the user, allowing them to approve or deny the request.
375 In SIP, this is done using the watcherinfo event package [<a href="#ref-7" title="&quot;A Watcher Information Event Template-Package for the Session Initiation Protocol (SIP)&quot;">7</a>]. This
376 package allows a user to learn the identity of the watcher, in order
377 to make an authorization decision.
378
379 Interestingly, this provides a vehicle for conveying information to a
380 user. By generating SUBSCRIBE requests from identities such as
381 sip:[email protected], brief messages can be
382 conveyed to the user, even though the sender does not have, and never
383 will receive, permission to access presence. As such, presence spam
384 can be viewed as a form of IM spam, where the amount of content to be
385
386
387
388 <span class="grey">Rosenberg &amp; Jennings Informational [Page 7]</span></pre>
389 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-8" ></span>
390 <span class="grey"><a href="./rfc5039">RFC 5039</a> SIP Spam January 2008</span>
391
392
393 conveyed is limited. The limit is equal to the amount of information
394 generated by the watcher that gets conveyed to the user through the
395 permission system.
396
397 This type of spam also shows up in consent frameworks used to prevent
398 call spam, as discussed in <a href="#section-3.4">Section 3.4</a>.
399
400 <span class="h2"><a class="selflink" id="section-3" href="#section-3">3</a>. Solution Space</span>
401
402 In this section, we consider the various solutions that might be
403 possible to deal with SIP spam. We primarily consider techniques
404 that have been employed to deal with email spam. It is important to
405 note that the solutions documented below are not meant to be an
406 exhaustive study of the spam solutions used for email but rather just
407 a representative set. We also consider some solutions that appear to
408 be SIP-specific.
409
410 <span class="h3"><a class="selflink" id="section-3.1" href="#section-3.1">3.1</a>. Content Filtering</span>
411
412 The most common form of spam protection used in email is based on
413 content filtering. Spam filters analyze the content of the email,
414 and look for clues that the email is spam. Bayesian spam filters are
415 in this category.
416
417 Unfortunately, this type of spam filtering, while successful for
418 email spam, is completely useless for call spam. There are two
419 reasons. First, in the case where the user answers the call, the
420 call is already established and the user is paying attention before
421 the content is delivered. The spam cannot be analyzed before the
422 user sees it. Second, if the content is stored before the user
423 accesses it (e.g., with voicemail), the content will be in the form
424 of recorded audio or video. Speech and video recognition technology
425 is not likely to be good enough to analyze the content and determine
426 whether or not it is spam. Indeed, if a system tried to perform
427 speech recognition on a recording in order to perform such an
428 analysis, it would be easy for the spammers to make calls with
429 background noises, poor grammar, and varied accents, all of which
430 will throw off recognition systems. Video recognition is even harder
431 to do and remains primarily an area of research.
432
433 IM spam, due to its similarity to email, can be countered with
434 content analysis tools. Indeed, the same tools and techniques used
435 for email will directly work for IM spam.
436
437
438
439
440
441
442
443
444 <span class="grey">Rosenberg &amp; Jennings Informational [Page 8]</span></pre>
445 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-9" ></span>
446 <span class="grey"><a href="./rfc5039">RFC 5039</a> SIP Spam January 2008</span>
447
448
449 <span class="h3"><a class="selflink" id="section-3.2" href="#section-3.2">3.2</a>. Black Lists</span>
450
451 Black listing is an approach whereby the spam filter maintains a list
452 of addresses that identify spammers. These addresses include both
453 usernames ([email protected]) and entire domains (example.com).
454 Pure blacklists are not very effective in email for two reasons.
455 First, email addresses are easy to spoof, making it easy for the
456 sender to pretend to be someone else. If the sender varies the
457 addresses they send from, the black list becomes almost completely
458 useless. The second problem is that, even if the sender doesn't
459 forge the From address, email addresses are in almost limitless
460 supply. Each domain contains an infinite supply of email addresses,
461 and new domains can be obtained for very low cost. Furthermore,
462 there will always be public providers that will allow users to obtain
463 identities for almost no cost (for example, Yahoo or AOL mail
464 accounts). The entire domain cannot be blacklisted because it
465 contains so many valid users. Blacklisting needs to be for
466 individual users. Those identities are easily changed.
467
468 As a result, as long as identities are easy to manufacture, or
469 zombies are used, black lists will have limited effectiveness for
470 email.
471
472 Blacklists are also likely to be ineffective for SIP spam.
473 Mechanisms for inter-domain authenticated identity for email and SIP
474 are discussed in <a href="#section-4">Section 4</a> and <a href="#section-5">Section 5</a>. Assuming these mechanisms
475 are used and enabled in inter-domain communications, it becomes
476 difficult to forge sender addresses. However, it still remains cheap
477 to obtain a nearly infinite supply of addresses.
478
479 <span class="h3"><a class="selflink" id="section-3.3" href="#section-3.3">3.3</a>. White Lists</span>
480
481 White lists are the opposite of black lists. It is a list of valid
482 senders that a user is willing to accept email from. Unlike black
483 lists, a spammer cannot change identities to get around the white
484 list. White lists are susceptible to address spoofing, but a strong
485 identity authentication mechanism can prevent that problem. As a
486 result, the combination of white lists and strong identity, as
487 described in <a href="#section-4.2">Section 4.2</a> and <a href="#section-5">Section 5</a>, are a good form of defense
488 against spam.
489
490 However, they are not a complete solution, since they would prohibit
491 a user from ever being able to receive email from someone who was not
492 explicitly put on the white list. As a result, white lists require a
493 solution to the "introduction problem" - how to meet someone for the
494 first time, and decide whether they should be placed in the white
495 list. In addition to the introduction problem, white lists demand
496 time from the user to manage.
497
498
499
500 <span class="grey">Rosenberg &amp; Jennings Informational [Page 9]</span></pre>
501 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-10" ></span>
502 <span class="grey"><a href="./rfc5039">RFC 5039</a> SIP Spam January 2008</span>
503
504
505 In IM systems, white lists have proven exceptionally useful at
506 preventing spam. This is due, in no small part, to the fact that the
507 white list exists naturally in the form of the buddy list. Users
508 don't have to manage this list just for the purposes of spam
509 prevention; it provides general utility, and assists in spam
510 prevention for free. Many popular IM systems also have strong
511 identity mechanisms since they do not allow communications with IM
512 systems in other administrative domains. The introduction problem in
513 these systems is solved with a consent framework, described below.
514
515 The success of white lists in IM systems has applicability to SIP as
516 well. This is because SIP also provides a buddy list concept and has
517 an advanced presence system as part of its specifications. The
518 introduction problem remains. In email, techniques like Turing tests
519 have been employed to address the introduction problem. Turing tests
520 are considered further in the sections below. As with email, a
521 technique for solving the introduction problem would need to be
522 applied in conjunction with a white list.
523
524 If a user's computer is compromised and used a zombie, that computer
525 can usually be used to send spam to anyone that has put the user on
526 their white list.
527
528 <span class="h3"><a class="selflink" id="section-3.4" href="#section-3.4">3.4</a>. Consent-Based Communications</span>
529
530 A consent-based solution is used in conjunction with white or black
531 lists. That is, if user A is not on user B's white or black list,
532 and user A attempts to communicate with user B, user A's attempt is
533 initially rejected, and they are told that consent is being
534 requested. Next time user B connects, user B is informed that user A
535 had attempted communications. User B can then authorize or reject
536 user A.
