blob: a6ec3535eda1ebbe08af5e146ec426c586eb72c0 [file] [log] [blame]
Denis Vlasenko8ec6ee42007-11-23 21:43:40 +00001
2
3
4
5
6
7Network Working Group M. Allman
8Request for Comments: 2428 NASA Lewis/Sterling Software
9Category: Standards Track S. Ostermann
10 Ohio University
11 C. Metz
12 The Inner Net
13 September 1998
14
15
16 FTP Extensions for IPv6 and NATs
17
18Status of this Memo
19
20 This document specifies an Internet standards track protocol for the
21 Internet community, and requests discussion and suggestions for
22 improvements. Please refer to the current edition of the "Internet
23 Official Protocol Standards" (STD 1) for the standardization state
24 and status of this protocol. Distribution of this memo is unlimited.
25
26Copyright Notice
27
28 Copyright (C) The Internet Society (1998). All Rights Reserved.
29
30Abstract
31
32 The specification for the File Transfer Protocol assumes that the
33 underlying network protocol uses a 32-bit network address
34 (specifically IP version 4). With the deployment of version 6 of the
35 Internet Protocol, network addresses will no longer be 32-bits. This
36 paper specifies extensions to FTP that will allow the protocol to
37 work over IPv4 and IPv6. In addition, the framework defined can
38 support additional network protocols in the future.
39
401. Introduction
41
42 The keywords, such as MUST and SHOULD, found in this document are
43 used as defined in RFC 2119 [Bra97].
44
45 The File Transfer Protocol [PR85] only provides the ability to
46 communicate information about IPv4 data connections. FTP assumes
47 network addresses will be 32 bits in length. However, with the
48 deployment of version 6 of the Internet Protocol [DH96] addresses
49 will no longer be 32 bits long. RFC 1639 [Pis94] specifies
50 extensions to FTP to enable its use over various network protocols.
51 Unfortunately, the mechanism can fail in a multi-protocol
52 environment. During the transition between IPv4 and IPv6, FTP needs
53 the ability to negotiate the network protocol that will be used for
54 data transfer.
55
56
57
58Allman, et. al. Standards Track [Page 1]
59
60RFC 2428 FTP Extensions for IPv6 and NATs September 1998
61
62
63 This document provides a specification for a way that FTP can
64 communicate data connection endpoint information for network
65 protocols other than IPv4. In this specification, the FTP commands
66 PORT and PASV are replaced with EPRT and EPSV, respectively. This
67 document is organized as follows. Section 2 outlines the EPRT
68 command and Section 3 outlines the EPSV command. Section 4 defines
69 the utilization of these two new FTP commands. Section 5 briefly
70 presents security considerations. Finally, Section 6 provides
71 conclusions.
72
732. The EPRT Command
74
75 The EPRT command allows for the specification of an extended address
76 for the data connection. The extended address MUST consist of the
77 network protocol as well as the network and transport addresses. The
78 format of EPRT is:
79
80 EPRT<space><d><net-prt><d><net-addr><d><tcp-port><d>
81
82 The EPRT command keyword MUST be followed by a single space (ASCII
83 32). Following the space, a delimiter character (<d>) MUST be
84 specified. The delimiter character MUST be one of the ASCII
85 characters in range 33-126 inclusive. The character "|" (ASCII 124)
86 is recommended unless it coincides with a character needed to encode
87 the network address.
88
89 The <net-prt> argument MUST be an address family number defined by
90 IANA in the latest Assigned Numbers RFC (RFC 1700 [RP94] as of the
91 writing of this document). This number indicates the protocol to be
92 used (and, implicitly, the address length). This document will use
93 two of address family numbers from [RP94] as examples, according to
94 the following table:
95
96 AF Number Protocol
97 --------- --------
98 1 Internet Protocol, Version 4 [Pos81a]
99 2 Internet Protocol, Version 6 [DH96]
100
101 The <net-addr> is a protocol specific string representation of the
102 network address. For the two address families specified above (AF
103 Number 1 and 2), addresses MUST be in the following format:
104
105 AF Number Address Format Example
106 --------- -------------- -------
107 1 dotted decimal 132.235.1.2
108 2 IPv6 string 1080::8:800:200C:417A
109 representations
110 defined in [HD96]
111
112
113
114Allman, et. al. Standards Track [Page 2]
115
116RFC 2428 FTP Extensions for IPv6 and NATs September 1998
117
118
119 The <tcp-port> argument must be the string representation of the
120 number of the TCP port on which the host is listening for the data
121 connection.
