3 Tor Protocol Specification
8 Note: This document aims to specify Tor as implemented in 0.2.1.x. Future
9 versions of Tor may implement improved protocols, and compatibility is not
10 guaranteed. Compatibility notes are given for versions 0.1.1.15-rc and
11 later; earlier versions are not compatible with the Tor network as of this
14 This specification is not a design document; most design criteria
15 are not examined. For more information on why Tor acts as it does,
20 0.1. Notation and encoding
24 K -- a key for a symmetric cypher.
26 a|b -- concatenation of 'a' and 'b'.
28 [A0 B1 C2] -- a three-byte sequence, containing the bytes with
29 hexadecimal values A0, B1, and C2, in that order.
31 All numeric values are encoded in network (big-endian) order.
33 H(m) -- a cryptographic hash of m.
35 0.2. Security parameters
37 Tor uses a stream cipher, a public-key cipher, the Diffie-Hellman
38 protocol, and a hash function.
40 KEY_LEN -- the length of the stream cipher's key, in bytes.
42 PK_ENC_LEN -- the length of a public-key encrypted message, in bytes.
43 PK_PAD_LEN -- the number of bytes added in padding for public-key
44 encryption, in bytes. (The largest number of bytes that can be encrypted
45 in a single public-key operation is therefore PK_ENC_LEN-PK_PAD_LEN.)
47 DH_LEN -- the number of bytes used to represent a member of the
49 DH_SEC_LEN -- the number of bytes used in a Diffie-Hellman private key (x).
51 HASH_LEN -- the length of the hash function's output, in bytes.
53 PAYLOAD_LEN -- The longest allowable cell payload, in bytes. (509)
55 CELL_LEN -- The length of a Tor cell, in bytes.
59 For a stream cipher, we use 128-bit AES in counter mode, with an IV of all
62 For a public-key cipher, we use RSA with 1024-bit keys and a fixed
63 exponent of 65537. We use OAEP-MGF1 padding, with SHA-1 as its digest
64 function. We leave the optional "Label" parameter unset. (For OAEP
65 padding, see ftp://ftp.rsasecurity.com/pub/pkcs/pkcs-1/pkcs-1v2-1.pdf)
67 For Diffie-Hellman, we use a generator (g) of 2. For the modulus (p), we
68 use the 1024-bit safe prime from rfc2409 section 6.2 whose hex
71 "FFFFFFFFFFFFFFFFC90FDAA22168C234C4C6628B80DC1CD129024E08"
72 "8A67CC74020BBEA63B139B22514A08798E3404DDEF9519B3CD3A431B"
73 "302B0A6DF25F14374FE1356D6D51C245E485B576625E7EC6F44C42E9"
74 "A637ED6B0BFF5CB6F406B7EDEE386BFB5A899FA5AE9F24117C4B1FE6"
75 "49286651ECE65381FFFFFFFFFFFFFFFF"
77 As an optimization, implementations SHOULD choose DH private keys (x) of
78 320 bits. Implementations that do this MUST never use any DH key more
80 [May other implementations reuse their DH keys?? -RD]
81 [Probably not. Conceivably, you could get away with changing DH keys once
82 per second, but there are too many oddball attacks for me to be
83 comfortable that this is safe. -NM]
85 For a hash function, we use SHA-1.
88 DH_LEN=128; DH_SEC_LEN=40.
89 PK_ENC_LEN=128; PK_PAD_LEN=42.
92 When we refer to "the hash of a public key", we mean the SHA-1 hash of the
93 DER encoding of an ASN.1 RSA public key (as specified in PKCS.1).
95 All "random" values should be generated with a cryptographically strong
96 random number generator, unless otherwise noted.
98 The "hybrid encryption" of a byte sequence M with a public key PK is
100 1. If M is less than PK_ENC_LEN-PK_PAD_LEN, pad and encrypt M with PK.
101 2. Otherwise, generate a KEY_LEN byte random key K.
102 Let M1 = the first PK_ENC_LEN-PK_PAD_LEN-KEY_LEN bytes of M,
103 and let M2 = the rest of M.
104 Pad and encrypt K|M1 with PK. Encrypt M2 with our stream cipher,
105 using the key K. Concatenate these encrypted values.
106 [XXX Note that this "hybrid encryption" approach does not prevent
107 an attacker from adding or removing bytes to the end of M. It also
108 allows attackers to modify the bytes not covered by the OAEP --
109 see Goldberg's PET2006 paper for details. We will add a MAC to this
112 0.4. Other parameter values
118 Tor is a distributed overlay network designed to anonymize
119 low-latency TCP-based applications such as web browsing, secure shell,
120 and instant messaging. Clients choose a path through the network and
121 build a ``circuit'', in which each node (or ``onion router'' or ``OR'')
122 in the path knows its predecessor and successor, but no other nodes in
123 the circuit. Traffic flowing down the circuit is sent in fixed-size
124 ``cells'', which are unwrapped by a symmetric key at each node (like
125 the layers of an onion) and relayed downstream.
129 Every Tor server has multiple public/private keypairs:
131 - A long-term signing-only "Identity key" used to sign documents and
132 certificates, and used to establish server identity.