537
538 These kinds of consent-based systems are used widely in presence and
539 IM. Since most of today's popular IM systems only allow
540 communications within a single administrative domain, sender
541 identities can be authenticated. Email often uses similar consent-
542 based systems for mailing lists. They use a form of authentication
543 based on sending cookies to an email address to verify that a user
544 can receive mail at that address.
545
546 This kind of consent-based communications has been standardized in
547 SIP for presence, using the watcher information event package [<a href="#ref-7" title="&quot;A Watcher Information Event Template-Package for the Session Initiation Protocol (SIP)&quot;">7</a>] and
548 data format [<a href="#ref-8" title="&quot;An Extensible Markup Language (XML) Based Format for Watcher Information&quot;">8</a>], which allow a user to find out that someone has
549 subscribed. Then, the XML Configuration Access Protocol (XCAP) [<a href="#ref-10" title="&quot;The Extensible Markup Language (XML) Configuration Access Protocol (XCAP)&quot;">10</a>]
550 is used, along with the XML format for presence authorization [<a href="#ref-11" title="&quot;Presence Authorization Rules&quot;">11</a>] to
551 provide permission for the user to communicate.
552
553
554
555
556 <span class="grey">Rosenberg &amp; Jennings Informational [Page 10]</span></pre>
557 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-11" ></span>
558 <span class="grey"><a href="./rfc5039">RFC 5039</a> SIP Spam January 2008</span>
559
560
561 A consent framework has also been developed that is applicable to
562 other forms of SIP communications [<a href="#ref-12" title="&quot;A Framework for Consent-Based Communications in the Session Initiation Protocol (SIP)&quot;">12</a>]. However, this framework
563 focuses on authorizing the addition of users to "mailing lists",
564 known as exploders in SIP terminology. Though spammers typically use
565 such exploder functions, presumably one run by a spammer would not
566 use this technique. Consequently, this consent framework is not
567 directly applicable to the spam problem. It is, however, useful as a
568 tool for managing a white list. Through the PUBLISH mechanism, it
569 allows a user to upload a permission document [<a href="#ref-13" title="&quot;A Document Format for Requesting Consent&quot;">13</a>] that indicates
570 that they will only accept incoming calls from a particular sender.
571
572 Can a consent framework, like the ones used for presence, help solve
573 call spam? At first glance, it would seem to help a lot. However,
574 it might just change the nature of the spam. Instead of being
575 bothered with content, in the form of call spam or IM spam, users are
576 bothered with consent requests. A user's "communications inbox"
577 might instead be filled with requests for communications from a
578 multiplicity of users. Those requests for communications don't
579 convey much useful content to the user, but they can convey some. At
580 the very least, they will convey the identity of the requester. The
581 user part of the SIP URI allows for limited free form text, and thus
582 could be used to convey brief messages. One can imagine receiving
583 consent requests with identities like
584 "sip:[email protected]", for
585 example. Fortunately, it is possible to apply traditional content
586 filtering systems to the header fields in the SIP messages, thus
587 reducing these kinds of consent request attacks.
588
589 In order for the spammer to convey more extensive content to the
590 user, the user must explicitly accept the request, and only then can
591 the spammer convey the full content. This is unlike email spam,
592 where, even though much spam is automatically deleted, some
593 percentage of the content does get through, and is seen by users,
594 without their explicit consent that they want to see it. Thus, if
595 consent is required first, the value in sending spam is reduced, and
596 perhaps it will cease for those spam cases where consent is not given
597 to spammers.
598
599 As such, the real question is whether or not the consent system would
600 make it possible for a user to give consent to non-spammers and
601 reject spammers. Authenticated identity can help. A user in an
602 enterprise would know to give consent to senders in other enterprises
603 in the same industry, for example. However, in the consumer space,
604 if sip:[email protected] tries to communicate with a user, how does
605 that user determine whether Bob is a spammer or a long-lost friend
606 from high school? There is no way based on the identity alone. In
607 such a case, a useful technique is to grant permission for Bob to
608 communicate but to ensure that the permission is extremely limited.
609
610
611
612 <span class="grey">Rosenberg &amp; Jennings Informational [Page 11]</span></pre>
613 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-12" ></span>
614 <span class="grey"><a href="./rfc5039">RFC 5039</a> SIP Spam January 2008</span>
615
616
617 In particular, Bob may be granted permission to send no more than 200
618 words of text in a single IM, which he can use to identify himself,
619 so that the user can determine whether or not more permissions are
620 appropriate. It may even be possible that an automated system could
621 do some form of content analysis on this initial short message.
622 However, this 200 words of text may be enough for a spammer to convey
623 their message, in much the same way they might convey it in the user
624 part of the SIP URI.
625
626 Thus, it seems that a consent-based framework, along with white lists
627 and black lists, cannot fully solve the problem for SIP, although it
628 does appear to help.
629
630 <span class="h3"><a class="selflink" id="section-3.5" href="#section-3.5">3.5</a>. Reputation Systems</span>
631
632 A reputation system is also used in conjunction with white or black
633 lists. Assume that user A is not on user B's white list, and A
634 attempts to contact user B. If a consent-based system is used, B is
635 prompted to consent to communications from A, and along with the
636 consent, a reputation score might be displayed in order to help B
637 decide whether or not they should accept communications from A.
638
639 Traditionally, reputation systems are implemented in highly
640 centralized messaging architectures; the most widespread reputation
641 systems in messaging today have been deployed by monolithic instant
642 messaging providers (though many Web sites with a high degree of
643 interactivity employ very similar concepts of reputation).
644 Reputation is calculated based on user feedback. For example, a
645 button on the user interface of the messaging client might empower
646 users to inform the system that a particular user is abusive. Of
647 course, the input of any single user has to be insufficient to ruin
648 one's reputation, but consistent negative feedback would give the
649 abusive user a negative reputation score.
650
651 Reputation systems have been successful in systems where
652 centralization of resources (user identities, authentication, etc.)
653 and monolithic control dominate. Examples of these include the large
654 instant messaging providers that run IM systems that do not exchange
655 messages with other administrative domains. That control, first of
656 all, provides a relatively strong identity assertion for users (since
657 all users trust a common provider, and the common provider is the
658 arbiter of authentication and identity). Secondly, it provides a
659 single place where reputation can be managed.
660
661 Reputation systems based on negative reputation scores suffer from
662 many of the same problems as black lists, since effectively the
663 consequence of having a negative reputation is that you are
664 blacklisted. If identities are very easy to acquire, a user with a
665
666
667
668 <span class="grey">Rosenberg &amp; Jennings Informational [Page 12]</span></pre>
669 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-13" ></span>
670 <span class="grey"><a href="./rfc5039">RFC 5039</a> SIP Spam January 2008</span>
671
672
673 negative reputation will simply acquire a new identity. Moreover,
674 negative reputation is generated by tattling, which requires users to
675 be annoyed enough to click the warning button -- a process that can
676 be abused. In some reputation systems, "reputation mafias"
677 consisting of large numbers of users routinely bully or extort
678 victims by threatening collectively to give victims a negative
679 reputation.
680
681 Reputation systems based on positive reputation, where users praise
682 each other for being good, rather than tattling on each other for
683 being bad, have some similar drawbacks. Collectives of spammers, or
684 just one spammer who acquires a large number identities, could praise
685 one another in order to create an artificial positive reputation.
