| blob | fe2ed5d744b1ac35fb593aa2906c70142cf64267 |
2 Tor directory protocol, version 3
4 0. Scope and preliminaries
6 This directory protocol is used by Tor version 0.2.0.x-alpha and later.
7 See dir-spec-v1.txt for information on the protocol used up to the
8 0.1.0.x series, and dir-spec-v2.txt for information on the protocol
9 used by the 0.1.1.x and 0.1.2.x series.
11 Caches and authorities must still support older versions of the
12 directory protocols, until the versions of Tor that require them are
13 finally out of commission.
15 This document merges and supersedes the following proposals:
17 101 Voting on the Tor Directory System
18 103 Splitting identity key from regularly used signing key
19 104 Long and Short Router Descriptors
21 XXX when to download certificates.
22 XXX timeline
23 XXX fill in XXXXs
25 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL
26 NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and
27 "OPTIONAL" in this document are to be interpreted as described in
28 RFC 2119.
30 0.1. History
32 The earliest versions of Onion Routing shipped with a list of known
33 routers and their keys. When the set of routers changed, users needed to
34 fetch a new list.
36 The Version 1 Directory protocol
37 --------------------------------
39 Early versions of Tor (0.0.2) introduced "Directory authorities": servers
40 that served signed "directory" documents containing a list of signed
41 "router descriptors", along with short summary of the status of each
42 router. Thus, clients could get up-to-date information on the state of
43 the network automatically, and be certain that the list they were getting
44 was attested by a trusted directory authority.
46 Later versions (0.0.8) added directory caches, which download
47 directories from the authorities and serve them to clients. Non-caches
48 fetch from the caches in preference to fetching from the authorities, thus
49 distributing bandwidth requirements.
51 Also added during the version 1 directory protocol were "router status"
52 documents: short documents that listed only the up/down status of the
53 routers on the network, rather than a complete list of all the
54 descriptors. Clients and caches would fetch these documents far more
55 frequently than they would fetch full directories.
57 The Version 2 Directory Protocol
58 --------------------------------
60 During the Tor 0.1.1.x series, Tor revised its handling of directory
61 documents in order to address two major problems:
63 * Directories had grown quite large (over 1MB), and most directory
64 downloads consisted mainly of router descriptors that clients
65 already had.
67 * Every directory authority was a trust bottleneck: if a single
68 directory authority lied, it could make clients believe for a time
69 an arbitrarily distorted view of the Tor network. (Clients
70 trusted the most recent signed document they downloaded.) Thus,
71 adding more authorities would make the system less secure, not
72 more.
74 To address these, we extended the directory protocol so that
75 authorities now published signed "network status" documents. Each
76 network status listed, for every router in the network: a hash of its
77 identity key, a hash of its most recent descriptor, and a summary of
78 what the authority believed about its status. Clients would download
79 the authorities' network status documents in turn, and believe
80 statements about routers iff they were attested to by more than half of
81 the authorities.
83 Instead of downloading all router descriptors at once, clients
84 downloaded only the descriptors that they did not have. Descriptors
85 were indexed by their digests, in order to prevent malicious caches
86 from giving different versions of a router descriptor to different
87 clients.
89 Routers began working harder to upload new descriptors only when their
90 contents were substantially changed.
93 0.2. Goals of the version 3 protocol
95 Version 3 of the Tor directory protocol tries to solve the following
96 issues:
98 * A great deal of bandwidth used to transmit router descriptors was
99 used by two fields that are not actually used by Tor routers
100 (namely read-history and write-history). We save about 60% by
101 moving them into a separate document that most clients do not
102 fetch or use.
104 * It was possible under certain perverse circumstances for clients
105 to download an unusual set of network status documents, thus
106 partitioning themselves from clients who have a more recent and/or
107 typical set of documents. Even under the best of circumstances,
108 clients were sensitive to the ages of the network status documents
109 they downloaded. Therefore, instead of having the clients
110 correlate multiple network status documents, we have the
111 authorities collectively vote on a single consensus network status
112 document.
114 * The most sensitive data in the entire network (the identity keys
115 of the directory authorities) needed to be stored unencrypted so
116 that the authorities can sign network-status documents on the fly.
117 Now, the authorities' identity keys are stored offline, and used
118 to certify medium-term signing keys that can be rotated.
120 0.3. Some Remaining questions
122 Things we could solve on a v3 timeframe:
124 The SHA-1 hash is showing its age. We should do something about our
125 dependency on it. We could probably future-proof ourselves here in
126 this revision, at least so far as documents from the authorities are
127 concerned.
129 Too many things about the authorities are hardcoded by IP.
131 Perhaps we should start accepting longer identity keys for routers
132 too.
134 Things to solve eventually:
136 Requiring every client to know about every router won't scale forever.
138 Requiring every directory cache to know every router won't scale
139 forever.
142 1. Outline
144 There is a small set (say, around 5-10) of semi-trusted directory
145 authorities. A default list of authorities is shipped with the Tor
146 software. Users can change this list, but are encouraged not to do so,
147 in order to avoid partitioning attacks.
149 Every authority has a very-secret, long-term "Authority Identity Key".
150 This is stored encrypted and/or offline, and is used to sign "key
151 certificate" documents. Every key certificate contains a medium-term
152 (3-12 months) "authority signing key", that is used by the authority to
153 sign other directory information. (Note that the authority identity
154 key is distinct from the router identity key that the authority uses
155 in its role as an ordinary router.)
157 Routers periodically upload signed "routers descriptors" to the
158 directory authorities describing their keys, capabilities, and other
159 information. Routers may also upload signed "extra info documents"
160 containing information that is not required for the Tor protocol.
161 Directory authorities serve router descriptors indexed by router
162 identity, or by hash of the descriptor.
164 Routers may act as directory caches to reduce load on the directory
165 authorities. They announce this in their descriptors.
167 Periodically, each directory authority generates a view of
168 the current descriptors and status for known routers. They send a
169 signed summary of this view (a "status vote") to the other
170 authorities. The authorities compute the result of this vote, and sign
171 a "consensus status" document containing the result of the vote.
173 Directory caches download, cache, and re-serve consensus documents.
175 Clients, directory caches, and directory authorities all use consensus
176 documents to find out when their list of routers is out-of-date.
177 (Directory authorities also use vote statuses.) If it is, they download
178 any missing router descriptors. Clients download missing descriptors
179 from caches; caches and authorities download from authorities.
180 Descriptors are downloaded by the hash of the descriptor, not by the
181 relay's identity key: this prevents directory servers from attacking
182 clients by giving them descriptors nobody else uses.
184 All directory information is uploaded and downloaded with HTTP.
186 [Authorities also generate and caches also cache documents produced and
187 used by earlier versions of this protocol; see dir-spec-v1.txt and
188 dir-spec-v2.txt for notes on those versions.]
190 1.1. What's different from version 2?
192 Clients used to download multiple network status documents,
193 corresponding roughly to "status votes" above. They would compute the
194 result of the vote on the client side.
196 Authorities used to sign documents using the same private keys they used
197 for their roles as routers. This forced them to keep these extremely
198 sensitive keys in memory unencrypted.
200 All of the information in extra-info documents used to be kept in the
201 main descriptors.
203 1.2. Document meta-format
205 Router descriptors, directories, and running-routers documents all obey the
206 following lightweight extensible information format.
208 The highest level object is a Document, which consists of one or more
209 Items. Every Item begins with a KeywordLine, followed by zero or more
210 Objects. A KeywordLine begins with a Keyword, optionally followed by
211 whitespace and more non-newline characters, and ends with a newline. A
212 Keyword is a sequence of one or more characters in the set [A-Za-z0-9-].
213 An Object is a block of encoded data in pseudo-Open-PGP-style
214 armor. (cf. RFC 2440)
216 More formally:
218 NL = The ascii LF character (hex value 0x0a).
219 Document ::= (Item | NL)+
220 Item ::= KeywordLine Object*
221 KeywordLine ::= Keyword NL | Keyword WS ArgumentChar+ NL
222 Keyword = KeywordChar+
223 KeywordChar ::= 'A' ... 'Z' | 'a' ... 'z' | '0' ... '9' | '-'
224 ArgumentChar ::= any printing ASCII character except NL.
225 WS = (SP | TAB)+
226 Object ::= BeginLine Base-64-encoded-data EndLine
227 BeginLine ::= "-----BEGIN " Keyword "-----" NL
228 EndLine ::= "-----END " Keyword "-----" NL
230 The BeginLine and EndLine of an Object must use the same keyword.
232 When interpreting a Document, software MUST ignore any KeywordLine that
233 starts with a keyword it doesn't recognize; future implementations MUST NOT
234 require current clients to understand any KeywordLine not currently
235 described.
237 The "opt" keyword was used until Tor 0.1.2.5-alpha for non-critical future
238 extensions. All implementations MUST ignore any item of the form "opt
239 keyword ....." when they would not recognize "keyword ....."; and MUST
240 treat "opt keyword ....." as synonymous with "keyword ......" when keyword
241 is recognized.
243 Implementations before 0.1.2.5-alpha rejected any document with a
244 KeywordLine that started with a keyword that they didn't recognize.
245 When generating documents that need to be read by older versions of Tor,
246 implementations MUST prefix items not recognized by older versions of
247 Tor with an "opt" until those versions of Tor are obsolete. [Note that
248 key certificates, status vote documents, extra info documents, and
249 status consensus documents will never be read by older versions of Tor.]
251 Other implementations that want to extend Tor's directory format MAY
252 introduce their own items. The keywords for extension items SHOULD start
253 with the characters "x-" or "X-", to guarantee that they will not conflict
254 with keywords used by future versions of Tor.
256 In our document descriptions below, we tag Items with a multiplicity in
257 brackets. Possible tags are:
259 "At start, exactly once": These items MUST occur in every instance of
260 the document type, and MUST appear exactly once, and MUST be the
261 first item in their documents.
263 "Exactly once": These items MUST occur exactly one time in every
264 instance of the document type.
266 "At end, exactly once": These items MUST occur in every instance of
267 the document type, and MUST appear exactly once, and MUST be the
268 last item in their documents.
270 "At most once": These items MAY occur zero or one times in any
271 instance of the document type, but MUST NOT occur more than once.
273 "Any number": These items MAY occur zero, one, or more times in any
274 instance of the document type.
276 "Once or more": These items MUST occur at least once in any instance
277 of the document type, and MAY occur more.
279 1.3. Signing documents
281 Every signable document below is signed in a similar manner, using a
282 given "Initial Item", a final "Signature Item", a digest algorithm, and
283 a signing key.
285 The Initial Item must be the first item in the document.
287 The Signature Item has the following format:
289 <signature item keyword> [arguments] NL SIGNATURE NL
291 The "SIGNATURE" Object contains a signature (using the signing key) of
292 the PKCS1-padded digest of the entire document, taken from the
293 beginning of the Initial item, through the newline after the Signature
294 Item's keyword and its arguments.
296 Unless otherwise, the digest algorithm is SHA-1.
298 All documents are invalid unless signed with the correct signing key.
300 The "Digest" of a document, unless stated otherwise, is its digest *as
301 signed by this signature scheme*.
303 1.4. Voting timeline
305 Every consensus document has a "valid-after" (VA) time, a "fresh-until"
306 (FU) time and a "valid-until" (VU) time. VA MUST precede FU, which MUST
307 in turn precede VU. Times are chosen so that every consensus will be
308 "fresh" until the next consensus becomes valid, and "valid" for a while
309 after. At least 3 consensuses should be valid at any given time.
311 The timeline for a given consensus is as follows:
313 VA-DistSeconds-VoteSeconds: The authorities exchange votes.
315 VA-DistSeconds-VoteSeconds/2: The authorities try to download any
316 votes they don't have.
318 VA-DistSeconds: The authorities calculate the consensus and exchange
319 signatures.
321 VA-DistSeconds/2: The authorities try to download any signatures
322 they don't have.
324 VA: All authorities have a multiply signed consensus.
326 VA ... FU: Caches download the consensus. (Note that since caches have
327 no way of telling what VA and FU are until they have downloaded
328 the consensus, they assume that the present consensus's VA is
329 equal to the previous one's FU, and that its FU is one interval after
330 that.)
332 FU: The consensus is no longer the freshest consensus.
334 FU ... (the current consensus's VU): Clients download the consensus.
335 (See note above: clients guess that the next consensus's FU will be
336 two intervals after the current VA.)
338 VU: The consensus is no longer valid.
340 VoteSeconds and DistSeconds MUST each be at least 20 seconds; FU-VA and
341 VU-FU MUST each be at least 5 minutes.
343 2. Router operation and formats
345 ORs SHOULD generate a new router descriptor and a new extra-info
346 document whenever any of the following events have occurred:
348 - A period of time (18 hrs by default) has passed since the last
349 time a descriptor was generated.
351 - A descriptor field other than bandwidth or uptime has changed.
353 - Bandwidth has changed by a factor of 2 from the last time a
354 descriptor was generated, and at least a given interval of time
355 (20 mins by default) has passed since then.
357 - Its uptime has been reset (by restarting).
359 [XXX this list is incomplete; see router_differences_are_cosmetic()
360 in routerlist.c for others]
362 ORs SHOULD NOT publish a new router descriptor or extra-info document
363 if none of the above events have occurred and not much time has passed
364 (12 hours by default).
366 After generating a descriptor, ORs upload them to every directory
367 authority they know, by posting them (in order) to the URL
369 http://<hostname:port>/tor/
371 2.1. Router descriptor format
373 Router descriptors consist of the following items. For backward
374 compatibility, there should be an extra NL at the end of each router
375 descriptor.