122
123 The following are sample EPRT commands:
124
125 EPRT |1|132.235.1.2|6275|
126
127 EPRT |2|1080::8:800:200C:417A|5282|
128
129 The first command specifies that the server should use IPv4 to open a
130 data connection to the host "132.235.1.2" on TCP port 6275. The
131 second command specifies that the server should use the IPv6 network
132 protocol and the network address "1080::8:800:200C:417A" to open a
133 TCP data connection on port 5282.
134
135 Upon receipt of a valid EPRT command, the server MUST return a code
136 of 200 (Command OK). The standard negative error code 500 and 501
137 [PR85] are sufficient to handle most errors (e.g., syntax errors)
138 involving the EPRT command. However, an additional error code is
139 needed. The response code 522 indicates that the server does not
140 support the requested network protocol. The interpretation of this
141 new error code is:
142
143 5yz Negative Completion
144 x2z Connections
145 xy2 Extended Port Failure - unknown network protocol
146
147 The text portion of the response MUST indicate which network
148 protocols the server does support. If the network protocol is
149 unsupported, the format of the response string MUST be:
150
151 <text stating that the network protocol is unsupported> \
152 (prot1,prot2,...,protn)
153
154 Both the numeric code specified above and the protocol information
155 between the characters '(' and ')' are intended for the software
156 automata receiving the response; the textual message between the
157 numeric code and the '(' is intended for the human user and can be
158 any arbitrary text, but MUST NOT include the characters '(' and ')'.
159 In the above case, the text SHOULD indicate that the network protocol
160 in the EPRT command is not supported by the server. The list of
161 protocols inside the parenthesis MUST be a comma separated list of
162 address family numbers. Two example response strings follow:
163
164 Network protocol not supported, use (1)
165
166 Network protocol not supported, use (1,2)
167
168
169
170Allman, et. al. Standards Track [Page 3]
171
172RFC 2428 FTP Extensions for IPv6 and NATs September 1998
173
174
1753. The EPSV Command
176
177 The EPSV command requests that a server listen on a data port and
178 wait for a connection. The EPSV command takes an optional argument.
179 The response to this command includes only the TCP port number of the
180 listening connection. The format of the response, however, is
181 similar to the argument of the EPRT command. This allows the same
182 parsing routines to be used for both commands. In addition, the
183 format leaves a place holder for the network protocol and/or network
184 address, which may be needed in the EPSV response in the future. The
185 response code for entering passive mode using an extended address
186 MUST be 229. The interpretation of this code, according to [PR85]
187 is:
188
189 2yz Positive Completion
190 x2z Connections
191 xy9 Extended Passive Mode Entered
192
193 The text returned in response to the EPSV command MUST be:
194
195 <text indicating server is entering extended passive mode> \
196 (<d><d><d><tcp-port><d>)
197
198 The portion of the string enclosed in parentheses MUST be the exact
199 string needed by the EPRT command to open the data connection, as
200 specified above.
201
202 The first two fields contained in the parenthesis MUST be blank. The
203 third field MUST be the string representation of the TCP port number
204 on which the server is listening for a data connection. The network
205 protocol used by the data connection will be the same network
206 protocol used by the control connection. In addition, the network
207 address used to establish the data connection will be the same
208 network address used for the control connection. An example response
209 string follows:
210
211 Entering Extended Passive Mode (|||6446|)
212
213 The standard negative error codes 500 and 501 are sufficient to
214 handle all errors involving the EPSV command (e.g., syntax errors).
215
216 When the EPSV command is issued with no argument, the server will
217 choose the network protocol for the data connection based on the
218 protocol used for the control connection. However, in the case of
219 proxy FTP, this protocol might not be appropriate for communication
220 between the two servers. Therefore, the client needs to be able to
221 request a specific protocol. If the server returns a protocol that
222 is not supported by the host that will be connecting to the port, the
223
224
225
226Allman, et. al. Standards Track [Page 4]
227
228RFC 2428 FTP Extensions for IPv6 and NATs September 1998
229
230
231 client MUST issue an ABOR (abort) command to allow the server to
232 close down the listening connection. The client can then send an
233 EPSV command requesting the use of a specific network protocol, as
234 follows:
235
236 EPSV<space><net-prt>
237
238 If the requested protocol is supported by the server, it SHOULD use
239 the protocol. If not, the server MUST return the 522 error messages
240 as outlined in section 2.
241
242 Finally, the EPSV command can be used with the argument "ALL" to
243 inform Network Address Translators that the EPRT command (as well as
244 other data commands) will no longer be used. An example of this
245 command follows:
246
247 EPSV<space>ALL
248
249 Upon receipt of an EPSV ALL command, the server MUST reject all data
250 connection setup commands other than EPSV (i.e., EPRT, PORT, PASV, et
251 al.). This use of the EPSV command is further explained in section
252 4.
253
2544. Command Usage
255
256 For all FTP transfers where the control and data connection(s) are
257 being established between the same two machines, the EPSV command
258 MUST be used. Using the EPSV command benefits performance of
259 transfers that traverse firewalls or Network Address Translators
260 (NATs). RFC 1579 [Bel94] recommends using the passive command when
261 behind firewalls since firewalls do not generally allow incoming
262 connections (which are required when using the PORT (EPRT) command).