686 Users similarly have to overcome the inertia required to press the
687 "praise" button. Unlike negative reputation systems, however,
688 positive reputation is not circumvented when users acquire a new
689 identity, since basing authorization decisions on positive reputation
690 is essentially a form of white listing.
691
692 So, while positive reputation systems are superior to negative
693 reputation systems, they are far from perfect. Intriguingly, though,
694 combining presence-based systems with reputation systems leads to an
695 interesting fusion. The "buddy-list" concept of presence is, in
696 effect, a white list - and one can infer that the users on one's
697 buddy list are people whom you are "praising". This eliminates the
698 problem of user inertia in the use of the "praise" button, and
699 automates the initial establishment of reputation.
700
701 And of course, your buddies in turn have buddies. Collectively, you
702 and your buddies (and their buddies, and so on) constitute a social
703 network of reputation. If there were a way to leverage this social
704 network, it would eliminate the need for centralization of the
705 reputation system. Your perception of a particular user's reputation
706 might be dependent on your relationship to them in the social
707 network: are they one buddy removed (strong reputation), four buddies
708 removed (weaker reputation), three buddies removed but connected to
709 you through several of your buddies, etc. This web of trust
710 furthermore would have the very desirable property that circles of
711 spammers adding one another to their own buddy lists would not affect
712 your perception of their reputation unless their circle linked to
713 your own social network.
714
715 If a users machine is compromised and turned into a zombie, this
716 allows SPAM to be sent and may impact their reputation in a negative
717 way. Once their reputation decreases, it becomes extremely difficult
718 to reestablish a positive reputation.
719
720
721
722
723
724 <span class="grey">Rosenberg &amp; Jennings Informational [Page 13]</span></pre>
725 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-14" ></span>
726 <span class="grey"><a href="./rfc5039">RFC 5039</a> SIP Spam January 2008</span>
727
728
729 <span class="h3"><a class="selflink" id="section-3.6" href="#section-3.6">3.6</a>. Address Obfuscation</span>
730
731 Spammers build up their spam lists by gathering email addresses from
732 Web sites and other public sources of information. One way to
733 minimize spam is to make your address difficult or impossible to
734 gather. Spam bots typically look for text in pages of the form
735 "user@domain", and assume that anything of that form is an email
736 address. To hide from such spam bots, many Web sites have recently
737 begun placing email addresses in an obfuscated form, usable to humans
738 but difficult for an automata to read as an email address. Examples
739 include forms such as, "user at example dot com" or "j d r o s e n a
740 t e x a m p l e d o t c o m".
741
742 These techniques are equally applicable to prevention of SIP spam,
743 and are likely to be as equally effective or ineffective in its
744 prevention.
745
746 It is worth mentioning that the source of addresses need not be a Web
747 site - any publicly accessible service containing addresses will
748 suffice. As a result, ENUM [<a href="#ref-9" title="&quot;The E.164 to Uniform Resource Identifiers (URI) Dynamic Delegation Discovery System (DDDS) Application (ENUM)&quot;">9</a>] has been cited as a potential gold
749 mine for spammers. It would allow a spammer to collect SIP and other
750 URIs by traversing the tree in e164.arpa and mining it for data.
751 This problem is mitigated in part if only number prefixes, as opposed
752 to actual numbers, appear in the DNS. Even in that case, however, it
753 provides a technique for a spammer to learn which phone numbers are
754 reachable through cheaper direct SIP connectivity.
755
756 <span class="h3"><a class="selflink" id="section-3.7" href="#section-3.7">3.7</a>. Limited-Use Addresses</span>
757
758 A related technique to address obfuscation is limited-use addresses.
759 In this technique, a user has a large number of email addresses at
760 their disposal, each of which has constraints on its applicability.
761 A limited-use address can be time-bound, so that it expires after a
762 fixed period. Or, a different email address can be given to each
763 correspondent. When spam arrives from that correspondent, the
764 limited-use address they were given is terminated. In another
765 variation, the same limited-use address is given to multiple users
766 that share some property; for example, all work colleagues, all
767 coworkers from different companies, all retailers, and so on. Should
768 spam begin arriving on one of the addresses, it is invalidated,
769 preventing communications from anyone else that received the limited
770 use address.
771
772 This technique is equally applicable to SIP. One of the drawbacks of
773 the approach is that it can make it hard for people to reach you; if
774 an email address you hand out to a friend becomes spammed, changing
775 it requires you to inform your friend of the new address. SIP can
776 help solve this problem in part, by making use of presence [<a href="#ref-6" title="&quot;A Presence Event Package for the Session Initiation Protocol (SIP)&quot;">6</a>].
777
778
779
780 <span class="grey">Rosenberg &amp; Jennings Informational [Page 14]</span></pre>
781 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-15" ></span>
782 <span class="grey"><a href="./rfc5039">RFC 5039</a> SIP Spam January 2008</span>
783
784
785 Instead of handing out your email address to your friends, you would
786 hand out your presence URI. When a friend wants to send you an
787 email, they subscribe to your presence (indeed, they are likely to be
788 continuously subscribed from a buddy list application). The presence
789 data can include an email address where you can be reached. This
790 email address can be obfuscated and be of single use, different for
791 each buddy who requests your presence. They can also be constantly
792 changed, as these changes are pushed directly to your buddies. In a
793 sense, the buddy list represents an automatically updated address
794 book, and would therefore eliminate the problem.
795
796 Another approach is to give a different address to each and every
797 correspondent, so that it is never necessary to tell a "good" user
798 that an address needs to be changed. This is an extreme form of
799 limited-use addresses, which can be called a single-use address.
800 Mechanisms are available in SIP for the generation of [<a href="#ref-16" title="&quot;Applying Loose Routing to Session Initiation Protocol (SIP) User Agents (UA)&quot;">16</a>] an
801 infinite supply of single use addresses. However, the hard part
802 remains a useful mechanism for distribution and management of those
803 addresses.
804
805 <span class="h3"><a class="selflink" id="section-3.8" href="#section-3.8">3.8</a>. Turing Tests</span>
806
807 In email, Turing tests are mechanisms whereby the sender of the
808 message is given some kind of puzzle or challenge, which only a human
809 can answer (since Turing tests rely on video or audio puzzles, they
810 sometimes cannot be solved by individuals with handicaps). These
811 tests are also known as captchas (Completely Automated Public Turing
812 test to tell Computers and Humans Apart). If the puzzle is answered
813 correctly, the sender is placed on the user's white list. These
814 puzzles frequently take the form of recognizing a word or sequence of
815 numbers in an image with a lot of background noise. The tests need
816 to be designed such that automata cannot easily perform the image
817 recognition needed to extract the word or number sequence, but a
818 human user usually can. Designing such tests is not easy, since
819 ongoing advances in image processing and artificial intelligence
820 continually raise the bar. Consequently, the effectiveness of
821 captchas are tied to whether spammers can come up with or obtain
822 algorithms for automatically solving them.
823
824 Like many of the other email techniques, Turing tests are dependent
825 on sender identity, which cannot easily be authenticated in email.
826
827 Turing tests can be used to prevent IM spam in much the same way they
828 can be used to prevent email spam.
829
830 Turing tests can be applied to call spam as well, although not
831 directly, because call spam does not usually involve the transfer of
832 images and other content that can be used to verify that a human is
833
834
835
836 <span class="grey">Rosenberg &amp; Jennings Informational [Page 15]</span></pre>
837 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-16" ></span>
838 <span class="grey"><a href="./rfc5039">RFC 5039</a> SIP Spam January 2008</span>
839
840
841 on the other end. If most of the calls are voice, the technique
842 needs to be adapted to voice. This is not that difficult to do.