377 In lines that take multiple arguments, extra arguments SHOULD be
378 accepted and ignored. Many of the nonterminals below are defined in
379 section 2.3.
381 "router" nickname address ORPort SOCKSPort DirPort NL
383 [At start, exactly once.]
385 Indicates the beginning of a router descriptor. "nickname" must be a
386 valid router nickname as specified in 2.3. "address" must be an IPv4
387 address in dotted-quad format. The last three numbers indicate the
388 TCP ports at which this OR exposes functionality. ORPort is a port at
389 which this OR accepts TLS connections for the main OR protocol;
390 SOCKSPort is deprecated and should always be 0; and DirPort is the
391 port at which this OR accepts directory-related HTTP connections. If
392 any port is not supported, the value 0 is given instead of a port
393 number. (At least one of DirPort and ORPort SHOULD be set;
394 authorities MAY reject any descriptor with both DirPort and ORPort of
395 0.)
397 "bandwidth" bandwidth-avg bandwidth-burst bandwidth-observed NL
399 [Exactly once]
401 Estimated bandwidth for this router, in bytes per second. The
402 "average" bandwidth is the volume per second that the OR is willing to
403 sustain over long periods; the "burst" bandwidth is the volume that
404 the OR is willing to sustain in very short intervals. The "observed"
405 value is an estimate of the capacity this relay can handle. The
406 relay remembers the max bandwidth sustained output over any ten
407 second period in the past day, and another sustained input. The
408 "observed" value is the lesser of these two numbers.
410 "platform" string NL
412 [At most once]
414 A human-readable string describing the system on which this OR is
415 running. This MAY include the operating system, and SHOULD include
416 the name and version of the software implementing the Tor protocol.
418 "published" YYYY-MM-DD HH:MM:SS NL
420 [Exactly once]
422 The time, in GMT, when this descriptor (and its corresponding
423 extra-info document if any) was generated.
425 "fingerprint" fingerprint NL
427 [At most once]
429 A fingerprint (a HASH_LEN-byte of asn1 encoded public key, encoded in
430 hex, with a single space after every 4 characters) for this router's
431 identity key. A descriptor is considered invalid (and MUST be
432 rejected) if the fingerprint line does not match the public key.
434 [We didn't start parsing this line until Tor 0.1.0.6-rc; it should
435 be marked with "opt" until earlier versions of Tor are obsolete.]
437 "hibernating" bool NL
439 [At most once]
441 If the value is 1, then the Tor relay was hibernating when the
442 descriptor was published, and shouldn't be used to build circuits.
444 [We didn't start parsing this line until Tor 0.1.0.6-rc; it should be
445 marked with "opt" until earlier versions of Tor are obsolete.]
447 "uptime" number NL
449 [At most once]
451 The number of seconds that this OR process has been running.
453 "onion-key" NL a public key in PEM format
455 [Exactly once]
457 This key is used to encrypt EXTEND cells for this OR. The key MUST be
458 accepted for at least 1 week after any new key is published in a
459 subsequent descriptor. It MUST be 1024 bits.
461 "signing-key" NL a public key in PEM format
463 [Exactly once]
465 The OR's long-term identity key. It MUST be 1024 bits.
467 "accept" exitpattern NL
468 "reject" exitpattern NL
470 [Any number]
472 These lines describe an "exit policy": the rules that an OR follows
473 when deciding whether to allow a new stream to a given address. The
474 'exitpattern' syntax is described below. There MUST be at least one
475 such entry. The rules are considered in order; if no rule matches,
476 the address will be accepted. For clarity, the last such entry SHOULD
477 be accept *:* or reject *:*.
479 "router-signature" NL Signature NL
481 [At end, exactly once]
483 The "SIGNATURE" object contains a signature of the PKCS1-padded
484 hash of the entire router descriptor, taken from the beginning of the
485 "router" line, through the newline after the "router-signature" line.
486 The router descriptor is invalid unless the signature is performed
487 with the router's identity key.
489 "contact" info NL
491 [At most once]
493 Describes a way to contact the relay's administrator, preferably
494 including an email address and a PGP key fingerprint.
496 "family" names NL
498 [At most once]
500 'Names' is a space-separated list of relay nicknames or
501 hexdigests. If two ORs list one another in their "family" entries,
502 then OPs should treat them as a single OR for the purpose of path
503 selection.
505 For example, if node A's descriptor contains "family B", and node B's
506 descriptor contains "family A", then node A and node B should never
507 be used on the same circuit.
509 "read-history" YYYY-MM-DD HH:MM:SS (NSEC s) NUM,NUM,NUM,NUM,NUM... NL
510 [At most once]
511 "write-history" YYYY-MM-DD HH:MM:SS (NSEC s) NUM,NUM,NUM,NUM,NUM... NL
512 [At most once]
514 Declare how much bandwidth the OR has used recently. Usage is divided
515 into intervals of NSEC seconds. The YYYY-MM-DD HH:MM:SS field
516 defines the end of the most recent interval. The numbers are the
517 number of bytes used in the most recent intervals, ordered from
518 oldest to newest.
520 [We didn't start parsing these lines until Tor 0.1.0.6-rc; they should
521 be marked with "opt" until earlier versions of Tor are obsolete.]
523 [See also migration notes in section 2.2.1.]
525 "eventdns" bool NL
527 [At most once]
529 Declare whether this version of Tor is using the newer enhanced
530 dns logic. Versions of Tor with this field set to false SHOULD NOT
531 be used for reverse hostname lookups.
533 [This option is obsolete. All Tor current relays should be presumed
534 to have the evdns backend.]
536 "caches-extra-info" NL
538 [At most once.]
540 Present only if this router is a directory cache that provides
541 extra-info documents.
543 [Versions before 0.2.0.1-alpha don't recognize this, and versions
544 before 0.1.2.5-alpha will reject descriptors containing it unless
545 it is prefixed with "opt"; it should be so prefixed until these
546 versions are obsolete.]
548 "extra-info-digest" digest NL
550 [At most once]
552 "Digest" is a hex-encoded digest (using upper-case characters) of the
553 router's extra-info document, as signed in the router's extra-info
554 (that is, not including the signature). (If this field is absent, the
555 router is not uploading a corresponding extra-info document.)
557 [Versions before 0.2.0.1-alpha don't recognize this, and versions
558 before 0.1.2.5-alpha will reject descriptors containing it unless
559 it is prefixed with "opt"; it should be so prefixed until these
560 versions are obsolete.]
562 "hidden-service-dir" *(SP VersionNum) NL
564 [At most once.]
566 Present only if this router stores and serves hidden service
567 descriptors. If any VersionNum(s) are specified, this router
568 supports those descriptor versions. If none are specified, it
569 defaults to version 2 descriptors.
571 [Versions of Tor before 0.1.2.5-alpha rejected router descriptors
572 with unrecognized items; the protocols line should be preceded with
573 an "opt" until these Tors are obsolete.]
575 "protocols" SP "Link" SP LINK-VERSION-LIST SP "Circuit" SP
576 CIRCUIT-VERSION-LIST NL
578 [At most once.]
580 Both lists are space-separated sequences of numbers, to indicate which
581 protocols the server supports. As of 30 Mar 2008, specified
582 protocols are "Link 1 2 Circuit 1". See section 4.1 of tor-spec.txt
583 for more information about link protocol versions.
585 [Versions of Tor before 0.1.2.5-alpha rejected router descriptors
586 with unrecognized items; the protocols line should be preceded with
587 an "opt" until these Tors are obsolete.]
589 "allow-single-hop-exits" NL
591 [At most once.]
593 Present only if the router allows single-hop circuits to make exit
594 connections. Most Tor relays do not support this: this is
595 included for specialized controllers designed to support perspective
596 access and such.
598 "or-address" SP ADDRESS ":" PORTLIST NL
600 [Any number]
602 ADDRESS = IP6ADDR | IP4ADDR
603 IPV6ADDR = an ipv6 address, surrounded by square brackets.
604 IPV4ADDR = an ipv4 address, represented as a dotted quad.
605 PORTLIST = PORTSPEC | PORTSPEC "," PORTLIST
606 PORTSPEC = PORT
607 PORT = a number between 1 and 65535 inclusive.
609 An alternative for the address and ORPort of the "router" line, but with
610 two added capabilities:
612 * or-address can be either an IPv4 or IPv6 address
613 * or-address allows for multiple ORPorts and addresses
615 A descriptor SHOULD NOT include an or-address line that does nothing but
616 duplicate the address:port pair from its "router" line.
618 The ordering of or-address lines and their PORT entries matter because
619 Tor MAY accept a limited number of addresses or ports. As of Tor 0.2.3.x
620 only the first address and the first port are used.
622 2.2. Extra-info documents
624 Extra-info documents consist of the following items:
626 "extra-info" Nickname Fingerprint NL
627 [At start, exactly once.]
629 Identifies what router this is an extra info descriptor for.
630 Fingerprint is encoded in hex (using upper-case letters), with
631 no spaces.
633 "published" YYYY-MM-DD HH:MM:SS NL
635 [Exactly once.]
637 The time, in GMT, when this document (and its corresponding router
638 descriptor if any) was generated. It MUST match the published time
639 in the corresponding router descriptor.
641 "read-history" YYYY-MM-DD HH:MM:SS (NSEC s) NUM,NUM,NUM,NUM,NUM... NL
642 [At most once.]
643 "write-history" YYYY-MM-DD HH:MM:SS (NSEC s) NUM,NUM,NUM,NUM,NUM... NL
644 [At most once.]
646 As documented in 2.1 above. See migration notes in section 2.2.1.
648 "geoip-db-digest" Digest NL
649 [At most once.]
651 SHA1 digest of the GeoIP database file that is used to resolve IP
652 addresses to country codes.
654 ("geoip-start-time" YYYY-MM-DD HH:MM:SS NL)
655 ("geoip-client-origins" CC=N,CC=N,... NL)
657 Only generated by bridge routers (see blocking.pdf), and only
658 when they have been configured with a geoip database.
659 Non-bridges SHOULD NOT generate these fields. Contains a list
660 of mappings from two-letter country codes (CC) to the number
661 of clients that have connected to that bridge from that
662 country (approximate, and rounded up to the nearest multiple of 8
663 in order to hamper traffic analysis). A country is included
664 only if it has at least one address. The time in
665 "geoip-start-time" is the time at which we began collecting geoip
666 statistics.
668 "geoip-start-time" and "geoip-client-origins" have been replaced by
669 "bridge-stats-end" and "bridge-stats-ips" in 0.2.2.4-alpha. The
670 reason is that the measurement interval with "geoip-stats" as
671 determined by subtracting "geoip-start-time" from "published" could
672 have had a variable length, whereas the measurement interval in
673 0.2.2.4-alpha and later is set to be exactly 24 hours long. In
674 order to clearly distinguish the new measurement intervals from
675 the old ones, the new keywords have been introduced.
677 "bridge-stats-end" YYYY-MM-DD HH:MM:SS (NSEC s) NL
678 [At most once.]
680 YYYY-MM-DD HH:MM:SS defines the end of the included measurement
681 interval of length NSEC seconds (86400 seconds by default).
683 A "bridge-stats-end" line, as well as any other "bridge-*" line,
684 is only added when the relay has been running as a bridge for at
685 least 24 hours.
687 "bridge-ips" CC=N,CC=N,... NL
688 [At most once.]
690 List of mappings from two-letter country codes to the number of
691 unique IP addresses that have connected from that country to the
692 bridge and which are no known relays, rounded up to the nearest
693 multiple of 8.
695 "dirreq-stats-end" YYYY-MM-DD HH:MM:SS (NSEC s) NL
696 [At most once.]
698 YYYY-MM-DD HH:MM:SS defines the end of the included measurement
699 interval of length NSEC seconds (86400 seconds by default).
701 A "dirreq-stats-end" line, as well as any other "dirreq-*" line,
702 is only added when the relay has opened its Dir port and after 24
703 hours of measuring directory requests.
705 "dirreq-v2-ips" CC=N,CC=N,... NL
706 [At most once.]
707 "dirreq-v3-ips" CC=N,CC=N,... NL
708 [At most once.]
710 List of mappings from two-letter country codes to the number of
711 unique IP addresses that have connected from that country to
712 request a v2/v3 network status, rounded up to the nearest multiple
713 of 8. Only those IP addresses are counted that the directory can
714 answer with a 200 OK status code.
716 "dirreq-v2-reqs" CC=N,CC=N,... NL
717 [At most once.]
718 "dirreq-v3-reqs" CC=N,CC=N,... NL
719 [At most once.]
721 List of mappings from two-letter country codes to the number of
722 requests for v2/v3 network statuses from that country, rounded up
723 to the nearest multiple of 8. Only those requests are counted that
724 the directory can answer with a 200 OK status code.
726 "dirreq-v2-share" num% NL
727 [At most once.]
728 "dirreq-v3-share" num% NL
729 [At most once.]
731 The share of v2/v3 network status requests that the directory
732 expects to receive from clients based on its advertised bandwidth
733 compared to the overall network bandwidth capacity. Shares are
734 formatted in percent with two decimal places. Shares are
735 calculated as means over the whole 24-hour interval.
737 "dirreq-v2-resp" status=num,... NL
738 [At most once.]
739 "dirreq-v3-resp" status=nul,... NL
740 [At most once.]