133 - A medium-term "Onion key" used to decrypt onion skins when accepting
134 circuit extend attempts. (See 5.1.) Old keys MUST be accepted for at
135 least one week after they are no longer advertised. Because of this,
136 servers MUST retain old keys for a while after they're rotated.
137 - A short-term "Connection key" used to negotiate TLS connections.
138 Tor implementations MAY rotate this key as often as they like, and
139 SHOULD rotate this key at least once a day.
141 Tor servers are also identified by "nicknames"; these are specified in
146 Connections between two Tor servers, or between a client and a server,
147 use TLS/SSLv3 for link authentication and encryption. All
148 implementations MUST support the SSLv3 ciphersuite
149 "SSL_DHE_RSA_WITH_3DES_EDE_CBC_SHA", and SHOULD support the TLS
150 ciphersuite "TLS_DHE_RSA_WITH_AES_128_CBC_SHA" if it is available.
152 There are three acceptable ways to perform a TLS handshake when
153 connecting to a Tor server: "certificates up-front", "renegotiation", and
154 "backwards-compatible renegotiation". ("Backwards-compatible
155 renegotiation" is, as the name implies, compatible with both other
158 Before Tor 0.2.0.21, only "certificates up-front" was supported. In Tor
159 0.2.0.21 or later, "backwards-compatible renegotiation" is used.
161 In "certificates up-front", the connection initiator always sends a
162 two-certificate chain, consisting of an X.509 certificate using a
163 short-term connection public key and a second, self- signed X.509
164 certificate containing its identity key. The other party sends a similar
165 certificate chain. The initiator's ClientHello MUST NOT include any
166 ciphersuites other than:
167 TLS_DHE_RSA_WITH_AES_256_CBC_SHA
168 TLS_DHE_RSA_WITH_AES_128_CBC_SHA
169 SSL_DHE_RSA_WITH_3DES_EDE_CBC_SHA
170 SSL_DHE_DSS_WITH_3DES_EDE_CBC_SHA
172 In "renegotiation", the connection initiator sends no certificates, and
173 the responder sends a single connection certificate. Once the TLS
174 handshake is complete, the initiator renegotiates the handshake, with each
175 parties sending a two-certificate chain as in "certificates up-front".
176 The initiator's ClientHello MUST include at least once ciphersuite not in
177 the list above. The responder SHOULD NOT select any ciphersuite besides
178 those in the list above.
179 [The above "should not" is because some of the ciphers that
180 clients list may be fake.]
182 In "backwards-compatible renegotiation", the connection initiator's
183 ClientHello MUST include at least one ciphersuite other than those listed
184 above. The connection responder examines the initiator's ciphersuite list
185 to see whether it includes any ciphers other than those included in the
186 list above. If extra ciphers are included, the responder proceeds as in
187 "renegotiation": it sends a single certificate and does not request
188 client certificates. Otherwise (in the case that no extra ciphersuites
189 are included in the ClientHello) the responder proceeds as in
190 "certificates up-front": it requests client certificates, and sends a
191 two-certificate chain. In either case, once the responder has sent its
192 certificate or certificates, the initiator counts them. If two
193 certificates have been sent, it proceeds as in "certificates up-front";
194 otherwise, it proceeds as in "renegotiation".
196 All new implementations of the Tor server protocol MUST support
197 "backwards-compatible renegotiation"; clients SHOULD do this too. If
198 this is not possible, new client implementations MUST support both
199 "renegotiation" and "certificates up-front" and use the router's
200 published link protocols list (see dir-spec.txt on the "protocols" entry)
201 to decide which to use.
203 In all of the above handshake variants, certificates sent in the clear
204 SHOULD NOT include any strings to identify the host as a Tor server. In
205 the "renegotation" and "backwards-compatible renegotiation", the
206 initiator SHOULD chose a list of ciphersuites and TLS extensions chosen
207 to mimic one used by a popular web browser.
209 Responders MUST NOT select any TLS ciphersuite that lacks ephemeral keys,
210 or whose symmetric keys are less then KEY_LEN bits, or whose digests are
211 less than HASH_LEN bits. Responders SHOULD NOT select any SSLv3
212 ciphersuite other than those listed above.
214 Even though the connection protocol is identical, we will think of the
215 initiator as either an onion router (OR) if it is willing to relay
216 traffic for other Tor users, or an onion proxy (OP) if it only handles
217 local requests. Onion proxies SHOULD NOT provide long-term-trackable
218 identifiers in their handshakes.
220 In all handshake variants, once all certificates are exchanged, all
221 parties receiving certificates must confirm that the identity key is as
222 expected. (When initiating a connection, the expected identity key is
223 the one given in the directory; when creating a connection because of an
224 EXTEND cell, the expected identity key is the one given in the cell.) If
225 the key is not as expected, the party must close the connection.
227 When connecting to an OR, all parties SHOULD reject the connection if that
228 OR has a malformed or missing certificate. When accepting an incoming
229 connection, an OR SHOULD NOT reject incoming connections from parties with
230 malformed or missing certificates. (However, an OR should not believe
231 that an incoming connection is from another OR unless the certificates
232 are present and well-formed.)