843 Here is how it could be done. User A calls user B and is not on user
844 B's white or black list. User A is transferred to an Interactive
845 Voice Response (IVR) system. The IVR system tells the user that they
846 are going to hear a series of numbers (say 5 of them), and that they
847 have to enter those numbers on the keypad. The IVR system reads out
848 the numbers while background music is playing, making it difficult
849 for an automated speech recognition system to be applied to the
850 media. The user then enters the numbers on their keypad. If they
851 are entered correctly, the user is added to the white list.
852
853 This kind of voice-based Turing test is easily extended to a variety
854 of media, such as video and text, and user interfaces by making use
855 of the SIP application interaction framework [<a href="#ref-14" title="&quot;A Framework for Application Interaction in the Session Initiation Protocol (SIP)&quot;">14</a>]. This framework
856 allows client devices to interact with applications in the network,
857 where such interaction is done with stimulus signaling, including
858 keypads (supported with the Keypad Markup Language [<a href="#ref-15" title="&quot;A Session Initiation Protocol (SIP) Event Package for Key Press Stimulus (KPML)&quot;">15</a>]), but also
859 including Web browsers, voice recognition, and so on. The framework
860 allows the application to determine the media capabilities of the
861 device (or user, in cases where they are handicapped) and interact
862 with them appropriately.
863
864 In the case of voice, the Turing test would need to be made to run in
865 the language of the caller. This is possible in SIP, using the
866 Accept-Language header field, though this is not widely used at the
867 moment, and meant for languages of SIP message components, not the
868 media streams.
869
870 The primary problem with the voice Turing test is the same one that
871 email tests have: instead of having an automata process the test, a
872 spammer can pay cheap workers to take the tests. Assuming cheap
873 labor in a poor country can be obtained for about 60 cents per hour,
874 and assuming a Turing test of a 30-second duration, this is about
875 0.50 cents per test and thus 0.50 cents per message to send an IM
876 spam. Lower labor rates would reduce this further; the number quoted
877 here is based on real online bids in September of 2006 made for
878 actual work of this type.
879
880 As an alternative to paying cheap workers to take the tests, the
881 tests can be taken by human users that are tricked into completing
882 the tests in order to gain access to what they believe is a
883 legitimate resource. This was done by a spambot that posted the
884 tests on a pornography site, and required users to complete the tests
885 in order to gain access to content.
886
887 Due to these limitations, Turing tests may never completely solve the
888 problem.
889
890
891
892 <span class="grey">Rosenberg &amp; Jennings Informational [Page 16]</span></pre>
893 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-17" ></span>
894 <span class="grey"><a href="./rfc5039">RFC 5039</a> SIP Spam January 2008</span>
895
896
897 <span class="h3"><a class="selflink" id="section-3.9" href="#section-3.9">3.9</a>. Computational Puzzles</span>
898
899 This technique is similar to Turing tests. When user A tries to
900 communicate with user B, user B asks user A to perform a computation
901 and pass the result back. This computation has to be something a
902 human user cannot perform and something expensive enough to increase
903 user A's cost to communicate. This cost increase has to be high
904 enough to make it prohibitively expensive for spammers but
905 inconsequential for legitimate users.
906
907 One of the problems with the technique is that there is wide
908 variation in the computational power of the various clients that
909 might legitimately communicate. The CPU speed on a low-end cell
910 phone is around 50 MHz, while a high-end PC approaches 5 GHz. This
911 represents almost two orders of magnitude difference. Thus, if the
912 test is designed to be reasonable for a cell phone to perform, it is
913 two orders of magnitude cheaper to perform for a spammer on a high-
914 end machine. Recent research has focused on defining computational
915 puzzles that challenge the CPU/memory bandwidth, as opposed to just
916 the CPU [<a href="#ref-26" title="&quot;Moderately Hard, Memory Bound Functions, NDSS 2003&quot;">26</a>]. It seems that there is less variety in the CPU/memory
917 bandwidth across devices, roughly a single order of magnitude.
918
919 Recent work [<a href="#ref-28" title="&quot;Proof of Work Proves not to Work, Third Annual Workshop on Economics and Information Security&quot;">28</a>] suggests that, due to the ability of spammers to use
920 virus-infected machines (also known as zombies) to generate the spam,
921 the amount of computational power available to the spammers is
922 substantial, and it may be impossible to have them compute a puzzle
923 that is sufficiently hard that will not also block normal emails. If
924 combined with white listing, computational puzzles would only be
925 utilized for new communications partners. Of course, if the partner
926 on the white list is a zombie, spam will come from that source. The
927 frequency of communications with new partners is arguably higher for
928 email than for multimedia, and thus the computational puzzle
929 techniques may be more effective for SIP than for email in dealing
930 with the introduction problem.
931
932 These techniques are an active area of research right now, and any
933 results for email are likely to be usable for SIP.
934
935 <span class="h3"><a class="selflink" id="section-3.10" href="#section-3.10">3.10</a>. Payments at Risk</span>
936
937 This approach has been proposed for email [<a href="#ref-27" title="&quot;Bankable Postage for Network Services, Proceedings of the 8th Asian Computing Science Conference, Mumbai, India&quot;">27</a>]. When user A sends
938 email to user B, user A deposits a small amount of money (say, one
939 dollar) into user B's account. If user B decides that the message is
940 not spam, user B refunds this money back to user A. If the message
941 is spam, user B keeps the money. This technique requires two
942 transactions to complete: a transfer from A to B, and a transfer from
943 B back to A. The first transfer has to occur before the message can
944 be received in order to avoid reuse of "pending payments" across
945
946
947
948 <span class="grey">Rosenberg &amp; Jennings Informational [Page 17]</span></pre>
949 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-18" ></span>
950 <span class="grey"><a href="./rfc5039">RFC 5039</a> SIP Spam January 2008</span>
951
952
953 several messages, which would eliminate the utility of the solution.
954 The second one then needs to occur when the message is found not to
955 be spam.
956
957 This technique appears just as applicable to call spam and IM spam as
958 it is to email spam. Like many of the other techniques, this
959 exchange would only happen the first time you talk to people. Its
960 proper operation therefore requires a good authenticated identity
961 infrastructure.
962
963 This technique has the potential to make it arbitrarily expensive to
964 send spam of any sort. However, it relies on cheap micro-payment
965 techniques on the Internet. Traditional costs for Internet payments
966 are around 25 cents per transaction, which would probably be
967 prohibitive. However, recent providers have been willing to charge
968 15% of the transaction for small transactions, as small as one cent.
969 This cost would have to be shouldered by users of the system. The
970 cost that would need to be shouldered per user is equal to the number
971 of messages from unknown senders (that is, senders not on the white
972 list) that are received. For a busy user, assume about 10 new
973 senders per day. If the deposit is 5 cents, the transaction provider
974 would take 0.75 cents and deliver 4.25 cents. If the sender is
975 allowed, the recipient returns 4.25 cents, the provider takes 0.64
976 cents, and returns 3.6 cents. This costs the sender 0.65 cents on
977 each transaction, if it was legitimate. If there are ten new
978 recipients per day, that is US $1.95 per month, which is relatively
979 inexpensive.
980
981 Assuming a micro-payment infrastructure exists, another problem with
982 payment-at-risk is that it loses effectiveness when there are strong
983 inequities in the value of currency between sender and recipient.