742 List of mappings from response statuses to the number of requests
743 for v2/v3 network statuses that were answered with that response
744 status, rounded up to the nearest multiple of 4. Only response
745 statuses with at least 1 response are reported. New response
746 statuses can be added at any time. The current list of response
747 statuses is as follows:
749 "ok": a network status request is answered; this number
750 corresponds to the sum of all requests as reported in
751 "dirreq-v2-reqs" or "dirreq-v3-reqs", respectively, before
752 rounding up.
753 "not-enough-sigs: a version 3 network status is not signed by a
754 sufficient number of requested authorities.
755 "unavailable": a requested network status object is unavailable.
756 "not-found": a requested network status is not found.
757 "not-modified": a network status has not been modified since the
758 If-Modified-Since time that is included in the request.
759 "busy": the directory is busy.
761 "dirreq-v2-direct-dl" key=val,... NL
762 [At most once.]
763 "dirreq-v3-direct-dl" key=val,... NL
764 [At most once.]
765 "dirreq-v2-tunneled-dl" key=val,... NL
766 [At most once.]
767 "dirreq-v3-tunneled-dl" key=val,... NL
768 [At most once.]
770 List of statistics about possible failures in the download process
771 of v2/v3 network statuses. Requests are either "direct"
772 HTTP-encoded requests over the relay's directory port, or
773 "tunneled" requests using a BEGIN_DIR cell over the relay's OR
774 port. The list of possible statistics can change, and statistics
775 can be left out from reporting. The current list of statistics is
776 as follows:
778 Successful downloads and failures:
780 "complete": a client has finished the download successfully.
781 "timeout": a download did not finish within 10 minutes after
782 starting to send the response.
783 "running": a download is still running at the end of the
784 measurement period for less than 10 minutes after starting to
785 send the response.
787 Download times:
789 "min", "max": smallest and largest measured bandwidth in B/s.
790 "d[1-4,6-9]": 1st to 4th and 6th to 9th decile of measured
791 bandwidth in B/s. For a given decile i, i/10 of all downloads
792 had a smaller bandwidth than di, and (10-i)/10 of all downloads
793 had a larger bandwidth than di.
794 "q[1,3]": 1st and 3rd quartile of measured bandwidth in B/s. One
795 fourth of all downloads had a smaller bandwidth than q1, one
796 fourth of all downloads had a larger bandwidth than q3, and the
797 remaining half of all downloads had a bandwidth between q1 and
798 q3.
799 "md": median of measured bandwidth in B/s. Half of the downloads
800 had a smaller bandwidth than md, the other half had a larger
801 bandwidth than md.
803 "dirreq-read-history" YYYY-MM-DD HH:MM:SS (NSEC s) NUM,NUM,NUM... NL
804 [At most once]
805 "dirreq-write-history" YYYY-MM-DD HH:MM:SS (NSEC s) NUM,NUM,NUM... NL
806 [At most once]
808 Declare how much bandwidth the OR has spent on answering directory
809 requests. Usage is divided into intervals of NSEC seconds. The
810 YYYY-MM-DD HH:MM:SS field defines the end of the most recent
811 interval. The numbers are the number of bytes used in the most
812 recent intervals, ordered from oldest to newest.
814 "entry-stats-end" YYYY-MM-DD HH:MM:SS (NSEC s) NL
815 [At most once.]
817 YYYY-MM-DD HH:MM:SS defines the end of the included measurement
818 interval of length NSEC seconds (86400 seconds by default).
820 An "entry-stats-end" line, as well as any other "entry-*"
821 line, is first added after the relay has been running for at least
822 24 hours.
824 "entry-ips" CC=N,CC=N,... NL
825 [At most once.]
827 List of mappings from two-letter country codes to the number of
828 unique IP addresses that have connected from that country to the
829 relay and which are no known other relays, rounded up to the
830 nearest multiple of 8.
832 "cell-stats-end" YYYY-MM-DD HH:MM:SS (NSEC s) NL
833 [At most once.]
835 YYYY-MM-DD HH:MM:SS defines the end of the included measurement
836 interval of length NSEC seconds (86400 seconds by default).
838 A "cell-stats-end" line, as well as any other "cell-*" line,
839 is first added after the relay has been running for at least 24
840 hours.
842 "cell-processed-cells" num,...,num NL
843 [At most once.]
845 Mean number of processed cells per circuit, subdivided into
846 deciles of circuits by the number of cells they have processed in
847 descending order from loudest to quietest circuits.
849 "cell-queued-cells" num,...,num NL
850 [At most once.]
852 Mean number of cells contained in queues by circuit decile. These
853 means are calculated by 1) determining the mean number of cells in
854 a single circuit between its creation and its termination and 2)
855 calculating the mean for all circuits in a given decile as
856 determined in "cell-processed-cells". Numbers have a precision of
857 two decimal places.
859 "cell-time-in-queue" num,...,num NL
860 [At most once.]
862 Mean time cells spend in circuit queues in milliseconds. Times are
863 calculated by 1) determining the mean time cells spend in the
864 queue of a single circuit and 2) calculating the mean for all
865 circuits in a given decile as determined in
866 "cell-processed-cells".
868 "cell-circuits-per-decile" num NL
869 [At most once.]
871 Mean number of circuits that are included in any of the deciles,
872 rounded up to the next integer.
874 "conn-bi-direct" YYYY-MM-DD HH:MM:SS (NSEC s) BELOW,READ,WRITE,BOTH NL
875 [At most once]
877 Number of connections, split into 10-second intervals, that are
878 used uni-directionally or bi-directionally as observed in the NSEC
879 seconds (usually 86400 seconds) before YYYY-MM-DD HH:MM:SS. Every
880 10 seconds, we determine for every connection whether we read and
881 wrote less than a threshold of 20 KiB (BELOW), read at least 10
882 times more than we wrote (READ), wrote at least 10 times more than
883 we read (WRITE), or read and wrote more than the threshold, but
884 not 10 times more in either direction (BOTH). After classifying a
885 connection, read and write counters are reset for the next
886 10-second interval.
888 "exit-stats-end" YYYY-MM-DD HH:MM:SS (NSEC s) NL
889 [At most once.]
891 YYYY-MM-DD HH:MM:SS defines the end of the included measurement
892 interval of length NSEC seconds (86400 seconds by default).
894 An "exit-stats-end" line, as well as any other "exit-*" line, is
895 first added after the relay has been running for at least 24 hours
896 and only if the relay permits exiting (where exiting to a single
897 port and IP address is sufficient).
899 "exit-kibibytes-written" port=N,port=N,... NL
900 [At most once.]
901 "exit-kibibytes-read" port=N,port=N,... NL
902 [At most once.]
904 List of mappings from ports to the number of kibibytes that the
905 relay has written to or read from exit connections to that port,
906 rounded up to the next full kibibyte. Relays may limit the
907 number of listed ports and subsume any remaining kibibytes under
908 port "other".
910 "exit-streams-opened" port=N,port=N,... NL
911 [At most once.]
913 List of mappings from ports to the number of opened exit streams
914 to that port, rounded up to the nearest multiple of 4. Relays may
915 limit the number of listed ports and subsume any remaining opened
916 streams under port "other".
918 "router-signature" NL Signature NL
919 [At end, exactly once.]
921 A document signature as documented in section 1.3, using the
922 initial item "extra-info" and the final item "router-signature",
923 signed with the router's identity key.
925 2.2.1. Moving history fields to extra-info documents.
927 Tools that want to use the read-history and write-history values SHOULD
928 download extra-info documents as well as router descriptors. Such
929 tools SHOULD accept history values from both sources; if they appear in
930 both documents, the values in the extra-info documents are authoritative.
932 New versions of Tor no longer generate router descriptors
933 containing read-history or write-history. Tools should continue to
934 accept read-history and write-history values in router descriptors
935 produced by older versions of Tor until all Tor versions earlier
936 than 0.2.0.x are obsolete.
938 2.3. Nonterminals in router descriptors
940 nickname ::= between 1 and 19 alphanumeric characters ([A-Za-z0-9]),
941 case-insensitive.
942 hexdigest ::= a '$', followed by 40 hexadecimal characters
943 ([A-Fa-f0-9]). [Represents a relay by the digest of its identity
944 key.]
946 exitpattern ::= addrspec ":" portspec
947 portspec ::= "*" | port | port "-" port
948 port ::= an integer between 1 and 65535, inclusive.
950 [Some implementations incorrectly generate ports with value 0.
951 Implementations SHOULD accept this, and SHOULD NOT generate it.
952 Connections to port 0 are never permitted.]
954 addrspec ::= "*" | ip4spec | ip6spec
955 ipv4spec ::= ip4 | ip4 "/" num_ip4_bits | ip4 "/" ip4mask
956 ip4 ::= an IPv4 address in dotted-quad format
957 ip4mask ::= an IPv4 mask in dotted-quad format
958 num_ip4_bits ::= an integer between 0 and 32
959 ip6spec ::= ip6 | ip6 "/" num_ip6_bits
960 ip6 ::= an IPv6 address, surrounded by square brackets.
961 num_ip6_bits ::= an integer between 0 and 128
963 bool ::= "0" | "1"
965 3. Formats produced by directory authorities.
967 Every authority has two keys used in this protocol: a signing key, and
968 an authority identity key. (Authorities also have a router identity
969 key used in their role as a router and by earlier versions of the
970 directory protocol.) The identity key is used from time to time to
971 sign new key certificates using new signing keys; it is very sensitive.
972 The signing key is used to sign key certificates and status documents.
974 There are three kinds of documents generated by directory authorities:
976 Key certificates
977 Status votes
978 Status consensuses
980 Each is discussed below.
982 3.1. Key certificates
984 Key certificates consist of the following items:
986 "dir-key-certificate-version" version NL
988 [At start, exactly once.]
990 Determines the version of the key certificate. MUST be "3" for
991 the protocol described in this document. Implementations MUST
992 reject formats they don't understand.
994 "dir-address" IPPort NL
995 [At most once]
997 An IP:Port for this authority's directory port.
999 "fingerprint" fingerprint NL
1001 [Exactly once.]
1003 Hexadecimal encoding without spaces based on the authority's
1004 identity key.
1006 "dir-identity-key" NL a public key in PEM format
1008 [Exactly once.]
1010 The long-term authority identity key for this authority. This key
1011 SHOULD be at least 2048 bits long; it MUST NOT be shorter than
1012 1024 bits.
1014 "dir-key-published" YYYY-MM-DD HH:MM:SS NL
1016 [Exactly once.]
1018 The time (in GMT) when this document and corresponding key were
1019 last generated.
1021 "dir-key-expires" YYYY-MM-DD HH:MM:SS NL
1023 [Exactly once.]
1025 A time (in GMT) after which this key is no longer valid.
1027 "dir-signing-key" NL a key in PEM format
1029 [Exactly once.]
1031 The directory server's public signing key. This key MUST be at
1032 least 1024 bits, and MAY be longer.
1034 "dir-key-crosscert" NL CrossSignature NL
1036 [At most once.]
1038 NOTE: Authorities MUST include this field in all newly generated
1039 certificates. A future version of this specification will make
1040 the field required.
1042 CrossSignature is a signature, made using the certificate's signing
1043 key, of the digest of the PKCS1-padded hash of the certificate's
1044 identity key. For backward compatibility with broken versions of the
1045 parser, we wrap the base64-encoded signature in -----BEGIN ID
1046 SIGNATURE---- and -----END ID SIGNATURE----- tags. Implementations
1047 MUST allow the "ID " portion to be omitted, however.
1049 When encountering a certificate with a dir-key-crosscert entry,
1050 implementations MUST verify that the signature is a correct signature
1051 of the hash of the identity key using the signing key.
1053 "dir-key-certification" NL Signature NL
1055 [At end, exactly once.]
1057 A document signature as documented in section 1.3, using the
1058 initial item "dir-key-certificate-version" and the final item
1059 "dir-key-certification", signed with the authority identity key.
1061 Authorities MUST generate a new signing key and corresponding
1062 certificate before the key expires.
1064 3.2. Microdescriptors
1066 Microdescriptors are a stripped-down version of router descriptors
1067 generated by the directory authorities which may additionally contain
1068 authority-generated information. Microdescriptors contain only the
1069 most relevant parts that clients care about. Microdescriptors are
1070 expected to be relatively static and only change about once per week.
1071 Microdescriptors do not contain any information that clients need to
1072 use to decide which servers to fetch information about, or which
1073 servers to fetch information from.
1075 Microdescriptors are a straight transform from the router descriptor
1076 and the consensus method. Microdescriptors have no header or footer.
1077 Microdescriptors are identified by the hash of its concatenated
1078 elements without a signature by the router. Microdescriptors do not
1079 contain any version information, because their version is determined
1080 by the consensus method.
1082 3.2.1. Microdescriptors in consensus method 8 or later
1084 Starting with consensus method 8, microdescriptors contain the
1085 following elements taken from or based on the router descriptor. Order
1086 matters here, because different directory authorities must be able to
1087 transform a given router descriptor and consensus method into the exact
1088 same microdescriptor.
1090 "onion-key" NL a public key in PEM format
1092 [Exactly once, at start]
1094 The "onion-key" element as specified in 2.1.
1096 "family" names NL
1098 [At most once]
1100 The "family" element as specified in 2.1.
1102 "p" SP ("accept" / "reject") SP PortList NL
1104 [At most once]
1106 The exit-policy summary as specified in 3.3 and 3.5.2. A missing
1107 "p" line is equivalent to "p reject 1-65535".