234 [Before version 0.1.2.8-rc, ORs rejected incoming connections from ORs and
235 OPs alike if their certificates were missing or malformed.]
237 Once a TLS connection is established, the two sides send cells
238 (specified below) to one another. Cells are sent serially. All
239 cells are CELL_LEN bytes long. Cells may be sent embedded in TLS
240 records of any size or divided across TLS records, but the framing
241 of TLS records MUST NOT leak information about the type or contents
244 TLS connections are not permanent. Either side MAY close a connection
245 if there are no circuits running over it and an amount of time
246 (KeepalivePeriod, defaults to 5 minutes) has passed since the last time
247 any traffic was transmitted over the TLS connection. Clients SHOULD
248 also hold a TLS connection with no circuits open, if it is likely that a
249 circuit will be built soon using that connection.
251 (As an exception, directory servers may try to stay connected to all of
252 the ORs -- though this will be phased out for the Tor 0.1.2.x release.)
254 3. Cell Packet format
256 The basic unit of communication for onion routers and onion
257 proxies is a fixed-width "cell".
259 On a version 1 connection, each cell contains the following
264 Payload (padded with 0 bytes) [PAYLOAD_LEN bytes]
266 On a version 2 connection, all cells are as in version 1 connections,
267 except for the initial VERSIONS cell, whose format is:
269 Circuit [2 octets; set to 0]
270 Command [1 octet; set to 7 for VERSIONS]
271 Length [2 octets; big-endian integer]
272 Payload [Length bytes]
274 The CircID field determines which circuit, if any, the cell is
277 The 'Command' field holds one of the following values:
278 0 -- PADDING (Padding) (See Sec 7.2)
279 1 -- CREATE (Create a circuit) (See Sec 5.1)
280 2 -- CREATED (Acknowledge create) (See Sec 5.1)
281 3 -- RELAY (End-to-end data) (See Sec 5.5 and 6)
282 4 -- DESTROY (Stop using a circuit) (See Sec 5.4)
283 5 -- CREATE_FAST (Create a circuit, no PK) (See Sec 5.1)
284 6 -- CREATED_FAST (Circuit created, no PK) (See Sec 5.1)
285 7 -- VERSIONS (Negotiate proto version) (See Sec 4)
286 8 -- NETINFO (Time and address info) (See Sec 4)
288 The interpretation of 'Payload' depends on the type of the cell.
289 PADDING: Payload is unused.
290 CREATE: Payload contains the handshake challenge.
291 CREATED: Payload contains the handshake response.
292 RELAY: Payload contains the relay header and relay body.
293 DESTROY: Payload contains a reason for closing the circuit.
295 Upon receiving any other value for the command field, an OR must
296 drop the cell. Since more cell types may be added in the future, ORs
297 should generally not warn when encountering unrecognized commands.
299 The payload is padded with 0 bytes.
301 PADDING cells are currently used to implement connection keepalive.
302 If there is no other traffic, ORs and OPs send one another a PADDING
303 cell every few minutes.
305 CREATE, CREATED, and DESTROY cells are used to manage circuits;
308 RELAY cells are used to send commands and data along a circuit; see
311 VERSIONS and NETINFO cells are used to set up connections. See section 4
314 4. Negotiating and initializing connections
316 4.1. Negotiating versions with VERSIONS cells
318 There are multiple instances of the Tor link connection protocol. Any
319 connection negotiated using the "certificates up front" handshake (see
320 section 2 above) is "version 1". In any connection where both parties
321 have behaved as in the "renegotiation" handshake, the link protocol
322 version is 2 or higher.
324 To determine the version, in any connection where the "renegotiation"
325 handshake was used (that is, where the server sent only one certificate
326 at first and where the client did not send any certificates until
327 renegotiation), both parties MUST send a VERSIONS cell immediately after
328 the renegotiation is finished, before any other cells are sent. Parties
329 MUST NOT send any other cells on a connection until they have received a
332 The payload in a VERSIONS cell is a series of big-endian two-byte
333 integers. Both parties MUST select as the link protocol version the
334 highest number contained both in the VERSIONS cell they sent and in the
335 versions cell they received. If they have no such version in common,
336 they cannot communicate and MUST close the connection.
338 Since the version 1 link protocol does not use the "renegotiation"
339 handshake, implementations MUST NOT list version 1 in their VERSIONS
344 If version 2 or higher is negotiated, each party sends the other a
345 NETINFO cell. The cell's payload is:
348 Other OR's address [variable]
349 Number of addresses [1 byte]
350 This OR's addresses [variable]
352 The address format is a type/length/value sequence as given in section
353 6.4 below. The timestamp is a big-endian unsigned integer number of
354 seconds since the unix epoch.
356 Implementations MAY use the timestamp value to help decide if their
357 clocks are skewed. Initiators MAY use "other OR's address" to help
358 learn which address their connections are originating from, if they do
359 not know it. Initiators SHOULD use "this OR's address" to make sure
360 that they have connected to another OR at its canonical address.
362 [As of 0.2.0.23-rc, implementations use none of the above values.]