984 For example, a poor person in a Third World country might keep the
985 money in each mail message, regardless of whether it is spam.
986 Similarly, a poor person might not be willing to include money in an
987 email, even if legitimate, for fear that the recipient might keep it.
988 If the amount of money is lowered to help handle these problems, it
989 might become sufficiently small that spammers can just afford to
990 spend it.
991
992 <span class="h3"><a class="selflink" id="section-3.11" href="#section-3.11">3.11</a>. Legal Action</span>
993
994 In this solution, countries pass laws that prohibit spam. These laws
995 could apply to IM or call spam just as easily as they could apply to
996 email spam. There is a lot of debate about whether these laws would
997 really be effective in preventing spam.
998
999 As a recent example in the US, "do not call" lists seem to be
1000 effective. However, due to the current cost of long-distance phone
1001
1002
1003
1004 <span class="grey">Rosenberg &amp; Jennings Informational [Page 18]</span></pre>
1005 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-19" ></span>
1006 <span class="grey"><a href="./rfc5039">RFC 5039</a> SIP Spam January 2008</span>
1007
1008
1009 calls, the telemarketing is coming from companies within the US. As
1010 such, calls from such telemarketers can be traced. If a telemarketer
1011 violates the "do not call" list, the trace allows legal action to be
1012 taken against them. A similar "do not irritate" list for VoIP or for
1013 email would be less likely to work because the spam is likely to come
1014 from international sources. This problem could be obviated if there
1015 was a strong way to identify the sender's legal entity, and then
1016 determine whether it was in a jurisdiction where it was practical to
1017 take legal action against them. If the spammer is not in such a
1018 jurisdiction, the SIP spam could be rejected.
1019
1020 There are also schemes that cause laws other than anti-spam laws to
1021 be broken if spam is sent. This does not inherently reduce SPAM, but
1022 it allows more legal options to be brought to bear against the
1023 spammer. For example, Habeas &lt;<a href="http://www.habeas.com">http://www.habeas.com</a>&gt; inserts
1024 material in the header that, if it was inserted by a spammer without
1025 an appropriate license, would allegedly causes the spammer to violate
1026 US copyright and trademark laws, possibly reciprocal laws, and
1027 similar laws in many countries.
1028
1029 <span class="h3"><a class="selflink" id="section-3.12" href="#section-3.12">3.12</a>. Circles of Trust</span>
1030
1031 In this model, a group of domains (e.g., a set of enterprises) all
1032 get together. They agree to exchange SIP calls amongst each other,
1033 and they also agree to introduce a fine should any one of them be
1034 caught spamming. Each company would then enact measures to terminate
1035 employees who spam from their accounts.
1036
1037 This technique relies on secure inter-domain authentication - that
1038 is, domain B can know that messages are received from domain A. In
1039 SIP, this is readily provided by usage of the mutually authenticated
1040 Transport Level Security (TLS)[<a href="#ref-22" title="&quot;The Transport Layer Security (TLS) Protocol Version 1.1&quot;">22</a>] between providers or SIP Identity
1041 [<a href="#ref-17" title="&quot;Enhancements for Authenticated Identity Management in the Session Initiation Protocol (SIP)&quot;">17</a>].
1042
1043 This kind of technique works well for small domains or small sets of
1044 providers, where these policies can be easily enforced. However, it
1045 is unclear how well it scales up. Could a very large domain truly
1046 prevent its users from spamming? At what point would the network be
1047 large enough that it would be worthwhile to send spam and just pay
1048 the fine? How would the pricing be structured to allow both small
1049 and large domains alike to participate?
1050
1051 <span class="h3"><a class="selflink" id="section-3.13" href="#section-3.13">3.13</a>. Centralized SIP Providers</span>
1052
1053 This technique is a variation on the circles of trust described in
1054 <a href="#section-3.12">Section 3.12</a>. A small number of providers get established as "inter-
1055 domain SIP providers". These providers act as a SIP-equivalent to
1056 the interexchange carriers in the PSTN. Every enterprise, consumer
1057
1058
1059
1060 <span class="grey">Rosenberg &amp; Jennings Informational [Page 19]</span></pre>
1061 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-20" ></span>
1062 <span class="grey"><a href="./rfc5039">RFC 5039</a> SIP Spam January 2008</span>
1063
1064
1065 SIP provider, or other SIP network (call these the local SIP
1066 providers) connects to one of these inter-domain providers. The
1067 local SIP providers only accept SIP messages from their chosen inter-
1068 domain provider. The inter-domain provider charges the local
1069 provider, per SIP message, for the delivery of SIP messages to other
1070 local providers. The local provider can choose to pass on this cost
1071 to its own customers if it so chooses.
1072
1073 The inter-domain SIP providers then form bi-lateral agreements with
1074 each other, exchanging SIP messages according to strict contracts.
1075 These contracts require that each of the inter-domain providers be
1076 responsible for charging a minimum per-message fee to their own
1077 customers. Extensive auditing procedures can be put into place to
1078 verify this. Besides such contracts, there may or may not be a flow
1079 of funds between the inter-domain providers.
1080
1081 The result of such a system is that a fixed cost can be associated
1082 with sending a SIP message, and that this cost does not require
1083 micro-payments to be exchanged between local providers, as it does in
1084 <a href="#section-3.10">Section 3.10</a>. Since all of the relationships are pre-established and
1085 negotiated, cheaper techniques for monetary transactions (such as
1086 monthly post-paid transactions) can be used.
1087
1088 This technique can be made to work in SIP, whereas it cannot in
1089 email, because inter-domain SIP connectivity has not yet been broadly
1090 established. In email, there already exists a no-cost form of inter-
1091 domain connectivity that cannot be eliminated without destroying the
1092 utility of email. If, however, SIP inter-domain communications get
1093 established from the start using this structure, there is a path to
1094 deployment.
1095
1096 This structure is more or less the same as the one in place for the
1097 PSTN today, and since there is relatively little spam on the PSTN
1098 (compared to email!), there is some proof that this kind of
1099 arrangement can work. However, centralized architectures as these
1100 are deliberately eschewed because they put back into SIP much of the
1101 complexity and monopolistic structures that the protocol aims to
1102 eliminate.
1103
1104 <span class="h2"><a class="selflink" id="section-4" href="#section-4">4</a>. Authenticated Identity in Email</span>
1105
1106 Though not a form of anti-spam in and of itself, authenticated or
1107 verifiable identities are a key part of making other anti-spam
1108 mechanisms work. Many of the techniques described above are most
1109 effective when combined with a white or black list, which itself
1110 requires a strong form of identity.
1111
1112
1113
1114
1115
1116 <span class="grey">Rosenberg &amp; Jennings Informational [Page 20]</span></pre>
1117 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-21" ></span>
1118 <span class="grey"><a href="./rfc5039">RFC 5039</a> SIP Spam January 2008</span>
1119
1120
1121 In email, two types of authenticated identity have been developed -
1122 sender checks and signature-based solutions.