1109 [With microdescriptors, clients don't learn exact exit policies:
1110 clients can only guess whether a relay accepts their request, try the
1111 BEGIN request, and might get end-reason-exit-policy if they guessed
1112 wrong, in which case they'll have to try elsewhere.]
1114 (Note that with microdescriptors, clients do not learn the identity of
1115 their routers: they only learn a hash of the identity key. This is all
1116 they need to confirm the actual identity key when doing a TLS handshake,
1117 and all they need to put the identity key digest in their cREATE cells.)
1119 3.3. Vote and consensus status documents
1121 Votes and consensuses are more strictly formatted then other documents
1122 in this specification, since different authorities must be able to
1123 generate exactly the same consensus given the same set of votes.
1125 The procedure for deciding when to generate vote and consensus status
1126 documents are described in section 1.4 on the voting timeline.
1128 Status documents contain a preamble, an authority section, a list of
1129 router status entries, and one or more footer signature, in that order.
1131 Unlike other formats described above, a SP in these documents must be a
1132 single space character (hex 20).
1134 Some items appear only in votes, and some items appear only in
1135 consensuses. Unless specified, items occur in both.
1137 The preamble contains the following items. They MUST occur in the
1138 order given here:
1140 "network-status-version" SP version NL.
1142 [At start, exactly once.]
1144 A document format version. For this specification, the version is
1145 "3".
1147 "vote-status" SP type NL
1149 [Exactly once.]
1151 The status MUST be "vote" or "consensus", depending on the type of
1152 the document.
1154 "consensus-methods" SP IntegerList NL
1156 [At most once for votes; does not occur in consensuses.]
1158 A space-separated list of supported methods for generating
1159 consensuses from votes. See section 3.5.1 for details. Method "1"
1160 MUST be included.
1162 "consensus-method" SP Integer NL
1164 [At most once for consensuses; does not occur in votes.]
1166 See section 3.5.1 for details.
1168 (Only included when the vote is generated with consensus-method 2 or
1169 later.)
1171 "published" SP YYYY-MM-DD SP HH:MM:SS NL
1173 [Exactly once for votes; does not occur in consensuses.]
1175 The publication time for this status document (if a vote).
1177 "valid-after" SP YYYY-MM-DD SP HH:MM:SS NL
1179 [Exactly once.]
1181 The start of the Interval for this vote. Before this time, the
1182 consensus document produced from this vote should not be used.
1183 See 1.4 for voting timeline information.
1185 "fresh-until" SP YYYY-MM-DD SP HH:MM:SS NL
1187 [Exactly once.]
1189 The time at which the next consensus should be produced; before this
1190 time, there is no point in downloading another consensus, since there
1191 won't be a new one. See 1.4 for voting timeline information.
1193 "valid-until" SP YYYY-MM-DD SP HH:MM:SS NL
1195 [Exactly once.]
1197 The end of the Interval for this vote. After this time, the
1198 consensus produced by this vote should not be used. See 1.4 for
1199 voting timeline information.
1201 "voting-delay" SP VoteSeconds SP DistSeconds NL
1203 [Exactly once.]
1205 VoteSeconds is the number of seconds that we will allow to collect
1206 votes from all authorities; DistSeconds is the number of seconds
1207 we'll allow to collect signatures from all authorities. See 1.4 for
1208 voting timeline information.
1210 "client-versions" SP VersionList NL
1212 [At most once.]
1214 A comma-separated list of recommended Tor versions for client
1215 usage, in ascending order. The versions are given as defined by
1216 version-spec.txt. If absent, no opinion is held about client
1217 versions.
1219 "server-versions" SP VersionList NL
1221 [At most once.]
1223 A comma-separated list of recommended Tor versions for relay
1224 usage, in ascending order. The versions are given as defined by
1225 version-spec.txt. If absent, no opinion is held about server
1226 versions.
1228 "known-flags" SP FlagList NL
1230 [Exactly once.]
1232 A space-separated list of all of the flags that this document
1233 might contain. A flag is "known" either because the authority
1234 knows about them and might set them (if in a vote), or because
1235 enough votes were counted for the consensus for an authoritative
1236 opinion to have been formed about their status.
1238 "params" SP [Parameters] NL
1240 [At most once]
1242 Parameter ::= Keyword '=' Int32
1243 Int32 ::= A decimal integer between -2147483648 and 2147483647.
1244 Parameters ::= Parameter | Parameters SP Parameter
1246 The parameters list, if present, contains a space-separated list of
1247 case-sensitive key-value pairs, sorted in lexical order by their
1248 keyword (as ASCII byte strings). Each parameter has its own meaning.
1250 (Only included when the vote is generated with consensus-method 7 or
1251 later.)
1253 Commonly used "param" arguments at this point include:
1255 "circwindow" -- the default package window that circuits should
1256 be established with. It started out at 1000 cells, but some
1257 research indicates that a lower value would mean fewer cells in
1258 transit in the network at any given time. Obeyed by Tor 0.2.1.20
1259 and later.
1260 Min: 100, Max: 1000
1262 "CircuitPriorityHalflifeMsec" -- the halflife parameter used when
1263 weighting which circuit will send the next cell. Obeyed by Tor
1264 0.2.2.10-alpha and later. (Versions of Tor between 0.2.2.7-alpha
1265 and 0.2.2.10-alpha recognized a "CircPriorityHalflifeMsec" parameter,
1266 but mishandled it badly.)
1267 Min: -1, Max: 2147483647 (INT32_MAX)
1269 "perconnbwrate" and "perconnbwburst" -- if set, each relay sets
1270 up a separate token bucket for every client OR connection,
1271 and rate limits that connection indepedently. Typically left
1272 unset, except when used for performance experiments around trac
1273 entry 1750. Only honored by relays running Tor 0.2.2.16-alpha
1274 and later. (Note that relays running 0.2.2.7-alpha through
1275 0.2.2.14-alpha looked for bwconnrate and bwconnburst, but then
1276 did the wrong thing with them; see bug 1830 for details.)
1277 Min: 1, Max: 2147483647 (INT32_MAX)
1279 "refuseunknownexits" -- if set to one, exit relays look at
1280 the previous hop of circuits that ask to open an exit stream,
1281 and refuse to exit if they don't recognize it as a relay. The
1282 goal is to make it harder for people to use them as one-hop
1283 proxies. See trac entry 1751 for details.
1284 Min: 0, Max: 1
1286 "cbtdisabled", "cbtnummodes", "cbtrecentcount", "cbtmaxtimeouts",
1287 "cbtmincircs", "cbtquantile", "cbtclosequantile", "cbttestfreq",
1288 "cbtmintimeout", and "cbtinitialtimeout" -- see "2.4.5. Consensus
1289 parameters governing behavior" in path-spec.txt for a series of
1290 circuit build time related consensus params.
1292 "UseOptimisticData" -- If set to zero, clients by default
1293 shouldn't try to send optimistic data to servers until they have
1294 received a RELAY_CONNECTED cell.
1295 Min: 0, Max: 1
1297 The authority section of a vote contains the following items, followed
1298 in turn by the authority's current key certificate:
1300 "dir-source" SP nickname SP identity SP address SP IP SP dirport SP
1301 orport NL
1303 [Exactly once, at start]
1305 Describes this authority. The nickname is a convenient identifier
1306 for the authority. The identity is an uppercase hex fingerprint of
1307 the authority's current (v3 authority) identity key. The address is
1308 the server's hostname. The IP is the server's current IP address,
1309 and dirport is its current directory port. XXXXorport
1311 "contact" SP string NL
1313 [Exactly once.]
1315 An arbitrary string describing how to contact the directory
1316 server's administrator. Administrators should include at least an
1317 email address and a PGP fingerprint.
1319 "legacy-dir-key" SP FINGERPRINT NL
1321 [At most once]
1323 Lists a fingerprint for an obsolete _identity_ key still used
1324 by this authority to keep older clients working. This option
1325 is used to keep key around for a little while in case the
1326 authorities need to migrate many identity keys at once.
1327 (Generally, this would only happen because of a security
1328 vulnerability that affected multiple authorities, like the
1329 Debian OpenSSL RNG bug of May 2008.)
1331 The authority section of a consensus contains groups the following items,
1332 in the order given, with one group for each authority that contributed to
1333 the consensus, with groups sorted by authority identity digest:
1335 "dir-source" SP nickname SP identity SP address SP IP SP dirport SP
1336 orport NL
1338 [Exactly once, at start]
1340 As in the authority section of a vote.
1342 "contact" SP string NL
1344 [Exactly once.]
1346 As in the authority section of a vote.
1348 "vote-digest" SP digest NL
1350 [Exactly once.]
1352 A digest of the vote from the authority that contributed to this
1353 consensus, as signed (that is, not including the signature).
1354 (Hex, upper-case.)
1356 Each router status entry contains the following items. Router status
1357 entries are sorted in ascending order by identity digest.
1359 "r" SP nickname SP identity SP digest SP publication SP IP SP ORPort
1360 SP DirPort NL
1362 [At start, exactly once.]
1364 "Nickname" is the OR's nickname. "Identity" is a hash of its
1365 identity key, encoded in base64, with trailing equals sign(s)
1366 removed. "Digest" is a hash of its most recent descriptor as
1367 signed (that is, not including the signature), encoded in base64.
1368 "Publication" is the
1369 publication time of its most recent descriptor, in the form
1370 YYYY-MM-DD HH:MM:SS, in GMT. "IP" is its current IP address;
1371 ORPort is its current OR port, "DirPort" is its current directory
1372 port, or "0" for "none".
1374 "s" SP Flags NL
1376 [Exactly once.]
1378 A series of space-separated status flags, in lexical order (as ASCII
1379 byte strings). Currently documented flags are:
1381 "Authority" if the router is a directory authority.
1382 "BadExit" if the router is believed to be useless as an exit node
1383 (because its ISP censors it, because it is behind a restrictive
1384 proxy, or for some similar reason).
1385 "BadDirectory" if the router is believed to be useless as a
1386 directory cache (because its directory port isn't working,
1387 its bandwidth is always throttled, or for some similar
1388 reason).
1389 "Exit" if the router is more useful for building
1390 general-purpose exit circuits than for relay circuits. The
1391 path building algorithm uses this flag; see path-spec.txt.
1392 "Fast" if the router is suitable for high-bandwidth circuits.
1393 "Guard" if the router is suitable for use as an entry guard.
1394 "HSDir" if the router is considered a v2 hidden service directory.
1395 "Named" if the router's identity-nickname mapping is canonical,
1396 and this authority binds names.
1397 "Stable" if the router is suitable for long-lived circuits.
1398 "Running" if the router is currently usable.
1399 "Unnamed" if another router has bound the name used by this
1400 router, and this authority binds names.
1401 "Valid" if the router has been 'validated'.
1402 "V2Dir" if the router implements the v2 directory protocol.
1403 "V3Dir" if the router implements this protocol.
1405 "v" SP version NL
1407 [At most once.]
1409 The version of the Tor protocol that this relay is running. If
1410 the value begins with "Tor" SP, the rest of the string is a Tor
1411 version number, and the protocol is "The Tor protocol as supported
1412 by the given version of Tor." Otherwise, if the value begins with
1413 some other string, Tor has upgraded to a more sophisticated
1414 protocol versioning system, and the protocol is "a version of the
1415 Tor protocol more recent than any we recognize."
1417 Directory authorities SHOULD omit version strings they receive from
1418 descriptors if they would cause "v" lines to be over 128 characters
1419 long.
1421 "w" SP "Bandwidth=" INT [SP "Measured=" INT] NL
1423 [At most once.]
1425 An estimate of the bandwidth of this relay, in an arbitrary
1426 unit (currently kilobytes per second). Used to weight router
1427 selection.
1429 Additionally, the Measured= keyword is present in votes by
1430 participating bandwidth measurement authorities to indicate
1431 a measured bandwidth currently produced by measuring stream
1432 capacities.
1434 Other weighting keywords may be added later.
1435 Clients MUST ignore keywords they do not recognize.
1437 "p" SP ("accept" / "reject") SP PortList NL
1439 [At most once.]
1441 PortList = PortOrRange
1442 PortList = PortList "," PortOrRange
1443 PortOrRange = INT "-" INT / INT
1445 A list of those ports that this router supports (if 'accept')
1446 or does not support (if 'reject') for exit to "most
1447 addresses".
1449 "m" SP methods 1*(SP algorithm "=" digest) NL
1451 [Any number, only in votes.]
1453 Microdescriptor hashes for all consensus methods that an authority
1454 supports and that use the same microdescriptor format. "methods"
1455 is a comma-separated list of the consensus methods that the
1456 authority believes will produce "digest". "algorithm" is the name
1457 of the hash algorithm producing "digest", which can be "sha256" or
1458 something else, depending on the consensus "methods" supporting
1459 this algorithm. "digest" is the base64 encoding of the hash of
1460 the router's microdescriptor with trailing =s omitted.
1462 The footer section is delineated in all votes and consensuses supporting
1463 consensus method 9 and above with the following:
1465 "directory-footer" NL
1467 It contains two subsections, a bandwidths-weights line and a
1468 directory-signature.
1470 The bandwidths-weights line appears At Most Once for a consensus. It does
1471 not appear in votes.