365 5. Circuit management
367 5.1. CREATE and CREATED cells
369 Users set up circuits incrementally, one hop at a time. To create a
370 new circuit, OPs send a CREATE cell to the first node, with the
371 first half of the DH handshake; that node responds with a CREATED
372 cell with the second half of the DH handshake plus the first 20 bytes
373 of derivative key data (see section 5.2). To extend a circuit past
374 the first hop, the OP sends an EXTEND relay cell (see section 5)
375 which instructs the last node in the circuit to send a CREATE cell
376 to extend the circuit.
378 The payload for a CREATE cell is an 'onion skin', which consists
379 of the first step of the DH handshake data (also known as g^x).
380 This value is hybrid-encrypted (see 0.3) to Bob's onion key, giving
383 Padding [PK_PAD_LEN bytes]
384 Symmetric key [KEY_LEN bytes]
385 First part of g^x [PK_ENC_LEN-PK_PAD_LEN-KEY_LEN bytes]
386 Symmetrically encrypted:
387 Second part of g^x [DH_LEN-(PK_ENC_LEN-PK_PAD_LEN-KEY_LEN)
390 The relay payload for an EXTEND relay cell consists of:
393 Onion skin [DH_LEN+KEY_LEN+PK_PAD_LEN bytes]
394 Identity fingerprint [HASH_LEN bytes]
396 The port and address field denote the IPV4 address and port of the next
397 onion router in the circuit; the public key hash is the hash of the PKCS#1
398 ASN1 encoding of the next onion router's identity (signing) key. (See 0.3
399 above.) Including this hash allows the extending OR verify that it is
400 indeed connected to the correct target OR, and prevents certain
401 man-in-the-middle attacks.
403 The payload for a CREATED cell, or the relay payload for an
404 EXTENDED cell, contains:
405 DH data (g^y) [DH_LEN bytes]
406 Derivative key data (KH) [HASH_LEN bytes] <see 5.2 below>
408 The CircID for a CREATE cell is an arbitrarily chosen 2-byte integer,
409 selected by the node (OP or OR) that sends the CREATE cell. To prevent
410 CircID collisions, when one node sends a CREATE cell to another, it chooses
411 from only one half of the possible values based on the ORs' public
412 identity keys: if the sending node has a lower key, it chooses a CircID with
413 an MSB of 0; otherwise, it chooses a CircID with an MSB of 1.
415 (An OP with no public key MAY choose any CircID it wishes, since an OP
416 never needs to process a CREATE cell.)
418 Public keys are compared numerically by modulus.
420 As usual with DH, x and y MUST be generated randomly.
422 5.1.1. CREATE_FAST/CREATED_FAST cells
424 When initializing the first hop of a circuit, the OP has already
425 established the OR's identity and negotiated a secret key using TLS.
426 Because of this, it is not always necessary for the OP to perform the
427 public key operations to create a circuit. In this case, the
428 OP MAY send a CREATE_FAST cell instead of a CREATE cell for the first
429 hop only. The OR responds with a CREATED_FAST cell, and the circuit is
432 A CREATE_FAST cell contains:
434 Key material (X) [HASH_LEN bytes]
436 A CREATED_FAST cell contains:
438 Key material (Y) [HASH_LEN bytes]
439 Derivative key data [HASH_LEN bytes] (See 5.2 below)
441 The values of X and Y must be generated randomly.
443 If an OR sees a circuit created with CREATE_FAST, the OR is sure to be the
444 first hop of a circuit. ORs SHOULD reject attempts to create streams with
445 RELAY_BEGIN exiting the circuit at the first hop: letting Tor be used as a
446 single hop proxy makes exit nodes a more attractive target for compromise.
448 5.2. Setting circuit keys
450 Once the handshake between the OP and an OR is completed, both can
451 now calculate g^xy with ordinary DH. Before computing g^xy, both client
452 and server MUST verify that the received g^x or g^y value is not degenerate;
453 that is, it must be strictly greater than 1 and strictly less than p-1
454 where p is the DH modulus. Implementations MUST NOT complete a handshake
455 with degenerate keys. Implementations MUST NOT discard other "weak"
458 (Discarding degenerate keys is critical for security; if bad keys
459 are not discarded, an attacker can substitute the server's CREATED
460 cell's g^y with 0 or 1, thus creating a known g^xy and impersonating
461 the server. Discarding other keys may allow attacks to learn bits of
464 If CREATE or EXTEND is used to extend a circuit, the client and server
465 base their key material on K0=g^xy, represented as a big-endian unsigned
468 If CREATE_FAST is used, the client and server base their key material on
471 From the base key material K0, they compute KEY_LEN*2+HASH_LEN*3 bytes of
472 derivative key data as
473 K = H(K0 | [00]) | H(K0 | [01]) | H(K0 | [02]) | ...
475 The first HASH_LEN bytes of K form KH; the next HASH_LEN form the forward
476 digest Df; the next HASH_LEN 41-60 form the backward digest Db; the next
477 KEY_LEN 61-76 form Kf, and the final KEY_LEN form Kb. Excess bytes from K
480 KH is used in the handshake response to demonstrate knowledge of the
481 computed shared key. Df is used to seed the integrity-checking hash
482 for the stream of data going from the OP to the OR, and Db seeds the
483 integrity-checking hash for the data stream from the OR to the OP. Kf
484 is used to encrypt the stream of data going from the OP to the OR, and
485 Kb is used to encrypt the stream of data going from the OR to the OP.
487 5.3. Creating circuits
489 When creating a circuit through the network, the circuit creator
490 (OP) performs the following steps:
492 1. Choose an onion router as an exit node (R_N), such that the onion
493 router's exit policy includes at least one pending stream that
494 needs a circuit (if there are any).