263 In addition, using EPSV as defined in this document does not require
264 NATs to change the network address in the traffic as it is forwarded.
265 The NAT would have to change the address if the EPRT command was
266 used. Finally, if the client issues an "EPSV ALL" command, NATs may
267 be able to put the connection on a "fast path" through the
268 translator, as the EPRT command will never be used and therefore,
269 translation of the data portion of the segments will never be needed.
270 When a client only expects to do two-way FTP transfers, it SHOULD
271 issue this command as soon as possible. If a client later finds that
272 it must do a three-way FTP transfer after issuing an EPSV ALL
273 command, a new FTP session MUST be started.
274
275
276
277
278
279
280
281
282Allman, et. al. Standards Track [Page 5]
283
284RFC 2428 FTP Extensions for IPv6 and NATs September 1998
285
286
2875. Security Issues
288
289 The authors do not believe that these changes to FTP introduce new
290 security problems. A companion Work in Progress [AO98] is a more
291 general discussion of FTP security issues and techniques to reduce
292 these security problems.
293
2946. Conclusions
295
296 The extensions specified in this paper will enable FTP to operate
297 over a variety of network protocols.
298
299References
300
301 [AO98] Allman, M., and S. Ostermann, "FTP Security
302 Considerations", Work in Progress.
303
304 [Bel94] Bellovin, S., "Firewall-Friendly FTP", RFC 1579, February
305 1994.
306
307 [Bra97] Bradner, S., "Key words for use in RFCs to Indicate
308 Requirement Levels", BCP 14, RFC 2119, March 1997.
309
310 [DH96] Deering, S., and R. Hinden, "Internet Protocol, Version 6
311 (IPv6) Specification", RFC 1883, December 1995.
312
313 [HD96] Hinden, R., and S. Deering, "IP Version 6 Addressing
314 Architecture", RFC 2373, July 1998.
315
316 [Pis94] Piscitello, D., "FTP Operation Over Big Address Records
317 (FOOBAR)", RFC 1639, June 1994.
318
319 [Pos81a] Postel, J., "Internet Protocol", STD 5, RFC 791, September
320 1981.
321
322 [Pos81b] Postel, J., "Transmission Control Protocol", STD 7, RFC 793,
323 September 1981.
324
325 [PR85] Postel, J., and J. Reynolds, "File Transfer Protocol (FTP)",
326 STD 9, RFC 959, October 1985.
327
328 [RP94] Reynolds, J., and J. Postel, "Assigned Numbers", STD 2, RFC
329 1700, October 1994. See also:
330 http://www.iana.org/numbers.html
331
332
333
334
335
336
337
338Allman, et. al. Standards Track [Page 6]
339
340RFC 2428 FTP Extensions for IPv6 and NATs September 1998
341
342
343Authors' Addresses
344
345 Mark Allman
346 NASA Lewis Research Center/Sterling Software
347 21000 Brookpark Rd. MS 54-2
348 Cleveland, OH 44135
349
350 Phone: (216) 433-6586
351 EMail: mallman@lerc.nasa.gov
352 http://gigahertz.lerc.nasa.gov/~mallman/
353
354
355 Shawn Ostermann
356 School of Electrical Engineering and Computer Science
357 Ohio University
358 416 Morton Hall
359 Athens, OH 45701
360
361 Phone: (740) 593-1234
362 EMail: ostermann@cs.ohiou.edu
363
364
365 Craig Metz
366 The Inner Net
367 Box 10314-1954
368 Blacksburg, VA 24062-0314
369
370 Phone: (DSN) 754-8590
371 EMail: cmetz@inner.net
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394Allman, et. al. Standards Track [Page 7]
395
396RFC 2428 FTP Extensions for IPv6 and NATs September 1998
397
398
399Full Copyright Statement
400
401 Copyright (C) The Internet Society (1998). All Rights Reserved.
402
403 This document and translations of it may be copied and furnished to
404 others, and derivative works that comment on or otherwise explain it
405 or assist in its implementation may be prepared, copied, published
406 and distributed, in whole or in part, without restriction of any
407 kind, provided that the above copyright notice and this paragraph are
408 included on all such copies and derivative works. However, this
409 document itself may not be modified in any way, such as by removing
410 the copyright notice or references to the Internet Society or other
411 Internet organizations, except as needed for the purpose of
412 developing Internet standards in which case the procedures for
413 copyrights defined in the Internet Standards process must be
414 followed, or as required to translate it into languages other than
415 English.
416
417 The limited permissions granted above are perpetual and will not be
418 revoked by the Internet Society or its successors or assigns.
419
420 This document and the information contained herein is provided on an
421 "AS IS" basis and THE INTERNET SOCIETY AND THE INTERNET ENGINEERING
422 TASK FORCE DISCLAIMS ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING
423 BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION
424 HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF
425 MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450Allman, et. al. Standards Track [Page 8]
451