1473 "bandwidth-weights" SP
1474 "Wbd=" INT SP "Wbe=" INT SP "Wbg=" INT SP "Wbm=" INT SP
1475 "Wdb=" INT SP
1476 "Web=" INT SP "Wed=" INT SP "Wee=" INT SP "Weg=" INT SP "Wem=" INT SP
1477 "Wgb=" INT SP "Wgd=" INT SP "Wgg=" INT SP "Wgm=" INT SP
1478 "Wmb=" INT SP "Wmd=" INT SP "Wme=" INT SP "Wmg=" INT SP "Wmm=" INT NL
1480 These values represent the weights to apply to router bandwidths
1481 during path selection. They are sorted in lexical order (as ASCII byte
1482 strings). The integer values are divided by BW_WEIGHT_SCALE=10000 or
1483 the consensus param "bwweightscale". They are:
1485 Wgg - Weight for Guard-flagged nodes in the guard position
1486 Wgm - Weight for non-flagged nodes in the guard Position
1487 Wgd - Weight for Guard+Exit-flagged nodes in the guard Position
1489 Wmg - Weight for Guard-flagged nodes in the middle Position
1490 Wmm - Weight for non-flagged nodes in the middle Position
1491 Wme - Weight for Exit-flagged nodes in the middle Position
1492 Wmd - Weight for Guard+Exit flagged nodes in the middle Position
1494 Weg - Weight for Guard flagged nodes in the exit Position
1495 Wem - Weight for non-flagged nodes in the exit Position
1496 Wee - Weight for Exit-flagged nodes in the exit Position
1497 Wed - Weight for Guard+Exit-flagged nodes in the exit Position
1499 Wgb - Weight for BEGIN_DIR-supporting Guard-flagged nodes
1500 Wmb - Weight for BEGIN_DIR-supporting non-flagged nodes
1501 Web - Weight for BEGIN_DIR-supporting Exit-flagged nodes
1502 Wdb - Weight for BEGIN_DIR-supporting Guard+Exit-flagged nodes
1504 Wbg - Weight for Guard flagged nodes for BEGIN_DIR requests
1505 Wbm - Weight for non-flagged nodes for BEGIN_DIR requests
1506 Wbe - Weight for Exit-flagged nodes for BEGIN_DIR requests
1507 Wbd - Weight for Guard+Exit-flagged nodes for BEGIN_DIR requests
1509 These values are calculated as specified in Section 3.5.3.
1511 The signature contains the following item, which appears Exactly Once
1512 for a vote, and At Least Once for a consensus.
1514 "directory-signature" SP identity SP signing-key-digest NL Signature
1516 This is a signature of the status document, with the initial item
1517 "network-status-version", and the signature item
1518 "directory-signature", using the signing key. (In this case, we take
1519 the hash through the _space_ after directory-signature, not the
1520 newline: this ensures that all authorities sign the same thing.)
1521 "identity" is the hex-encoded digest of the authority identity key of
1522 the signing authority, and "signing-key-digest" is the hex-encoded
1523 digest of the current authority signing key of the signing authority.
1525 3.4. Assigning flags in a vote
1527 (This section describes how directory authorities choose which status
1528 flags to apply to routers, as of Tor 0.2.0.0-alpha-dev. Later directory
1529 authorities MAY do things differently, so long as clients keep working
1530 well. Clients MUST NOT depend on the exact behaviors in this section.)
1532 In the below definitions, a router is considered "active" if it is
1533 running, valid, and not hibernating.
1535 "Valid" -- a router is 'Valid' if it is running a version of Tor not
1536 known to be broken, and the directory authority has not blacklisted
1537 it as suspicious.
1539 "Named" -- Directory authority administrators may decide to support name
1540 binding. If they do, then they must maintain a file of
1541 nickname-to-identity-key mappings, and try to keep this file consistent
1542 with other directory authorities. If they don't, they act as clients, and
1543 report bindings made by other directory authorities (name X is bound to
1544 identity Y if at least one binding directory lists it, and no directory
1545 binds X to some other Y'.) A router is called 'Named' if the router
1546 believes the given name should be bound to the given key.
1548 Two strategies exist on the current network for deciding on
1549 values for the Named flag. In the original version, relay
1550 operators were asked to send nickname-identity pairs to a
1551 mailing list of Naming directory authorities' operators. The
1552 operators were then supposed to add the pairs to their
1553 mapping files; in practice, they didn't get to this often.
1555 Newer Naming authorities run a script that registers routers
1556 in their mapping files once the routers have been online at
1557 least two weeks, no other router has that nickname, and no
1558 other router has wanted the nickname for a month. If a router
1559 has not been online for six months, the router is removed.
1561 "Unnamed" -- Directory authorities that support naming should vote for a
1562 router to be 'Unnamed' if its given nickname is mapped to a different
1563 identity.
1565 "Running" -- A router is 'Running' if the authority managed to connect to
1566 it successfully within the last 45 minutes.
1568 "Stable" -- A router is 'Stable' if it is active, and either its Weighted
1569 MTBF is at least the median for known active routers or its Weighted MTBF
1570 corresponds to at least 7 days. Routers are never called Stable if they are
1571 running a version of Tor known to drop circuits stupidly. (0.1.1.10-alpha
1572 through 0.1.1.16-rc are stupid this way.)
1574 To calculate weighted MTBF, compute the weighted mean of the lengths
1575 of all intervals when the router was observed to be up, weighting
1576 intervals by $\alpha^n$, where $n$ is the amount of time that has
1577 passed since the interval ended, and $\alpha$ is chosen so that
1578 measurements over approximately one month old no longer influence the
1579 weighted MTBF much.
1581 [XXXX what happens when we have less than 4 days of MTBF info.]
1583 "Exit" -- A router is called an 'Exit' iff it allows exits to at
1584 least two of the ports 80, 443, and 6667 and allows exits to at
1585 least one /8 address space.
1587 "Fast" -- A router is 'Fast' if it is active, and its bandwidth is
1588 either in the top 7/8ths for known active routers or at least some
1589 minimum (20KB/s until 0.2.3.7-alpha, and 100KB/s after that).
1591 "Guard" -- A router is a possible 'Guard' if its Weighted Fractional
1592 Uptime is at least the median for "familiar" active routers, and if
1593 its bandwidth is at least median or at least 250KB/s.
1595 To calculate weighted fractional uptime, compute the fraction
1596 of time that the router is up in any given day, weighting so that
1597 downtime and uptime in the past counts less.
1599 A node is 'familiar' if 1/8 of all active nodes have appeared more
1600 recently than it, OR it has been around for a few weeks.
1602 "Authority" -- A router is called an 'Authority' if the authority
1603 generating the network-status document believes it is an authority.
1605 "V2Dir" -- A router supports the v2 directory protocol if it has an open
1606 directory port, and it is running a version of the directory protocol that
1607 supports the functionality clients need. (Currently, this is
1608 0.1.1.9-alpha or later.)
1610 "V3Dir" -- A router supports the v3 directory protocol if it has an open
1611 directory port, and it is running a version of the directory protocol that
1612 supports the functionality clients need. (Currently, this is
1613 0.2.0.?????-alpha or later.)
1615 "HSDir" -- A router is a v2 hidden service directory if it stores and
1616 serves v2 hidden service descriptors and the authority managed to connect
1617 to it successfully within the last 24 hours.
1619 Directory server administrators may label some relays or IPs as
1620 blacklisted, and elect not to include them in their network-status lists.
1622 Authorities SHOULD 'disable' any relays in excess of 3 on any single IP.
1623 When there are more than 3 to choose from, authorities should first prefer
1624 authorities to non-authorities, then prefer Running to non-Running, and
1625 then prefer high-bandwidth to low-bandwidth. To 'disable' a relay, the
1626 authority *should* advertise it without the Running or Valid flag.
1628 Thus, the network-status vote includes all non-blacklisted,
1629 non-expired, non-superseded descriptors.
1631 The bandwidth in a "w" line should be taken as the best estimate
1632 of the router's actual capacity that the authority has. For now,
1633 this should be the lesser of the observed bandwidth and bandwidth
1634 rate limit from the router descriptor. It is given in kilobytes
1635 per second, and capped at some arbitrary value (currently 10 MB/s).
1637 The Measured= keyword on a "w" line vote is currently computed
1638 by multiplying the previous published consensus bandwidth by the
1639 ratio of the measured average node stream capacity to the network
1640 average. If 3 or more authorities provide a Measured= keyword for
1641 a router, the authorities produce a consensus containing a "w"
1642 Bandwidth= keyword equal to the median of the Measured= votes.
1644 The ports listed in a "p" line should be taken as those ports for
1645 which the router's exit policy permits 'most' addresses, ignoring any
1646 accept not for all addresses, ignoring all rejects for private
1647 netblocks. "Most" addresses are permitted if no more than 2^25
1648 IPv4 addresses (two /8 networks) were blocked. The list is encoded
1649 as described in 3.5.2.
1651 3.5. Computing a consensus from a set of votes
1653 Given a set of votes, authorities compute the contents of the consensus
1654 document as follows:
1656 The "valid-after", "valid-until", and "fresh-until" times are taken as
1657 the median of the respective values from all the votes.
1659 The times in the "voting-delay" line are taken as the median of the
1660 VoteSeconds and DistSeconds times in the votes.
1662 Known-flags is the union of all flags known by any voter.
1664 Entries are given on the "params" line for every keyword on which a
1665 majority of authorities (total authorities, not just those
1666 participating in this vote) voted on, or if at least three
1667 authorities voted for that parameter. The values given are the
1668 low-median of all votes on that keyword.
1670 Consensus methods 11 and before, entries are given on the "params"
1671 line for every keyword on which any authority voted, the value given
1672 being the low-median of all votes on that keyword.
1674 "client-versions" and "server-versions" are sorted in ascending
1675 order; A version is recommended in the consensus if it is recommended
1676 by more than half of the voting authorities that included a
1677 client-versions or server-versions lines in their votes.
1679 The authority item groups (dir-source, contact, fingerprint,
1680 vote-digest) are taken from the votes of the voting
1681 authorities. These groups are sorted by the digests of the
1682 authorities identity keys, in ascending order. If the consensus
1683 method is 3 or later, a dir-source line must be included for
1684 every vote with legacy-key entry, using the legacy-key's
1685 fingerprint, the voter's ordinary nickname with the string
1686 "-legacy" appended, and all other fields as from the original
1687 vote's dir-source line.
1689 A router status entry:
1690 * is included in the result if some router status entry with the same
1691 identity is included by more than half of the authorities (total
1692 authorities, not just those whose votes we have).
1694 * For any given identity, we include at most one router status entry.
1696 * A router entry has a flag set if that is included by more than half
1697 of the authorities who care about that flag.
1699 * Two router entries are "the same" if they have the same
1700 <descriptor digest, published time, nickname, IP, ports> tuple.
1701 We choose the tuple for a given router as whichever tuple appears
1702 for that router in the most votes. We break ties first in favor of
1703 the more recently published, then in favor of smaller server
1704 descriptor digest.
1706 * The Named flag appears if it is included for this routerstatus by
1707 _any_ authority, and if all authorities that list it list the same
1708 nickname. However, if consensus-method 2 or later is in use, and
1709 any authority calls this identity/nickname pair Unnamed, then
1710 this routerstatus does not get the Named flag.
1712 * If consensus-method 2 or later is in use, the Unnamed flag is
1713 set for a routerstatus if any authorities have voted for a different
1714 identities to be Named with that nickname, or if any authority
1715 lists that nickname/ID pair as Unnamed.
1717 (With consensus-method 1, Unnamed is set like any other flag.)
1719 * The version is given as whichever version is listed by the most
1720 voters, with ties decided in favor of more recent versions.
1722 * If consensus-method 4 or later is in use, then routers that
1723 do not have the Running flag are not listed at all.
1725 * If consensus-method 5 or later is in use, then the "w" line
1726 is generated using a low-median of the bandwidth values from
1727 the votes that included "w" lines for this router.
1729 * If consensus-method 5 or later is in use, then the "p" line
1730 is taken from the votes that have the same policy summary
1731 for the descriptor we are listing. (They should all be the
1732 same. If they are not, we pick the most commonly listed
1733 one, breaking ties in favor of the lexicographically larger
1734 vote.) The port list is encoded as specified in 3.5.2.
1736 * If consensus-method 6 or later is in use and if 3 or more
1737 authorities provide a Measured= keyword in their votes for
1738 a router, the authorities produce a consensus containing a
1739 Bandwidth= keyword equal to the median of the Measured= votes.
1741 * If consensus-method 7 or later is in use, the params line is
1742 included in the output.
1744 * If the consensus method is under 11, bad exits are considered as
1745 possible exits when computing bandwidth weights. Otherwise, if
1746 method 11 or later is in use, any router that is determined to get
1747 the BadExit flag doesn't count when we're calculating weights.
1749 * If consensus method 12 or later is used, only consensus
1750 parameters that more than half of the total number of
1751 authorities voted for are included in the consensus.
1753 The signatures at the end of a consensus document are sorted in
1754 ascending order by identity digest.
1756 All ties in computing medians are broken in favor of the smaller or
1757 earlier item.
1759 3.5.1. Forward compatibility
1761 Future versions of Tor will need to include new information in the
1762 consensus documents, but it is important that all authorities (or at least
1763 half) generate and sign the same signed consensus.
1765 To achieve this, authorities list in their votes their supported methods
1766 for generating consensuses from votes. Later methods will be assigned
1767 higher numbers. Currently recognized methods:
1768 "1" -- The first implemented version.
1769 "2" -- Added support for the Unnamed flag.
1770 "3" -- Added legacy ID key support to aid in authority ID key rollovers
1771 "4" -- No longer list routers that are not running in the consensus
1772 "5" -- adds support for "w" and "p" lines.