496 2. Choose a chain of (N-1) onion routers
497 (R_1...R_N-1) to constitute the path, such that no router
498 appears in the path twice.
500 3. If not already connected to the first router in the chain,
501 open a new connection to that router.
503 4. Choose a circID not already in use on the connection with the
504 first router in the chain; send a CREATE cell along the
505 connection, to be received by the first onion router.
507 5. Wait until a CREATED cell is received; finish the handshake
508 and extract the forward key Kf_1 and the backward key Kb_1.
510 6. For each subsequent onion router R (R_2 through R_N), extend
513 To extend the circuit by a single onion router R_M, the OP performs
516 1. Create an onion skin, encrypted to R_M's public onion key.
518 2. Send the onion skin in a relay EXTEND cell along
519 the circuit (see section 5).
521 3. When a relay EXTENDED cell is received, verify KH, and
522 calculate the shared keys. The circuit is now extended.
524 When an onion router receives an EXTEND relay cell, it sends a CREATE
525 cell to the next onion router, with the enclosed onion skin as its
526 payload. As special cases, if the extend cell includes a digest of
527 all zeroes, or asks to extend back to the relay that sent the extend
528 cell, the circuit will fail and be torn down. The initiating onion
529 router chooses some circID not yet used on the connection between the
530 two onion routers. (But see section 5.1. above, concerning choosing
531 circIDs based on lexicographic order of nicknames.)
533 When an onion router receives a CREATE cell, if it already has a
534 circuit on the given connection with the given circID, it drops the
535 cell. Otherwise, after receiving the CREATE cell, it completes the
536 DH handshake, and replies with a CREATED cell. Upon receiving a
537 CREATED cell, an onion router packs it payload into an EXTENDED relay
538 cell (see section 5), and sends that cell up the circuit. Upon
539 receiving the EXTENDED relay cell, the OP can retrieve g^y.
541 (As an optimization, OR implementations may delay processing onions
542 until a break in traffic allows time to do so without harming
543 network latency too greatly.)
545 5.3.1. Canonical connections
547 It is possible for an attacker to launch a man-in-the-middle attack
548 against a connection by telling OR Alice to extend to OR Bob at some
549 address X controlled by the attacker. The attacker cannot read the
550 encrypted traffic, but the attacker is now in a position to count all
551 bytes sent between Alice and Bob (assuming Alice was not already
554 To prevent this, when an OR we gets an extend request, it SHOULD use an
555 existing OR connection if the ID matches, and ANY of the following
557 - The IP matches the requested IP.
558 - The OR knows that the IP of the connection it's using is canonical
559 because it was listed in the NETINFO cell.
560 - The OR knows that the IP of the connection it's using is canonical
561 because it was listed in the server descriptor.
563 [This is not implemented in Tor 0.2.0.23-rc.]
565 5.4. Tearing down circuits
567 Circuits are torn down when an unrecoverable error occurs along
568 the circuit, or when all streams on a circuit are closed and the
569 circuit's intended lifetime is over. Circuits may be torn down
570 either completely or hop-by-hop.
572 To tear down a circuit completely, an OR or OP sends a DESTROY
573 cell to the adjacent nodes on that circuit, using the appropriate
576 Upon receiving an outgoing DESTROY cell, an OR frees resources
577 associated with the corresponding circuit. If it's not the end of
578 the circuit, it sends a DESTROY cell for that circuit to the next OR
579 in the circuit. If the node is the end of the circuit, then it tears
580 down any associated edge connections (see section 6.1).
582 After a DESTROY cell has been processed, an OR ignores all data or
583 destroy cells for the corresponding circuit.
585 To tear down part of a circuit, the OP may send a RELAY_TRUNCATE cell
586 signaling a given OR (Stream ID zero). That OR sends a DESTROY
587 cell to the next node in the circuit, and replies to the OP with a
588 RELAY_TRUNCATED cell.
590 When an unrecoverable error occurs along one connection in a
591 circuit, the nodes on either side of the connection should, if they
592 are able, act as follows: the node closer to the OP should send a
593 RELAY_TRUNCATED cell towards the OP; the node farther from the OP
594 should send a DESTROY cell down the circuit.
596 The payload of a RELAY_TRUNCATED or DESTROY cell contains a single octet,
597 describing why the circuit is being closed or truncated. When sending a
598 TRUNCATED or DESTROY cell because of another TRUNCATED or DESTROY cell,
599 the error code should be propagated. The origin of a circuit always sets
600 this error code to 0, to avoid leaking its version.
603 0 -- NONE (No reason given.)
604 1 -- PROTOCOL (Tor protocol violation.)
605 2 -- INTERNAL (Internal error.)
606 3 -- REQUESTED (A client sent a TRUNCATE command.)