1123
1124 <span class="h3"><a class="selflink" id="section-4.1" href="#section-4.1">4.1</a>. Sender Checks</span>
1125
1126 In email, DNS resource records have been defined that will allow a
1127 domain that receives a message to verify that the sender is a valid
1128 Message Transfer Agent (MTA) for the sending domain [<a href="#ref-18" title="&quot;SMTP Service Extension for Indicating the Responsible Submitter of an E-Mail Message&quot;">18</a>] [<a href="#ref-19" title="&quot;Sender ID: Authenticating E-Mail&quot;">19</a>] [<a href="#ref-20" title="&quot;Purported Responsible Address in E-Mail Messages&quot;">20</a>]
1129 [<a href="#ref-21" title="&quot;Sender Policy Framework (SPF) for Authorizing Use of Domains in E-Mail, Version 1&quot;">21</a>]. They don't prevent spam by themselves, but may help in
1130 preventing spoofed emails. As has been mentioned several times, a
1131 form of strong authenticated identity is key in making many other
1132 anti-spam techniques work.
1133
1134 Are these techniques useful for SIP? They can be used for SIP but
1135 are not necessary. In SIP, TLS with mutual authentication can be
1136 used inter-domain. A provider receiving a message can then reject
1137 any message coming from a domain that does not match the asserted
1138 identity of the sender of the message. Such a policy only works in
1139 the "trapezoid" model of SIP, whereby there are only two domains in
1140 any call - the sending domain, which is where the originator resides,
1141 and the receiving domain. These techniques are discussed in <a href="./rfc3261#section-26.3.2.2">Section</a>
1142 <a href="./rfc3261#section-26.3.2.2">26.3.2.2 of RFC 3261</a> [<a href="#ref-2" title="&quot;SIP: Session Initiation Protocol&quot;">2</a>]. In forwarding situations, the assumption
1143 no longer holds and these techniques no longer work. However, the
1144 authenticated identity mechanism for SIP, discussed in <a href="#section-5">Section 5</a>,
1145 does work in more complex network configurations and provides fairly
1146 strong assertion of identity.
1147
1148 <span class="h3"><a class="selflink" id="section-4.2" href="#section-4.2">4.2</a>. Signature-Based Techniques</span>
1149
1150 Domain Keys Identified Mail (DKIM) Signatures [<a href="#ref-23" title="&quot;DomainKeys Identified Mail (DKIM) Signatures&quot;">23</a>] (and several non-
1151 standard techniques that preceded it) provide strong identity
1152 assertions by allowing the sending domain to sign an email, and then
1153 providing mechanisms by which the receiving MTA or Mail User Agent
1154 (MUA) can validate the signature.
1155
1156 Unfortunately, when used with blacklists, this kind of authenticated
1157 identity is only as useful as the fraction of the emails that utilize
1158 it. This is partly true for white lists as well; if any
1159 unauthenticated email is accepted for an address on a white list, a
1160 spammer can spoof that address. However, a white list can be
1161 effective with limited deployment of DKIM if all the people on the
1162 white list are those whose domains are utilizing the mechanism, and
1163 the users on that white list aren't zombies.
1164
1165 This kind of identity mechanism is also applicable to SIP, and is in
1166 fact, exactly what is defined by SIP's authenticated identity
1167 mechanism [<a href="#ref-17" title="&quot;Enhancements for Authenticated Identity Management in the Session Initiation Protocol (SIP)&quot;">17</a>].
1168
1169
1170
1171
1172 <span class="grey">Rosenberg &amp; Jennings Informational [Page 21]</span></pre>
1173 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-22" ></span>
1174 <span class="grey"><a href="./rfc5039">RFC 5039</a> SIP Spam January 2008</span>
1175
1176
1177 Other signature-based approaches for email include S/MIME[24] and
1178 OpenPGP[25].
1179
1180 <span class="h2"><a class="selflink" id="section-5" href="#section-5">5</a>. Authenticated Identity in SIP</span>
1181
1182 One of the key parts of many of the solutions described above is the
1183 ability to securely identify the sender of a SIP message. SIP
1184 provides a secure solution for this problem, called SIP Identity
1185 [<a href="#ref-17" title="&quot;Enhancements for Authenticated Identity Management in the Session Initiation Protocol (SIP)&quot;">17</a>], and it is important to discuss it here.
1186
1187 The solution starts by having each domain authenticate its own users.
1188 SIP provides HTTP digest authentication as part of the core SIP
1189 specification, and all clients and servers are required to support
1190 it. Indeed, digest is widely deployed for SIP. However, digest
1191 alone has many known vulnerabilities, most notably offline dictionary
1192 attacks. These vulnerabilities are all resolved by having each
1193 client maintain a persistent TLS connection to the server. The
1194 client verifies the server identity using TLS, and then authenticates
1195 itself to the server using a digest exchange over TLS. This
1196 technique, which is also documented in <a href="./rfc3261">RFC 3261</a>, is very secure but
1197 not widely deployed yet. In the long term, this approach will be
1198 necessary for the security properties needed to prevent SIP spam.
1199
1200 Once a domain has authenticated the identity of a user, when it
1201 relays a message from that user to another domain, the sending domain
1202 can assert the identity of the sender, and include a signature to
1203 validate that assertion. This is done using the SIP identity
1204 mechanism [<a href="#ref-17" title="&quot;Enhancements for Authenticated Identity Management in the Session Initiation Protocol (SIP)&quot;">17</a>].
1205
1206 A weaker form of identity assertion is possible using the P-Asserted-
1207 Identity header field [<a href="#ref-5" title="&quot;Private Extensions to the Session Initiation Protocol (SIP) for Asserted Identity within Trusted Networks&quot;">5</a>], but this technique requires mutual trust
1208 among all domains. Unfortunately, this becomes exponentially harder
1209 to provide as the number of interconnected domains grows. As that
1210 happens, the value of the identity assertion becomes equal to the
1211 trustworthiness of the least trustworthy domain. Since spam is a
1212 consequence of the receiving domain not being able to trust the
1213 sending domains to disallow the hosts in the sending to send spam,
1214 the P-Asserted-Identity technique becomes ineffective at exactly the
1215 same levels of interconnectedness that introduce spam.
1216
1217 Consider the following example to help illustrate this fact. A
1218 malicious domain -- let us call them spam.example.com, would like to
1219 send SIP INVITE requests with false P-Asserted-Identity, indicating
1220 users outside of its own domain. spam.example.com finds a regional
1221 SIP provider in a small country who, due to its small size and
1222 disinterest in spam, accepts any P-Asserted-Identity from its
1223 customers without verification. This provider, in turn, connects to
1224 a larger, interconnect provider. They do ask each of their customers
1225
1226
1227
1228 <span class="grey">Rosenberg &amp; Jennings Informational [Page 22]</span></pre>
1229 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-23" ></span>
1230 <span class="grey"><a href="./rfc5039">RFC 5039</a> SIP Spam January 2008</span>
1231
1232
1233 to verify P-Asserted-Identity but have no easy way of enforcing it.
1234 This provider, in turn, connects to everyone else. As a consequence,
1235 the spam.example.com domain is able to inject calls with a spoofed
1236 caller ID. This request can be directed to any recipient reachable
1237 through the network (presumably everyone due to the large size of the
1238 root provider). There is no way for a recipient to know that this
1239 particular P-Asserted-Identity came from this bad spam.example.com
1240 domain. As the example shows, even though the central provider's
1241 policy is good, the overall effectiveness of P-Asserted-Identity is
1242 still only as good as the policies of the weakest link in the chain.
1243
1244 SIP also defines the usage of TLS between domains, using mutual
1245 authentication, as part of the base specification. This technique
1246 provides a way for one domain to securely determine that it is
1247 talking to a server that is a valid representative of another domain.