1773 "6" -- Prefers measured bandwidth values rather than advertised
1774 "7" -- Provides keyword=integer pairs of consensus parameters
1775 "8" -- Provides microdescriptor summaries
1776 "9" -- Provides weights for selecting flagged routers in paths
1777 "10" -- Fixes edge case bugs in router flag selection weights
1778 "11" -- Don't consider BadExits when calculating bandwidth weights
1779 "12" -- Params are only included if enough auths voted for them
1782 Before generating a consensus, an authority must decide which consensus
1783 method to use. To do this, it looks for the highest version number
1784 supported by more than 2/3 of the authorities voting. If it supports this
1785 method, then it uses it. Otherwise, it falls back to method 1.
1787 (The consensuses generated by new methods must be parsable by
1788 implementations that only understand the old methods, and must not cause
1789 those implementations to compromise their anonymity. This is a means for
1790 making changes in the contents of consensus; not for making
1791 backward-incompatible changes in their format.)
1793 3.5.2. Encoding port lists
1795 Whether the summary shows the list of accepted ports or the list of
1796 rejected ports depends on which list is shorter (has a shorter string
1797 representation). In case of ties we choose the list of accepted
1798 ports. As an exception to this rule an allow-all policy is
1799 represented as "accept 1-65535" instead of "reject " and a reject-all
1800 policy is similarly given as "reject 1-65535".
1802 Summary items are compressed, that is instead of "80-88,89-100" there
1803 only is a single item of "80-100", similarly instead of "20,21" a
1804 summary will say "20-21".
1806 Port lists are sorted in ascending order.
1808 The maximum allowed length of a policy summary (including the "accept "
1809 or "reject ") is 1000 characters. If a summary exceeds that length we
1810 use an accept-style summary and list as much of the port list as is
1811 possible within these 1000 bytes. [XXXX be more specific.]
1813 3.5.3. Computing Bandwidth Weights
1815 Let weight_scale = 10000
1817 Let G be the total bandwidth for Guard-flagged nodes.
1818 Let M be the total bandwidth for non-flagged nodes.
1819 Let E be the total bandwidth for Exit-flagged nodes.
1820 Let D be the total bandwidth for Guard+Exit-flagged nodes.
1821 Let T = G+M+E+D
1823 Let Wgd be the weight for choosing a Guard+Exit for the guard position.
1824 Let Wmd be the weight for choosing a Guard+Exit for the middle position.
1825 Let Wed be the weight for choosing a Guard+Exit for the exit position.
1827 Let Wme be the weight for choosing an Exit for the middle position.
1828 Let Wmg be the weight for choosing a Guard for the middle position.
1830 Let Wgg be the weight for choosing a Guard for the guard position.
1831 Let Wee be the weight for choosing an Exit for the exit position.
1833 Balanced network conditions then arise from solutions to the following
1834 system of equations:
1836 Wgg*G + Wgd*D == M + Wmd*D + Wme*E + Wmg*G (guard bw = middle bw)
1837 Wgg*G + Wgd*D == Wee*E + Wed*D (guard bw = exit bw)
1838 Wed*D + Wmd*D + Wgd*D == D (aka: Wed+Wmd+Wdg = 1)
1839 Wmg*G + Wgg*G == G (aka: Wgg = 1-Wmg)
1840 Wme*E + Wee*E == E (aka: Wee = 1-Wme)
1842 We are short 2 constraints with the above set. The remaining constraints
1843 come from examining different cases of network load. The following
1844 constraints are used in consensus method 10 and above. There are another
1845 incorrect and obsolete set of constraints used for these same cases in
1846 consensus method 9. For those, see dir-spec.txt in Tor 0.2.2.10-alpha
1847 to 0.2.2.16-alpha.
1849 Case 1: E >= T/3 && G >= T/3 (Neither Exit nor Guard Scarce)
1851 In this case, the additional two constraints are: Wmg == Wmd,
1852 Wed == 1/3.
1854 This leads to the solution:
1855 Wgd = weight_scale/3
1856 Wed = weight_scale/3
1857 Wmd = weight_scale/3
1858 Wee = (weight_scale*(E+G+M))/(3*E)
1859 Wme = weight_scale - Wee
1860 Wmg = (weight_scale*(2*G-E-M))/(3*G)
1861 Wgg = weight_scale - Wmg
1863 Case 2: E < T/3 && G < T/3 (Both are scarce)
1865 Let R denote the more scarce class (Rare) between Guard vs Exit.
1866 Let S denote the less scarce class.
1868 Subcase a: R+D < S
1870 In this subcase, we simply devote all of D bandwidth to the
1871 scarce class.
1873 Wgg = Wee = weight_scale
1874 Wmg = Wme = Wmd = 0;
1875 if E < G:
1876 Wed = weight_scale
1877 Wgd = 0
1878 else:
1879 Wed = 0
1880 Wgd = weight_scale
1882 Subcase b: R+D >= S
1884 In this case, if M <= T/3, we have enough bandwidth to try to achieve
1885 a balancing condition.
1887 Add constraints Wgg = 1, Wmd == Wgd to maximize bandwidth in the guard
1888 position while still allowing exits to be used as middle nodes:
1890 Wee = (weight_scale*(E - G + M))/E
1891 Wed = (weight_scale*(D - 2*E + 4*G - 2*M))/(3*D)
1892 Wme = (weight_scale*(G-M))/E
1893 Wmg = 0
1894 Wgg = weight_scale
1895 Wmd = (weight_scale - Wed)/2
1896 Wgd = (weight_scale - Wed)/2
1898 If this system ends up with any values out of range (ie negative, or
1899 above weight_scale), use the constraints Wgg == 1 and Wee == 1, since
1900 both those positions are scarce:
1902 Wgg = weight_scale
1903 Wee = weight_scale
1904 Wed = (weight_scale*(D - 2*E + G + M))/(3*D)
1905 Wmd = (weight_Scale*(D - 2*M + G + E))/(3*D)
1906 Wme = 0
1907 Wmg = 0
1908 Wgd = weight_scale - Wed - Wmd
1910 If M > T/3, then the Wmd weight above will become negative. Set it to 0
1911 in this case:
1912 Wmd = 0
1913 Wgd = weight_scale - Wed
1915 Case 3: One of E < T/3 or G < T/3
1917 Let S be the scarce class (of E or G).
1919 Subcase a: (S+D) < T/3:
1920 if G=S:
1921 Wgg = Wgd = weight_scale;
1922 Wmd = Wed = Wmg = 0;
1923 // Minor subcase, if E is more scarce than M,
1924 // keep its bandwidth in place.
1925 if (E < M) Wme = 0;
1926 else Wme = (weight_scale*(E-M))/(2*E);
1927 Wee = weight_scale-Wme;
1928 if E=S:
1929 Wee = Wed = weight_scale;
1930 Wmd = Wgd = Wme = 0;
1931 // Minor subcase, if G is more scarce than M,
1932 // keep its bandwidth in place.
1933 if (G < M) Wmg = 0;
1934 else Wmg = (weight_scale*(G-M))/(2*G);
1935 Wgg = weight_scale-Wmg;
1937 Subcase b: (S+D) >= T/3
1938 if G=S:
1939 Add constraints Wgg = 1, Wmd == Wed to maximize bandwidth
1940 in the guard position, while still allowing exits to be
1941 used as middle nodes:
1942 Wgg = weight_scale
1943 Wgd = (weight_scale*(D - 2*G + E + M))/(3*D)
1944 Wmg = 0
1945 Wee = (weight_scale*(E+M))/(2*E)
1946 Wme = weight_scale - Wee
1947 Wmd = (weight_scale - Wgd)/2
1948 Wed = (weight_scale - Wgd)/2
1949 if E=S:
1950 Add constraints Wee == 1, Wmd == Wgd to maximize bandwidth
1951 in the exit position:
1952 Wee = weight_scale;
1953 Wed = (weight_scale*(D - 2*E + G + M))/(3*D);
1954 Wme = 0;
1955 Wgg = (weight_scale*(G+M))/(2*G);
1956 Wmg = weight_scale - Wgg;
1957 Wmd = (weight_scale - Wed)/2;
1958 Wgd = (weight_scale - Wed)/2;
1960 To ensure consensus, all calculations are performed using integer math
1961 with a fixed precision determined by the bwweightscale consensus
1962 parameter (defaults at 10000, Min: 1, Max: INT32_MAX).
1964 For future balancing improvements, Tor clients support 11 additional weights
1965 for directory requests and middle weighting. These weights are currently
1966 set at weight_scale, with the exception of the following groups of
1967 assignments:
1969 Directory requests use middle weights:
1970 Wbd=Wmd, Wbg=Wmg, Wbe=Wme, Wbm=Wmm
1972 Handle bridges and strange exit policies:
1973 Wgm=Wgg, Wem=Wee, Weg=Wed
1975 3.6. Consensus flavors
1977 Consensus flavors are variants of the consensus that clients can choose
1978 to download and use instead of the unflavored consensus. The purpose
1979 of a consensus flavor is to remove or replace information in the
1980 unflavored consensus without forcing clients to download information
1981 they would not use anyway.
1983 Directory authorities can produce and serve an arbitrary number of
1984 flavors of the same consensus. A downside of creating too many new
1985 flavors is that clients will be distinguishable based on which flavor
1986 they download. A new flavor should not be created when adding a field
1987 instead wouldn't be too onerous.
1989 Examples for consensus flavors include:
1990 - Publishing hashes of microdescriptors instead of hashes of
1991 full descriptors (see 3.6.2).
1992 - Including different digests of descriptors, instead of the
1993 perhaps-soon-to-be-totally-broken SHA1.
1995 Consensus flavors are derived from the unflavored consensus once the
1996 voting process is complete. This is to avoid consensus synchronization
1997 problems.
1999 Every consensus flavor has a name consisting of a sequence of one
2000 or more alphanumeric characters and dashes. For compatibility,
2001 current descriptor flavor is called "ns".
2003 The supported consensus flavors are defined as part of the
2004 authorities' consensus method.
2006 All consensus flavors have in common that their first line is
2007 "network-status-version" where version is 3 or higher, and the flavor
2008 is a string consisting of alphanumeric characters and dashes:
2010 "network-status-version" SP version SP flavor NL
2012 3.6.1. ns consensus
2014 The ns consensus flavor is equivalent to the unflavored consensus
2015 except for its first line which states its consensus flavor name:
2017 "network-status-version" SP version SP "ns" NL
2019 [At start, exactly once.]
2021 3.6.2. Microdescriptor consensus
2023 The microdescriptor consensus is a consensus flavor that contains
2024 microdescriptor hashes instead of descriptor hashes and that omits
2025 exit-policy summaries which are contained in microdescriptors. The
2026 microdescriptor consensus was designed to contain elements that are
2027 small and frequently changing. Clients use the information in the
2028 microdescriptor consensus to decide which servers to fetch information
2029 about and which servers to fetch information from.
2031 The microdescriptor consensus is based on the unflavored consensus with
2032 the exceptions as follows:
2034 "network-status-version" SP version SP "microdesc" NL
2036 [At start, exactly once.]
2038 The flavor name of a microdescriptor consensus is "microdesc".
2040 Changes to router status entries are as follows:
2042 "r" SP nickname SP identity SP publication SP IP SP ORPort
2043 SP DirPort NL
2045 [At start, exactly once.]
2047 Similar to "r" lines in 3.3, but without the digest element.
2049 "p" ... NL
2051 [Zero times.]
2053 Exit policy summaries are contained in microdescriptors and
2054 therefore omitted in the microdescriptor consensus.
2056 "m" SP digest NL
2058 [Exactly once.]
2060 "digest" is the base64 of the SHA256 hash of the router's
2061 microdescriptor with trailing =s omitted. For a given router
2062 descriptor digest and consensus method there should only be a
2063 single microdescriptor digest in the "m" lines of all votes.
2064 If different votes have different microdescriptor digests for
2065 the same descriptor digest and consensus method, at least one
2066 of the authorities is broken. If this happens, the microdesc
2067 consensus should contain whichever microdescriptor digest is
2068 most common. If there is no winner, we break ties in the favor
2069 of the lexically earliest.
2071 3.7. Detached signatures
2073 Assuming full connectivity, every authority should compute and sign the
2074 same consensus including any flavors in each period. Therefore, it
2075 isn't necessary to download the consensus or any flavors of it computed
2076 by each authority; instead, the authorities only push/fetch each
2077 others' signatures. A "detached signature" document contains items as
2078 follows:
2080 "consensus-digest" SP Digest NL
2082 [At start, at most once.]
2084 The digest of the consensus being signed.
2086 "valid-after" SP YYYY-MM-DD SP HH:MM:SS NL
2087 "fresh-until" SP YYYY-MM-DD SP HH:MM:SS NL
2088 "valid-until" SP YYYY-MM-DD SP HH:MM:SS NL
2090 [As in the consensus]
2092 "additional-digest" SP flavor SP algname SP digest NL
2094 [Any number.]
2096 For each supported consensus flavor, every directory authority
2097 adds one or more "additional-digest" lines. "flavor" is the name
2098 of the consensus flavor, "algname" is the name of the hash
2099 algorithm that is used to generate the digest, and "digest" is the
2100 hex-encoded digest.
2102 The hash algorithm for the microdescriptor consensus flavor is
2103 defined as SHA256 with algname "sha256".
2105 "additional-signature" SP flavor SP algname SP identity SP
2106 signing-key-digest NL signature.
2108 [Any number.]