607 4 -- HIBERNATING (Not currently operating; trying to save bandwidth.)
608 5 -- RESOURCELIMIT (Out of memory, sockets, or circuit IDs.)
609 6 -- CONNECTFAILED (Unable to reach server.)
610 7 -- OR_IDENTITY (Connected to server, but its OR identity was not
612 8 -- OR_CONN_CLOSED (The OR connection that was carrying this circuit
614 9 -- FINISHED (The circuit has expired for being dirty or old.)
615 10 -- TIMEOUT (Circuit construction took too long)
616 11 -- DESTROYED (The circuit was destroyed w/o client TRUNCATE)
617 12 -- NOSUCHSERVICE (Request for unknown hidden service)
619 5.5. Routing relay cells
621 When an OR receives a RELAY cell, it checks the cell's circID and
622 determines whether it has a corresponding circuit along that
623 connection. If not, the OR drops the RELAY cell.
625 Otherwise, if the OR is not at the OP edge of the circuit (that is,
626 either an 'exit node' or a non-edge node), it de/encrypts the payload
627 with the stream cipher, as follows:
628 'Forward' relay cell (same direction as CREATE):
629 Use Kf as key; decrypt.
630 'Back' relay cell (opposite direction from CREATE):
631 Use Kb as key; encrypt.
632 Note that in counter mode, decrypt and encrypt are the same operation.
634 The OR then decides whether it recognizes the relay cell, by
635 inspecting the payload as described in section 6.1 below. If the OR
636 recognizes the cell, it processes the contents of the relay cell.
637 Otherwise, it passes the decrypted relay cell along the circuit if
638 the circuit continues. If the OR at the end of the circuit
639 encounters an unrecognized relay cell, an error has occurred: the OR
640 sends a DESTROY cell to tear down the circuit.
642 When a relay cell arrives at an OP, the OP decrypts the payload
643 with the stream cipher as follows:
644 OP receives data cell:
646 Decrypt with Kb_I. If the payload is recognized (see
647 section 6..1), then stop and process the payload.
649 For more information, see section 6 below.
651 6. Application connections and stream management
655 Within a circuit, the OP and the exit node use the contents of
656 RELAY packets to tunnel end-to-end commands and TCP connections
657 ("Streams") across circuits. End-to-end commands can be initiated
658 by either edge; streams are initiated by the OP.
660 The payload of each unencrypted RELAY cell consists of:
661 Relay command [1 byte]
662 'Recognized' [2 bytes]
666 Data [CELL_LEN-14 bytes]
668 The relay commands are:
669 1 -- RELAY_BEGIN [forward]
670 2 -- RELAY_DATA [forward or backward]
671 3 -- RELAY_END [forward or backward]
672 4 -- RELAY_CONNECTED [backward]
673 5 -- RELAY_SENDME [forward or backward] [sometimes control]
674 6 -- RELAY_EXTEND [forward] [control]
675 7 -- RELAY_EXTENDED [backward] [control]
676 8 -- RELAY_TRUNCATE [forward] [control]
677 9 -- RELAY_TRUNCATED [backward] [control]
678 10 -- RELAY_DROP [forward or backward] [control]
679 11 -- RELAY_RESOLVE [forward]
680 12 -- RELAY_RESOLVED [backward]
681 13 -- RELAY_BEGIN_DIR [forward]
683 32..40 -- Used for hidden services; see rend-spec.txt.
685 Commands labelled as "forward" must only be sent by the originator
686 of the circuit. Commands labelled as "backward" must only be sent by
687 other nodes in the circuit back to the originator. Commands marked
688 as either can be sent either by the originator or other nodes.
690 The 'recognized' field in any unencrypted relay payload is always set
691 to zero; the 'digest' field is computed as the first four bytes of
692 the running digest of all the bytes that have been destined for
693 this hop of the circuit or originated from this hop of the circuit,
694 seeded from Df or Db respectively (obtained in section 5.2 above),
695 and including this RELAY cell's entire payload (taken with the digest
698 When the 'recognized' field of a RELAY cell is zero, and the digest
699 is correct, the cell is considered "recognized" for the purposes of
700 decryption (see section 5.5 above).
702 (The digest does not include any bytes from relay cells that do
703 not start or end at this hop of the circuit. That is, it does not
704 include forwarded data. Therefore if 'recognized' is zero but the
705 digest does not match, the running digest at that node should
706 not be updated, and the cell should be forwarded on.)
708 All RELAY cells pertaining to the same tunneled stream have the
709 same stream ID. StreamIDs are chosen arbitrarily by the OP. RELAY
710 cells that affect the entire circuit rather than a particular
711 stream use a StreamID of zero -- they are marked in the table above
712 as "[control]" style cells. (Sendme cells are marked as "sometimes
713 control" because they can take include a StreamID or not depending
714 on their purpose -- see Section 7.)
716 The 'Length' field of a relay cell contains the number of bytes in
717 the relay payload which contain real payload data. The remainder of
718 the payload is padded with NUL bytes.
720 If the RELAY cell is recognized but the relay command is not
721 understood, the cell must be dropped and ignored. Its contents
722 still count with respect to the digests, though.