1248
1249 <span class="h2"><a class="selflink" id="section-6" href="#section-6">6</a>. Framework for Anti-Spam in SIP</span>
1250
1251 Unfortunately, there is no magic bullet for preventing SIP spam, just
1252 as there is none for email spam. However, the combination of several
1253 techniques can provide a framework for dealing with spam in SIP.
1254 This section provides recommendations for network designers in order
1255 to help mitigate the risk of spam.
1256
1257 There are four core recommendations that can be made:
1258
1259 Strong Identity: Firstly, in almost all of the solutions discussed
1260 above, there is a dependency on the ability to authenticate the
1261 sender of a SIP message inter-domain. Consent, reputation
1262 systems, computational puzzles, and payments at risk, amongst
1263 others, all work best when applied only to new requests, and
1264 successful completion of an introduction results in the placement
1265 of a user on a white list. However, usage of white lists depends
1266 on strong identity assertions. Consequently, any network that
1267 interconnects with others should make use of strong SIP identity
1268 as described in <a href="./rfc4474">RFC 4474</a>. P-Asserted-Identity is not strong
1269 enough.
1270
1271 White Lists: Secondly, with a strong identity system in place,
1272 networks are recommended to make use of white lists. These are
1273 ideally built off existing buddy lists, if present. If not,
1274 separate white lists can be managed for spam. Placement on these
1275 lists can be manual or based on the successful completion of one
1276 or more introduction mechanisms.
1277
1278 Solve the Introduction Problem: This in turn leads to the final
1279 recommendation to be made. Network designers should make use of
1280 one or more mechanisms meant to solve the introduction problem.
1281
1282
1283
1284 <span class="grey">Rosenberg &amp; Jennings Informational [Page 23]</span></pre>
1285 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-24" ></span>
1286 <span class="grey"><a href="./rfc5039">RFC 5039</a> SIP Spam January 2008</span>
1287
1288
1289 Indeed, it is possible to use more than one and combine the
1290 results through some kind of weight. A user that successfully
1291 completes the introduction mechanism can be automatically added to
1292 the white list. Of course, that can only be done usefully if
1293 their identity is verified by SIP Identity. The set of mechanisms
1294 for solving the introduction problem, as described in this
1295 document, are based on some (but not all) of the techniques known
1296 and used at the time of writing. Providers of SIP services should
1297 keep tabs on solutions in email as they evolve, and utilize the
1298 best of what those techniques have to offer.
1299
1300 Don't Wait Until It's Too Late: But perhaps most importantly,
1301 providers should not ignore the spam problem until it happens! As
1302 soon as a provider inter-connects with other providers, or allows
1303 SIP messages from the open Internet, that provider must consider
1304 how they will deal with spam.
1305
1306 <span class="h2"><a class="selflink" id="section-7" href="#section-7">7</a>. Additional Work</span>
1307
1308 Though the above framework serves as a good foundation on which to
1309 deal with spam in SIP, there are gaps, some of which can be addressed
1310 by additional work that has yet to be undertaken.
1311
1312 One of the difficulties with the strong identity techniques is that a
1313 receiver of a SIP request without an authenticated identity cannot
1314 know whether the request lacked such an identity because the
1315 originating domain didn't support it, or because a man-in-the-middle
1316 removed it. As a result, transition mechanisms should be put in
1317 place to allow these to be differentiated. Without it, the value of
1318 the identity mechanism is much reduced.
1319
1320 <span class="h2"><a class="selflink" id="section-8" href="#section-8">8</a>. Security Considerations</span>
1321
1322 This document is entirely devoted to issues relating to spam in SIP
1323 and references a variety of security mechanisms in support of that
1324 goal.
1325
1326 <span class="h2"><a class="selflink" id="section-9" href="#section-9">9</a>. Acknowledgements</span>
1327
1328 The authors would like to thank Rohan Mahy for providing information
1329 on Habeas, Baruch Sterman for providing costs on VoIP termination
1330 services, and Gonzalo Camarillo and Vijay Gurbani for their reviews.
1331 Useful comments and feedback were provided by Nils Ohlmeir, Tony
1332 Finch, Randy Gellens, Lisa Dusseault, Sam Hartman, Chris Newman, Tim
1333 Polk, Donald Eastlake, and Yakov Shafranovich. Jon Peterson wrote
1334 some of the text in this document and has contributed to the work as
1335 it has moved along.
1336
1337
1338
1339
1340 <span class="grey">Rosenberg &amp; Jennings Informational [Page 24]</span></pre>
1341 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-25" ></span>
1342 <span class="grey"><a href="./rfc5039">RFC 5039</a> SIP Spam January 2008</span>
1343
1344
1345 <span class="h2"><a class="selflink" id="section-10" href="#section-10">10</a>. Informative References</span>
1346
1347 [<a id="ref-1">1</a>] Campbell, B., Mahy, R., and C. Jennings, "The Message Session
1348 Relay Protocol (MSRP)", <a href="./rfc4975">RFC 4975</a>, September 2007.
1349
1350 [<a id="ref-2">2</a>] Rosenberg, J., Schulzrinne, H., Camarillo, G., Johnston, A.,
1351 Peterson, J., Sparks, R., Handley, M., and E. Schooler, "SIP:
1352 Session Initiation Protocol", <a href="./rfc3261">RFC 3261</a>, June 2002.
1353
1354 [<a id="ref-3">3</a>] Campbell, B., Rosenberg, J., Schulzrinne, H., Huitema, C., and
1355 D. Gurle, "Session Initiation Protocol (SIP) Extension for
1356 Instant Messaging", <a href="./rfc3428">RFC 3428</a>, December 2002.
1357
1358 [<a id="ref-4">4</a>] Roach, A., "Session Initiation Protocol (SIP)-Specific Event
1359 Notification", <a href="./rfc3265">RFC 3265</a>, June 2002.
1360
1361 [<a id="ref-5">5</a>] Jennings, C., Peterson, J., and M. Watson, "Private Extensions
1362 to the Session Initiation Protocol (SIP) for Asserted Identity
1363 within Trusted Networks", <a href="./rfc3325">RFC 3325</a>, November 2002.
1364
1365 [<a id="ref-6">6</a>] Rosenberg, J., "A Presence Event Package for the Session
1366 Initiation Protocol (SIP)", <a href="./rfc3856">RFC 3856</a>, August 2004.
1367
1368 [<a id="ref-7">7</a>] Rosenberg, J., "A Watcher Information Event Template-Package
1369 for the Session Initiation Protocol (SIP)", <a href="./rfc3857">RFC 3857</a>,
1370 August 2004.
1371
1372 [<a id="ref-8">8</a>] Rosenberg, J., "An Extensible Markup Language (XML) Based
1373 Format for Watcher Information", <a href="./rfc3858">RFC 3858</a>, August 2004.
1374
1375 [<a id="ref-9">9</a>] Faltstrom, P. and M. Mealling, "The E.164 to Uniform Resource
1376 Identifiers (URI) Dynamic Delegation Discovery System (DDDS)
1377 Application (ENUM)", <a href="./rfc3761">RFC 3761</a>, April 2004.
1378
1379 [<a id="ref-10">10</a>] Rosenberg, J., "The Extensible Markup Language (XML)
1380 Configuration Access Protocol (XCAP)", <a href="./rfc4825">RFC 4825</a>, May 2007.
1381
1382 [<a id="ref-11">11</a>] Rosenberg, J., "Presence Authorization Rules", <a href="./rfc5025">RFC 5025</a>,
1383 October 2007.
1384
1385 [<a id="ref-12">12</a>] Rosenberg, J., "A Framework for Consent-Based Communications in
1386 the Session Initiation Protocol (SIP)", Work in Progress,
1387 October 2007.
1388
1389 [<a id="ref-13">13</a>] Camarillo, G., <a style="text-decoration: none" href='https://www.google.com/search?sitesearch=datatracker.ietf.org%2Fdoc%2Fhtml%2F&amp;q=inurl:draft-+%22A+Document+Format+for+Requesting+Consent%22'>"A Document Format for Requesting Consent"</a>, Work
1390 in Progress, October 2007.