2110 For each supported consensus flavor and defined digest algorithm,
2111 every directory authority adds an "additional-signature" line.
2112 "flavor" is the name of the consensus flavor. "algname" is the
2113 name of the algorithm that was used to hash the identity and
2114 signing keys, and to compute the signature. "identity" is the
2115 hex-encoded digest of the authority identity key of the signing
2116 authority, and "signing-key-digest" is the hex-encoded digest of
2117 the current authority signing key of the signing authority.
2119 The "sha256" signature format is defined as the RSA signature of
2120 the OAEP+-padded SHA256 digest of the item to be signed. When
2121 checking signatures, the signature MUST be treated as valid if the
2122 signature material begins with SHA256(document), so that other
2123 data can get added later.
2124 [To be honest, I didn't fully understand the previous paragraph
2125 and only copied it from the proposals. Review carefully. -KL]
2127 "directory-signature"
2129 [As in the consensus; the signature object is the same as in the
2130 consensus document.]
2132 4. Directory server operation
2134 All directory authorities and directory caches ("directory servers")
2135 implement this section, except as noted.
2137 4.1. Accepting uploads (authorities only)
2139 When a router posts a signed descriptor to a directory authority, the
2140 authority first checks whether it is well-formed and correctly
2141 self-signed. If it is, the authority next verifies that the nickname
2142 in question is not already assigned to a router with a different
2143 public key.
2144 Finally, the authority MAY check that the router is not blacklisted
2145 because of its key, IP, or another reason.
2147 If the descriptor passes these tests, and the authority does not already
2148 have a descriptor for a router with this public key, it accepts the
2149 descriptor and remembers it.
2151 If the authority _does_ have a descriptor with the same public key, the
2152 newly uploaded descriptor is remembered if its publication time is more
2153 recent than the most recent old descriptor for that router, and either:
2154 - There are non-cosmetic differences between the old descriptor and the
2155 new one.
2156 - Enough time has passed between the descriptors' publication times.
2157 (Currently, 12 hours.)
2159 Differences between router descriptors are "non-cosmetic" if they would be
2160 sufficient to force an upload as described in section 2 above.
2162 Note that the "cosmetic difference" test only applies to uploaded
2163 descriptors, not to descriptors that the authority downloads from other
2164 authorities.
2166 When a router posts a signed extra-info document to a directory authority,
2167 the authority again checks it for well-formedness and correct signature,
2168 and checks that its matches the extra-info-digest in some router
2169 descriptor that it believes is currently useful. If so, it accepts it and
2170 stores it and serves it as requested. If not, it drops it.
2172 4.2. Voting (authorities only)
2174 Authorities divide time into Intervals. Authority administrators SHOULD
2175 try to all pick the same interval length, and SHOULD pick intervals that
2176 are commonly used divisions of time (e.g., 5 minutes, 15 minutes, 30
2177 minutes, 60 minutes, 90 minutes). Voting intervals SHOULD be chosen to
2178 divide evenly into a 24-hour day.
2180 Authorities SHOULD act according to interval and delays in the
2181 latest consensus. Lacking a latest consensus, they SHOULD default to a
2182 30-minute Interval, a 5 minute VotingDelay, and a 5 minute DistDelay.
2184 Authorities MUST take pains to ensure that their clocks remain accurate
2185 within a few seconds. (Running NTP is usually sufficient.)
2187 The first voting period of each day begins at 00:00 (midnight) GMT. If
2188 the last period of the day would be truncated by one-half or more, it is
2189 merged with the second-to-last period.
2191 An authority SHOULD publish its vote immediately at the start of each voting
2192 period (minus VoteSeconds+DistSeconds). It does this by making it
2193 available at
2194 http://<hostname>/tor/status-vote/next/authority.z
2195 and sending it in an HTTP POST request to each other authority at the URL
2196 http://<hostname>/tor/post/vote
2198 If, at the start of the voting period, minus DistSeconds, an authority
2199 does not have a current statement from another authority, the first
2200 authority downloads the other's statement.
2202 Once an authority has a vote from another authority, it makes it available
2203 at
2204 http://<hostname>/tor/status-vote/next/<fp>.z
2205 where <fp> is the fingerprint of the other authority's identity key.
2206 And at
2207 http://<hostname>/tor/status-vote/next/d/<d>.z
2208 where <d> is the digest of the vote document.
2210 The consensus status, along with as many signatures as the server
2211 currently knows, should be available at
2212 http://<hostname>/tor/status-vote/next/consensus.z
2213 All of the detached signatures it knows for consensus status should be
2214 available at:
2215 http://<hostname>/tor/status-vote/next/consensus-signatures.z
2217 Once there are enough signatures, or once the voting period starts,
2218 these documents are available at
2219 http://<hostname>/tor/status-vote/current/consensus.z
2220 and
2221 http://<hostname>/tor/status-vote/current/consensus-signatures.z
2222 [XXX current/consensus-signatures is not currently implemented, as it
2223 is not used in the voting protocol.]
2225 The other vote documents are analogously made available under
2226 http://<hostname>/tor/status-vote/current/authority.z
2227 http://<hostname>/tor/status-vote/current/<fp>.z
2228 http://<hostname>/tor/status-vote/current/d/<d>.z
2229 once the consensus is complete.
2231 Once an authority has computed and signed a consensus network status, it
2232 should send its detached signature to each other authority in an HTTP POST
2233 request to the URL:
2234 http://<hostname>/tor/post/consensus-signature
2236 [XXX Note why we support push-and-then-pull.]
2238 [XXX possible future features include support for downloading old
2239 consensuses.]
2241 The authorities serve another consensus of each flavor "F" from the
2242 locations
2243 /tor/status-vote/(current|next)/consensus-F.z. and
2244 /tor/status-vote/(current|next)/consensus-F/<FP1>+....z.
2246 4.3. Downloading consensus status documents (caches only)
2248 All directory servers (authorities and caches) try to keep a recent
2249 network-status consensus document to serve to clients. A cache ALWAYS
2250 downloads a network-status consensus if any of the following are true:
2251 - The cache has no consensus document.
2252 - The cache's consensus document is no longer valid.
2253 Otherwise, the cache downloads a new consensus document at a randomly
2254 chosen time in the first half-interval after its current consensus
2255 stops being fresh. (This time is chosen at random to avoid swarming
2256 the authorities at the start of each period. The interval size is
2257 inferred from the difference between the valid-after time and the
2258 fresh-until time on the consensus.)
2260 [For example, if a cache has a consensus that became valid at 1:00,
2261 and is fresh until 2:00, that cache will fetch a new consensus at
2262 a random time between 2:00 and 2:30.]
2264 Directory caches also fetch consensus flavors from the authorities.
2265 Caches check the correctness of consensus flavors, but do not check
2266 anything about an unrecognized consensus document beyond its digest and
2267 length. Caches serve all consensus flavors from the same locations as
2268 the directory authorities.
2270 4.4. Downloading and storing router descriptors (authorities and caches)
2272 Periodically (currently, every 10 seconds), directory servers check
2273 whether there are any specific descriptors that they do not have and that
2274 they are not currently trying to download. Caches identify these
2275 descriptors by hash in the recent network-status consensus documents;
2276 authorities identify them by hash in vote (if publication date is more
2277 recent than the descriptor we currently have).
2279 [XXXX need a way to fetch descriptors ahead of the vote? v2 status docs can
2280 do that for now.]
2282 If so, the directory server launches requests to the authorities for these
2283 descriptors, such that each authority is only asked for descriptors listed
2284 in its most recent vote (if the requester is an authority) or in the
2285 consensus (if the requester is a cache). If we're an authority, and more
2286 than one authority lists the descriptor, we choose which to ask at random.
2288 If one of these downloads fails, we do not try to download that descriptor
2289 from the authority that failed to serve it again unless we receive a newer
2290 network-status (consensus or vote) from that authority that lists the same
2291 descriptor.
2293 Directory servers must potentially cache multiple descriptors for each
2294 router. Servers must not discard any descriptor listed by any recent
2295 consensus. If there is enough space to store additional descriptors,
2296 servers SHOULD try to hold those which clients are likely to download the
2297 most. (Currently, this is judged based on the interval for which each
2298 descriptor seemed newest.)
2299 [XXXX define recent]
2301 Authorities SHOULD NOT download descriptors for routers that they would
2302 immediately reject for reasons listed in 3.1.
2304 4.5. Downloading and storing microdescriptors (caches only)
2306 Directory mirrors should fetch, cache, and serve each microdescriptor
2307 from the authorities.
2309 The microdescriptors with base64 hashes <D1>,<D2>,<D3> are available
2310 at:
2311 http://<hostname>/tor/micro/d/<D1>-<D2>-<D3>[.z]
2313 <Dn> are base-64 encoded with trailing =s omitted for size and for
2314 consistency with the microdescriptor consensus format. -s are used
2315 instead of +s to separate items, since the + character is used in
2316 base64 encoding.
2318 All the microdescriptors from the current consensus should also be
2319 available at:
2320 http://<hostname>/tor/micro/all[.z]
2321 so a client that's bootstrapping doesn't need to send a 70KB URL just
2322 to name every microdescriptor it's looking for.
2323 [Note that /tor/micro/all[.z] is not implemented as of February 21,
2324 2012. -KL]
2326 Directory mirrors should check to make sure that the microdescriptors
2327 they're about to serve match the right hashes (either the hashes from
2328 the fetch URL or the hashes from the consensus, respectively).
2330 4.6. Downloading and storing extra-info documents
2332 All authorities, and any cache that chooses to cache extra-info documents,
2333 and any client that uses extra-info documents, should implement this
2334 section.
2336 Note that generally, clients don't need extra-info documents.
2338 Periodically, the Tor instance checks whether it is missing any extra-info
2339 documents: in other words, if it has any router descriptors with an
2340 extra-info-digest field that does not match any of the extra-info
2341 documents currently held. If so, it downloads whatever extra-info
2342 documents are missing. Caches download from authorities; non-caches try
2343 to download from caches. We follow the same splitting and back-off rules
2344 as in 4.4 (if a cache) or 5.3 (if a client).
2346 4.7. General-use HTTP URLs
2348 "Fingerprints" in these URLs are base-16-encoded SHA1 hashes.
2350 The most recent v3 consensus should be available at:
2351 http://<hostname>/tor/status-vote/current/consensus.z
2353 Starting with Tor version 0.2.1.1-alpha is also available at:
2354 http://<hostname>/tor/status-vote/current/consensus/<F1>+<F2>+<F3>.z
2356 Where F1, F2, etc. are authority identity fingerprints the client trusts.
2357 Servers will only return a consensus if more than half of the requested
2358 authorities have signed the document, otherwise a 404 error will be sent
2359 back. The fingerprints can be shortened to a length of any multiple of
2360 two, using only the leftmost part of the encoded fingerprint. Tor uses
2361 3 bytes (6 hex characters) of the fingerprint.
2363 Clients SHOULD sort the fingerprints in ascending order. Server MUST
2364 accept any order.
2366 Clients SHOULD use this format when requesting consensus documents from
2367 directory authority servers and from caches running a version of Tor
2368 that is known to support this URL format.
2370 A concatenated set of all the current key certificates should be available
2371 at:
2372 http://<hostname>/tor/keys/all.z
2374 The key certificate for this server (if it is an authority) should be
2375 available at:
2376 http://<hostname>/tor/keys/authority.z
2378 The key certificate for an authority whose authority identity fingerprint
2379 is <F> should be available at:
2380 http://<hostname>/tor/keys/fp/<F>.z
2382 The key certificate whose signing key fingerprint is <F> should be
2383 available at:
2384 http://<hostname>/tor/keys/sk/<F>.z
2386 The key certificate whose identity key fingerprint is <F> and whose signing
2387 key fingerprint is <S> should be available at:
2389 http://<hostname>/tor/keys/fp-sk/<F>-<S>.z
2391 (As usual, clients may request multiple certificates using:
2392 http://<hostname>/tor/keys/fp-sk/<F1>-<S1>+<F2>-<S2>.z )
2393 [The above fp-sk format was not supported before Tor 0.2.1.9-alpha.]
2395 The most recent descriptor for a server whose identity key has a
2396 fingerprint of <F> should be available at:
2397 http://<hostname>/tor/server/fp/<F>.z
2399 The most recent descriptors for servers with identity fingerprints
2400 <F1>,<F2>,<F3> should be available at:
2401 http://<hostname>/tor/server/fp/<F1>+<F2>+<F3>.z
2403 (NOTE: Implementations SHOULD NOT download descriptors by identity key
2404 fingerprint. This allows a corrupted server (in collusion with a cache) to
2405 provide a unique descriptor to a client, and thereby partition that client
2406 from the rest of the network.)
2408 The server descriptor with (descriptor) digest <D> (in hex) should be
2409 available at:
2410 http://<hostname>/tor/server/d/<D>.z
2412 The most recent descriptors with digests <D1>,<D2>,<D3> should be
2413 available at:
2414 http://<hostname>/tor/server/d/<D1>+<D2>+<D3>.z
2416 The most recent descriptor for this server should be at:
2417 http://<hostname>/tor/server/authority.z
2418 [Nothing in the Tor protocol uses this resource yet, but it is useful
2419 for debugging purposes. Also, the official Tor implementations
2420 (starting at 0.1.1.x) use this resource to test whether a server's
2421 own DirPort is reachable.]
2423 A concatenated set of the most recent descriptors for all known servers
2424 should be available at:
2425 http://<hostname>/tor/server/all.z
2427 Extra-info documents are available at the URLS
2428 http://<hostname>/tor/extra/d/...