724 6.2. Opening streams and transferring data
726 To open a new anonymized TCP connection, the OP chooses an open
727 circuit to an exit that may be able to connect to the destination
728 address, selects an arbitrary StreamID not yet used on that circuit,
729 and constructs a RELAY_BEGIN cell with a payload encoding the address
730 and port of the destination host. The payload format is:
732 ADDRESS | ':' | PORT | [00]
734 where ADDRESS can be a DNS hostname, or an IPv4 address in
735 dotted-quad format, or an IPv6 address surrounded by square brackets;
736 and where PORT is a decimal integer between 1 and 65535, inclusive.
738 [What is the [00] for? -NM]
739 [It's so the payload is easy to parse out with string funcs -RD]
741 Upon receiving this cell, the exit node resolves the address as
742 necessary, and opens a new TCP connection to the target port. If the
743 address cannot be resolved, or a connection can't be established, the
744 exit node replies with a RELAY_END cell. (See 6.4 below.)
745 Otherwise, the exit node replies with a RELAY_CONNECTED cell, whose
746 payload is in one of the following formats:
747 The IPv4 address to which the connection was made [4 octets]
748 A number of seconds (TTL) for which the address may be cached [4 octets]
750 Four zero-valued octets [4 octets]
751 An address type (6) [1 octet]
752 The IPv6 address to which the connection was made [16 octets]
753 A number of seconds (TTL) for which the address may be cached [4 octets]
754 [XXXX No version of Tor currently generates the IPv6 format.]
756 [Tor servers before 0.1.2.0 set the TTL field to a fixed value. Later
757 versions set the TTL to the last value seen from a DNS server, and expire
758 their own cached entries after a fixed interval. This prevents certain
761 The OP waits for a RELAY_CONNECTED cell before sending any data.
762 Once a connection has been established, the OP and exit node
763 package stream data in RELAY_DATA cells, and upon receiving such
764 cells, echo their contents to the corresponding TCP stream.
765 RELAY_DATA cells sent to unrecognized streams are dropped.
767 Relay RELAY_DROP cells are long-range dummies; upon receiving such
768 a cell, the OR or OP must drop it.
770 6.2.1. Opening a directory stream
772 If a Tor server is a directory server, it should respond to a
773 RELAY_BEGIN_DIR cell as if it had received a BEGIN cell requesting a
774 connection to its directory port. RELAY_BEGIN_DIR cells ignore exit
775 policy, since the stream is local to the Tor process.
777 If the Tor server is not running a directory service, it should respond
778 with a REASON_NOTDIRECTORY RELAY_END cell.
780 Clients MUST generate an all-zero payload for RELAY_BEGIN_DIR cells,
781 and servers MUST ignore the payload.
783 [RELAY_BEGIN_DIR was not supported before Tor 0.1.2.2-alpha; clients
784 SHOULD NOT send it to routers running earlier versions of Tor.]
788 When an anonymized TCP connection is closed, or an edge node
789 encounters error on any stream, it sends a 'RELAY_END' cell along the
790 circuit (if possible) and closes the TCP connection immediately. If
791 an edge node receives a 'RELAY_END' cell for any stream, it closes
792 the TCP connection completely, and sends nothing more along the
793 circuit for that stream.
795 The payload of a RELAY_END cell begins with a single 'reason' byte to
796 describe why the stream is closing, plus optional data (depending on
797 the reason.) The values are:
799 1 -- REASON_MISC (catch-all for unlisted reasons)
800 2 -- REASON_RESOLVEFAILED (couldn't look up hostname)
801 3 -- REASON_CONNECTREFUSED (remote host refused connection) [*]
802 4 -- REASON_EXITPOLICY (OR refuses to connect to host or port)
803 5 -- REASON_DESTROY (Circuit is being destroyed)
804 6 -- REASON_DONE (Anonymized TCP connection was closed)
805 7 -- REASON_TIMEOUT (Connection timed out, or OR timed out
807 8 -- (unallocated) [**]
808 9 -- REASON_HIBERNATING (OR is temporarily hibernating)
809 10 -- REASON_INTERNAL (Internal error at the OR)
810 11 -- REASON_RESOURCELIMIT (OR has no resources to fulfill request)
811 12 -- REASON_CONNRESET (Connection was unexpectedly reset)
812 13 -- REASON_TORPROTOCOL (Sent when closing connection because of
813 Tor protocol violations.)
814 14 -- REASON_NOTDIRECTORY (Client sent RELAY_BEGIN_DIR to a
815 non-directory server.)
817 (With REASON_EXITPOLICY, the 4-byte IPv4 address or 16-byte IPv6 address
818 forms the optional data, along with a 4-byte TTL; no other reason
819 currently has extra data.)
821 OPs and ORs MUST accept reasons not on the above list, since future
822 versions of Tor may provide more fine-grained reasons.
824 [*] Older versions of Tor also send this reason when connections are
826 [**] Due to a bug in versions of Tor through 0095, error reason 8 must
827 remain allocated until that version is obsolete.
829 --- [The rest of this section describes unimplemented functionality.]
831 Because TCP connections can be half-open, we follow an equivalent
832 to TCP's FIN/FIN-ACK/ACK protocol to close streams.