1391
1392
1393
1394
1395
1396 <span class="grey">Rosenberg &amp; Jennings Informational [Page 25]</span></pre>
1397 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-26" ></span>
1398 <span class="grey"><a href="./rfc5039">RFC 5039</a> SIP Spam January 2008</span>
1399
1400
1401 [<a id="ref-14">14</a>] Rosenberg, J., "A Framework for Application Interaction in the
1402 Session Initiation Protocol (SIP)", Work in Progress,
1403 October 2005.
1404
1405 [<a id="ref-15">15</a>] Burger, E. and M. Dolly, "A Session Initiation Protocol (SIP)
1406 Event Package for Key Press Stimulus (KPML)", <a href="./rfc4730">RFC 4730</a>,
1407 November 2006.
1408
1409 [<a id="ref-16">16</a>] Rosenberg, J., "Applying Loose Routing to Session Initiation
1410 Protocol (SIP) User Agents (UA)", Work in Progress, June 2007.
1411
1412 [<a id="ref-17">17</a>] Peterson, J. and C. Jennings, "Enhancements for Authenticated
1413 Identity Management in the Session Initiation Protocol (SIP)",
1414 <a href="./rfc4474">RFC 4474</a>, August 2006.
1415
1416 [<a id="ref-18">18</a>] Allman, E. and H. Katz, "SMTP Service Extension for Indicating
1417 the Responsible Submitter of an E-Mail Message", <a href="./rfc4405">RFC 4405</a>,
1418 April 2006.
1419
1420 [<a id="ref-19">19</a>] Lyon, J. and M. Wong, "Sender ID: Authenticating E-Mail",
1421 <a href="./rfc4406">RFC 4406</a>, April 2006.
1422
1423 [<a id="ref-20">20</a>] Lyon, J., "Purported Responsible Address in E-Mail Messages",
1424 <a href="./rfc4407">RFC 4407</a>, April 2006.
1425
1426 [<a id="ref-21">21</a>] Wong, M. and W. Schlitt, "Sender Policy Framework (SPF) for
1427 Authorizing Use of Domains in E-Mail, Version 1", <a href="./rfc4408">RFC 4408</a>,
1428 April 2006.
1429
1430 [<a id="ref-22">22</a>] Dierks, T. and E. Rescorla, "The Transport Layer Security (TLS)
1431 Protocol Version 1.1", <a href="./rfc4346">RFC 4346</a>, April 2006.
1432
1433 [<a id="ref-23">23</a>] Allman, E., Callas, J., Delany, M., Libbey, M., Fenton, J., and
1434 M. Thomas, "DomainKeys Identified Mail (DKIM) Signatures",
1435 <a href="./rfc4871">RFC 4871</a>, May 2007.
1436
1437 [<a id="ref-24">24</a>] Ramsdell, B., "Secure/Multipurpose Internet Mail Extensions
1438 (S/MIME) Version 3.1 Message Specification", <a href="./rfc3851">RFC 3851</a>,
1439 July 2004.
1440
1441 [<a id="ref-25">25</a>] Elkins, M., Del Torto, D., Levien, R., and T. Roessler, "MIME
1442 Security with OpenPGP", <a href="./rfc3156">RFC 3156</a>, August 2001.
1443
1444 [<a id="ref-26">26</a>] Abadi, M., Burrows, M., Manasse, M., and T. Wobber, "Moderately
1445 Hard, Memory Bound Functions, NDSS 2003", February 2003.
1446
1447
1448
1449
1450
1451
1452 <span class="grey">Rosenberg &amp; Jennings Informational [Page 26]</span></pre>
1453 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-27" ></span>
1454 <span class="grey"><a href="./rfc5039">RFC 5039</a> SIP Spam January 2008</span>
1455
1456
1457 [<a id="ref-27">27</a>] Abadi, M., Burrows, M., Birrell, A., Dabek, F., and T. Wobber,
1458 "Bankable Postage for Network Services, Proceedings of the 8th
1459 Asian Computing Science Conference, Mumbai, India",
1460 December 2003.
1461
1462 [<a id="ref-28">28</a>] Clayton, R. and B. Laurie, "Proof of Work Proves not to Work,
1463 Third Annual Workshop on Economics and Information Security",
1464 May 2004.
1465
1466 Authors' Addresses
1467
1468 Jonathan Rosenberg
1469 Cisco
1470 Edison, NJ
1471 US
1472
1473 EMail: [email protected]
1474 URI: <a href="http://www.jdrosen.net">http://www.jdrosen.net</a>
1475
1476
1477 Cullen Jennings
1478 Cisco
1479 170 West Tasman Dr.
1480 San Jose, CA 95134
1481 US
1482
1483 Phone: +1 408 421-9990
1484 EMail: [email protected]
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508 <span class="grey">Rosenberg &amp; Jennings Informational [Page 27]</span></pre>
1509 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-28" ></span>
1510 <span class="grey"><a href="./rfc5039">RFC 5039</a> SIP Spam January 2008</span>
1511
1512
1513 Full Copyright Statement
1514
1515 Copyright (C) The IETF Trust (2008).
1516
1517 This document is subject to the rights, licenses and restrictions
1518 contained in <a href="https://www.rfc-editor.org/bcp/bcp78">BCP 78</a>, and except as set forth therein, the authors
1519 retain all their rights.
1520
1521 This document and the information contained herein are provided on an
1522 "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS
1523 OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY, THE IETF TRUST AND
1524 THE INTERNET ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS
1525 OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF
1526 THE INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED
1527 WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.
1528
1529 Intellectual Property
1530
1531 The IETF takes no position regarding the validity or scope of any
1532 Intellectual Property Rights or other rights that might be claimed to
1533 pertain to the implementation or use of the technology described in
1534 this document or the extent to which any license under such rights
1535 might or might not be available; nor does it represent that it has
1536 made any independent effort to identify any such rights. Information
1537 on the procedures with respect to rights in RFC documents can be
1538 found in <a href="https://www.rfc-editor.org/bcp/bcp78">BCP 78</a> and <a href="https://www.rfc-editor.org/bcp/bcp79">BCP 79</a>.
1539
1540 Copies of IPR disclosures made to the IETF Secretariat and any
1541 assurances of licenses to be made available, or the result of an
1542 attempt made to obtain a general license or permission for the use of
1543 such proprietary rights by implementers or users of this
1544 specification can be obtained from the IETF on-line IPR repository at
1545 <a href="http://www.ietf.org/ipr">http://www.ietf.org/ipr</a>.
1546
1547 The IETF invites any interested party to bring to its attention any
1548 copyrights, patents or patent applications, or other proprietary
1549 rights that may cover technology that may be required to implement
1550 this standard. Please address the information to the IETF at
1551 [email protected].
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564 Rosenberg &amp; Jennings Informational [Page 28]
1565 </pre>