2429 http://<hostname>/tor/extra/fp/...
2430 http://<hostname>/tor/extra/all[.z]
2431 http://<hostname>/tor/extra/authority[.z]
2432 (As for /tor/server/ URLs: supports fetching extra-info
2433 documents by their digest, by the fingerprint of their servers,
2434 or all at once. When serving by fingerprint, we serve the
2435 extra-info that corresponds to the descriptor we would serve by
2436 that fingerprint. Only directory authorities of version
2437 0.2.0.1-alpha or later are guaranteed to support the first
2438 three classes of URLs. Caches may support them, and MUST
2439 support them if they have advertised "caches-extra-info".)
2441 For debugging, directories SHOULD expose non-compressed objects at URLs like
2442 the above, but without the final ".z".
2443 Clients MUST handle compressed concatenated information in two forms:
2444 - A concatenated list of zlib-compressed objects.
2445 - A zlib-compressed concatenated list of objects.
2446 Directory servers MAY generate either format: the former requires less
2447 CPU, but the latter requires less bandwidth.
2449 Clients SHOULD use upper case letters (A-F) when base16-encoding
2450 fingerprints. Servers MUST accept both upper and lower case fingerprints
2451 in requests.
2453 [XXX Add new URLs for microdescriptors, consensus flavors, and
2454 microdescriptor consensus. -KL]
2456 5. Client operation: downloading information
2458 Every Tor that is not a directory server (that is, those that do
2459 not have a DirPort set) implements this section.
2461 5.1. Downloading network-status documents
2463 Each client maintains a list of directory authorities. Insofar as
2464 possible, clients SHOULD all use the same list.
2466 Clients try to have a live consensus network-status document at all times.
2467 A network-status document is "live" if the time in its valid-until field
2468 has not passed.
2470 If a client is missing a live network-status document, it tries to fetch
2471 it from a directory cache (or from an authority if it knows no caches).
2472 On failure, the client waits briefly, then tries that network-status
2473 document again from another cache. The client does not build circuits
2474 until it has a live network-status consensus document, and it has
2475 descriptors for more than 1/4 of the routers that it believes are running.
2477 (Note: clients can and should pick caches based on the network-status
2478 information they have: once they have first fetched network-status info
2479 from an authority, they should not need to go to the authority directly
2480 again.)
2482 To avoid swarming the caches whenever a consensus expires, the
2483 clients download new consensuses at a randomly chosen time after the
2484 caches are expected to have a fresh consensus, but before their
2485 consensus will expire. (This time is chosen uniformly at random from
2486 the interval between the time 3/4 into the first interval after the
2487 consensus is no longer fresh, and 7/8 of the time remaining after
2488 that before the consensus is invalid.)
2490 [For example, if a cache has a consensus that became valid at 1:00,
2491 and is fresh until 2:00, and expires at 4:00, that cache will fetch
2492 a new consensus at a random time between 2:45 and 3:50, since 3/4
2493 of the one-hour interval is 45 minutes, and 7/8 of the remaining 75
2494 minutes is 65 minutes.]
2496 Clients may choose to download the microdescriptor consensus instead
2497 of the general network status consensus. In that case they should use
2498 the same update strategy as for the normal consensus. They should not
2499 download more than one consensus flavor.
2501 5.2. Downloading and storing router descriptors or microdescriptors
2503 Clients try to have the best descriptor for each router. A descriptor is
2504 "best" if:
2505 * It is listed in the consensus network-status document.
2507 Periodically (currently every 10 seconds) clients check whether there are
2508 any "downloadable" descriptors. A descriptor is downloadable if:
2509 - It is the "best" descriptor for some router.
2510 - The descriptor was published at least 10 minutes in the past.
2511 (This prevents clients from trying to fetch descriptors that the
2512 mirrors have probably not yet retrieved and cached.)
2513 - The client does not currently have it.
2514 - The client is not currently trying to download it.
2515 - The client would not discard it immediately upon receiving it.
2516 - The client thinks it is running and valid (see 6.1 below).
2518 If at least 16 known routers have downloadable descriptors, or if
2519 enough time (currently 10 minutes) has passed since the last time the
2520 client tried to download descriptors, it launches requests for all
2521 downloadable descriptors, as described in 5.3 below.
2523 When a descriptor download fails, the client notes it, and does not
2524 consider the descriptor downloadable again until a certain amount of time
2525 has passed. (Currently 0 seconds for the first failure, 60 seconds for the
2526 second, 5 minutes for the third, 10 minutes for the fourth, and 1 day
2527 thereafter.) Periodically (currently once an hour) clients reset the
2528 failure count.
2530 Clients retain the most recent descriptor they have downloaded for each
2531 router so long as it is not too old (currently, 48 hours), OR so long as
2532 no better descriptor has been downloaded for the same router.
2534 [Versions of Tor before 0.1.2.3-alpha would discard descriptors simply for
2535 being published too far in the past.] [The code seems to discard
2536 descriptors in all cases after they're 5 days old. True? -RD]
2538 Clients which chose to download the microdescriptor consensus instead
2539 of the general consensus must download the referenced microdescriptors
2540 instead of router descriptors. Clients fetch and cache
2541 microdescriptors preemptively from dir mirrors when starting up, like
2542 they currently fetch descriptors. After bootstrapping, clients only
2543 need to fetch the microdescriptors that have changed.
2545 Clients maintain a cache of microdescriptors along with metadata like
2546 when it was last referenced by a consensus, and which identity key
2547 it corresponds to. They keep a microdescriptor until it hasn't been
2548 mentioned in any consensus for a week. Future clients might cache them
2549 for longer or shorter times.
2551 5.3. Managing downloads
2553 When a client has no consensus network-status document, it downloads it
2554 from a randomly chosen authority. In all other cases, the client
2555 downloads from caches randomly chosen from among those believed to be V2
2556 directory servers. (This information comes from the network-status
2557 documents; see 6 below.)
2559 When downloading multiple router descriptors, the client chooses multiple
2560 mirrors so that:
2561 - At least 3 different mirrors are used, except when this would result
2562 in more than one request for under 4 descriptors.
2563 - No more than 128 descriptors are requested from a single mirror.
2564 - Otherwise, as few mirrors as possible are used.
2565 After choosing mirrors, the client divides the descriptors among them
2566 randomly.
2568 After receiving any response client MUST discard any network-status
2569 documents and descriptors that it did not request.
2571 When a client gets a new microdescriptor consensus, it looks to see if
2572 there are any microdescriptors it needs to learn. If it needs to learn
2573 more than half of the microdescriptors, it requests 'all', else it
2574 requests only the missing ones. Clients MAY try to determine whether
2575 the upload bandwidth for listing the microdescriptors they want is more
2576 or less than the download bandwidth for the microdescriptors they do
2577 not want.
2579 6. Using directory information
2581 Everyone besides directory authorities uses the approaches in this section
2582 to decide which relays to use and what their keys are likely to be.
2583 (Directory authorities just believe their own opinions, as in 3.1 above.)
2585 6.1. Choosing routers for circuits.
2587 Circuits SHOULD NOT be built until the client has enough directory
2588 information: a live consensus network status [XXXX fallback?] and
2589 descriptors for at least 1/4 of the relays believed to be running.
2591 A relay is "listed" if it is included by the consensus network-status
2592 document. Clients SHOULD NOT use unlisted relays.
2594 These flags are used as follows:
2596 - Clients SHOULD NOT use non-'Valid' or non-'Running' routers unless
2597 requested to do so.
2599 - Clients SHOULD NOT use non-'Fast' routers for any purpose other than
2600 very-low-bandwidth circuits (such as introduction circuits).
2602 - Clients SHOULD NOT use non-'Stable' routers for circuits that are
2603 likely to need to be open for a very long time (such as those used for
2604 IRC or SSH connections).
2606 - Clients SHOULD NOT choose non-'Guard' nodes when picking entry guard
2607 nodes.
2609 - Clients SHOULD NOT download directory information from non-'V2Dir'
2610 caches.
2612 See the "path-spec.txt" document for more details.
2614 6.2. Managing naming
2616 In order to provide human-memorable names for individual router
2617 identities, some directory servers bind names to IDs. Clients handle
2618 names in two ways:
2620 When a client encounters a name it has not mapped before:
2622 If the consensus lists any router with that name as "Named", or if
2623 consensus-method 2 or later is in use and the consensus lists any
2624 router with that name as having the "Unnamed" flag, then the name is
2625 bound. (It's bound to the ID listed in the entry with the Named,
2626 or to an unknown ID if no name is found.)
2628 When the user refers to a bound name, the implementation SHOULD provide
2629 only the router with ID bound to that name, and no other router, even
2630 if the router with the right ID can't be found.
2632 When a user tries to refer to a non-bound name, the implementation SHOULD
2633 warn the user. After warning the user, the implementation MAY use any
2634 router that advertises the name.
2636 Not every router needs a nickname. When a router doesn't configure a
2637 nickname, it publishes with the default nickname "Unnamed". Authorities
2638 SHOULD NOT ever mark a router with this nickname as Named; client software
2639 SHOULD NOT ever use a router in response to a user request for a router
2640 called "Unnamed".
2642 6.3. Software versions
2644 An implementation of Tor SHOULD warn when it has fetched a consensus
2645 network-status, and it is running a software version not listed.
2647 6.4. Warning about a router's status.
2649 If a router tries to publish its descriptor to a Naming authority
2650 that has its nickname mapped to another key, the router SHOULD
2651 warn the operator that it is either using the wrong key or is using
2652 an already claimed nickname.
2654 If a router has fetched a consensus document,, and the
2655 authorities do not publish a binding for the router's nickname, the
2656 router MAY remind the operator that the chosen nickname is not
2657 bound to this key at the authorities, and suggest contacting the
2658 authority operators.
2660 ...
2662 6.5. Router protocol versions
2664 A client should believe that a router supports a given feature if that
2665 feature is supported by the router or protocol versions in more than half
2666 of the live networkstatuses' "v" entries for that router. In other words,
2667 if the "v" entries for some router are:
2668 v Tor 0.0.8pre1 (from authority 1)
2669 v Tor 0.1.2.11 (from authority 2)
2670 v FutureProtocolDescription 99 (from authority 3)
2671 then the client should believe that the router supports any feature
2672 supported by 0.1.2.11.
2674 This is currently equivalent to believing the median declared version for
2675 a router in all live networkstatuses.
2677 7. Standards compliance
2679 All clients and servers MUST support HTTP 1.0. Clients and servers MAY
2680 support later versions of HTTP as well.
2682 7.1. HTTP headers
2684 Servers MAY set the Content-Length: header. Servers SHOULD set
2685 Content-Encoding to "deflate" or "identity".
2687 Servers MAY include an X-Your-Address-Is: header, whose value is the
2688 apparent IP address of the client connecting to them (as a dotted quad).
2689 For directory connections tunneled over a BEGIN_DIR stream, servers SHOULD
2690 report the IP from which the circuit carrying the BEGIN_DIR stream reached
2691 them. [Servers before version 0.1.2.5-alpha reported 127.0.0.1 for all
2692 BEGIN_DIR-tunneled connections.]
2694 Servers SHOULD disable caching of multiple network statuses or multiple
2695 router descriptors. Servers MAY enable caching of single descriptors,
2696 single network statuses, the list of all router descriptors, a v1
2697 directory, or a v1 running routers document. XXX mention times.
2699 7.2. HTTP status codes
2701 Tor delivers the following status codes. Some were chosen without much
2702 thought; other code SHOULD NOT rely on specific status codes yet.
2704 200 -- the operation completed successfully
2705 -- the user requested statuses or serverdescs, and none of the ones we
2706 requested were found (0.2.0.4-alpha and earlier).
2708 304 -- the client specified an if-modified-since time, and none of the
2709 requested resources have changed since that time.
2711 400 -- the request is malformed, or
2712 -- the URL is for a malformed variation of one of the URLs we support,
2713 or
2714 -- the client tried to post to a non-authority, or
2715 -- the authority rejected a malformed posted document, or
2717 404 -- the requested document was not found.
2718 -- the user requested statuses or serverdescs, and none of the ones
2719 requested were found (0.2.0.5-alpha and later).
2721 503 -- we are declining the request in order to save bandwidth
2722 -- user requested some items that we ordinarily generate or store,
2723 but we do not have any available.
2725 9. Backward compatibility and migration plans
2727 Until Tor versions before 0.1.1.x are completely obsolete, directory
2728 authorities should generate, and mirrors should download and cache, v1
2729 directories and running-routers lists, and allow old clients to download
2730 them. These documents and the rules for retrieving, serving, and caching
2731 them are described in dir-spec-v1.txt.
2733 Until Tor versions before 0.2.0.x are completely obsolete, directory
2734 authorities should generate, mirrors should download and cache, v2
2735 network-status documents, and allow old clients to download them.
2736 Additionally, all directory servers and caches should download, store, and
2737 serve any router descriptor that is required because of v2 network-status
2738 documents. These documents and the rules for retrieving, serving, and
2739 caching them are described in dir-spec-v1.txt.
2741 A. Consensus-negotiation timeline.
2743 Period begins: this is the Published time.
2744 Everybody sends votes
2745 Reconciliation: everybody tries to fetch missing votes.
2746 consensus may exist at this point.
2747 End of voting period:
2748 everyone swaps signatures.
2749 Now it's okay for caches to download
2750 Now it's okay for clients to download.
2752 Valid-after/valid-until switchover