834 An exit connection can have a TCP stream in one of three states:
835 'OPEN', 'DONE_PACKAGING', and 'DONE_DELIVERING'. For the purposes
836 of modeling transitions, we treat 'CLOSED' as a fourth state,
837 although connections in this state are not, in fact, tracked by the
840 A stream begins in the 'OPEN' state. Upon receiving a 'FIN' from
841 the corresponding TCP connection, the edge node sends a 'RELAY_FIN'
842 cell along the circuit and changes its state to 'DONE_PACKAGING'.
843 Upon receiving a 'RELAY_FIN' cell, an edge node sends a 'FIN' to
844 the corresponding TCP connection (e.g., by calling
845 shutdown(SHUT_WR)) and changing its state to 'DONE_DELIVERING'.
847 When a stream in already in 'DONE_DELIVERING' receives a 'FIN', it
848 also sends a 'RELAY_FIN' along the circuit, and changes its state
849 to 'CLOSED'. When a stream already in 'DONE_PACKAGING' receives a
850 'RELAY_FIN' cell, it sends a 'FIN' and changes its state to
853 If an edge node encounters an error on any stream, it sends a
854 'RELAY_END' cell (if possible) and closes the stream immediately.
856 6.4. Remote hostname lookup
858 To find the address associated with a hostname, the OP sends a
859 RELAY_RESOLVE cell containing the hostname to be resolved with a nul
860 terminating byte. (For a reverse lookup, the OP sends a RELAY_RESOLVE
861 cell containing an in-addr.arpa address.) The OR replies with a
862 RELAY_RESOLVED cell containing a status byte, and any number of
863 answers. Each answer is of the form:
866 Value (variable-width)
868 "Length" is the length of the Value field.
873 0xF0 -- Error, transient
874 0xF1 -- Error, nontransient
876 If any answer has a type of 'Error', then no other answer may be given.
878 The RELAY_RESOLVE cell must use a nonzero, distinct streamID; the
879 corresponding RELAY_RESOLVED cell must use the same streamID. No stream
880 is actually created by the OR when resolving the name.
886 Each client or relay should do appropriate bandwidth throttling to
889 Communicants rely on TCP's default flow control to push back when they
892 The mainline Tor implementation uses token buckets (one for reads,
893 one for writes) for the rate limiting.
895 Since 0.2.0.x, Tor has let the user specify an additional pair of
896 token buckets for "relayed" traffic, so people can deploy a Tor relay
897 with strict rate limiting, but also use the same Tor as a client. To
898 avoid partitioning concerns we combine both classes of traffic over a
899 given OR connection, and keep track of the last time we read or wrote
900 a high-priority (non-relayed) cell. If it's been less than N seconds
901 (currently N=30), we give the whole connection high priority, else we
902 give the whole connection low priority. We also give low priority
903 to reads and writes for connections that are serving directory
904 information. See proposal 111 for details.
908 Link padding can be created by sending PADDING cells along the
909 connection; relay cells of type "DROP" can be used for long-range
912 Currently nodes are not required to do any sort of link padding or
913 dummy traffic. Because strong attacks exist even with link padding,
914 and because link padding greatly increases the bandwidth requirements
915 for running a node, we plan to leave out link padding until this
916 tradeoff is better understood.
918 7.3. Circuit-level flow control
920 To control a circuit's bandwidth usage, each OR keeps track of two
921 'windows', consisting of how many RELAY_DATA cells it is allowed to
922 originate (package for transmission), and how many RELAY_DATA cells
923 it is willing to consume (receive for local streams). These limits
924 do not apply to cells that the OR receives from one host and relays
927 Each 'window' value is initially set to 1000 data cells
928 in each direction (cells that are not data cells do not affect
929 the window). When an OR is willing to deliver more cells, it sends a
930 RELAY_SENDME cell towards the OP, with Stream ID zero. When an OR
931 receives a RELAY_SENDME cell with stream ID zero, it increments its
934 Each of these cells increments the corresponding window by 100.
936 The OP behaves identically, except that it must track a packaging
937 window and a delivery window for every OR in the circuit.
939 An OR or OP sends cells to increment its delivery window when the
940 corresponding window value falls under some threshold (900).
942 If a packaging window reaches 0, the OR or OP stops reading from
943 TCP connections for all streams on the corresponding circuit, and
944 sends no more RELAY_DATA cells until receiving a RELAY_SENDME cell.
945 [this stuff is badly worded; copy in the tor-design section -RD]
947 7.4. Stream-level flow control
949 Edge nodes use RELAY_SENDME cells to implement end-to-end flow
950 control for individual connections across circuits. Similarly to
951 circuit-level flow control, edge nodes begin with a window of cells
952 (500) per stream, and increment the window by a fixed value (50)
953 upon receiving a RELAY_SENDME cell. Edge nodes initiate RELAY_SENDME
954 cells when both a) the window is <= 450, and b) there are less than
955 ten cell payloads remaining to be flushed at that edge.
957 A.1. Differences between spec and implementation
959 - The current specification requires all ORs to have IPv4 addresses, but
960 allows servers to exit and resolve to IPv6 addresses, and to declare IPv6
961 addresses in their exit policies. The current codebase has no IPv6