File Transfer Protocol RANG Command for Byte Ranges
draft-bryan-ftp-range-00

Abstract

The File Transfer Protocol offers REST, to designate a starting point for a transfer, but does not currently offer any method to specify an end point. This document specifies a new FTP RANG command to be used by clients to designate a start and end point for STREAM mode transfers with RETR.

Status of this Memo

This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.

Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at http://datatracker.ietf.org/drafts/current/.

Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as “work in progress.”

This Internet-Draft will expire on June 23, 2011.

Copyright Notice

Copyright (c) 2010 IETF Trust and the persons identified as the document authors. All rights reserved.

This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (http://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License.

Table of Contents

1.  Introduction
2.  Document Conventions
    2.1.  Basic Tokens
    2.2.  Server Replies
3.  Byte Ranges in STREAM Mode
4.  The RANGe Command (RANG)
    4.1.  FEAT Command Response for RANG Command
    4.2.  User-PI usage of RANG
    4.3.  RANG Command Errors
5.  IANA Considerations
6.  Security Considerations
7.  References
    7.1.  Normative References
    7.2.  Informative References
Appendix A.  Acknowledgements and Contributors
Appendix B.  Document History
§  Authors' Addresses

1.  Introduction

The File Transfer Protocol offers REST [RFC3659] (Hethmon, P., “Extensions to FTP,” March 2007.), to designate a starting point for a transfer, but does not currently offer any method to specify an end point. This document specifies a new FTP RANG command to be used by clients to designate a start and end point for STREAM mode transfers with RETR.

The current alternatives, without being able to specify an end point, are to issue an ABOR command or close the data connection.

HTTP offers similar functionality with the Range: header in Section 14.35 of [RFC2616] (Fielding, R., Gettys, J., Mogul, J., Frystyk, H., Masinter, L., Leach, P., and T. Berners-Lee, “Hypertext Transfer Protocol -- HTTP/1.1,” June 1999.), where a specific byte range can optionally be requested.

[[ Discussion of this draft should take place on the FTPEXT2 Working Group mailing list, although this draft is not a WG item: ftpext@ietf.org. ]]

2.  Document Conventions

This specification describes conformance of File Transfer Protocol Extension for RANG, and end point in a byte range.

The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in BCP 14, [RFC2119] (Bradner, S., “Key words for use in RFCs to Indicate Requirement Levels,” March 1997.), as scoped to those conformance targets.

This document also uses notation defined in STD 9, [RFC0959] (Postel, J. and J. Reynolds, “File Transfer Protocol,” October 1985.). In particular, the terms or commands "reply", "user", "file", "FTP commands", "user-PI" (user protocol interpreter), "server-FTP process", "server-PI", "mode", "Image type", "Stream transfer mode", "type", "STOR", "RETR", and "ASCII", are all used here as defined there. The command "REST" is used as defined in Section 5 of [RFC3659] (Hethmon, P., “Extensions to FTP,” March 2007.).

In the examples of FTP dialogs presented in this document, lines that begin "C> " were sent over the control connection from the user-PI to the server-PI, and lines that begin "S> " were sent over the control connection from the server-PI to the user-PI. In all cases, the prefixes shown above, including the one space, have been added for the purposes of this document, and are not a part of the data exchanged between client and server.

Syntax required is defined using the Augmented BNF defined in [RFC5234] (Crocker, D. and P. Overell, “Augmented BNF for Syntax Specifications: ABNF,” January 2008.).

2.1.  Basic Tokens

This document imports the core definitions given in Appendix B of [RFC5234] (Crocker, D. and P. Overell, “Augmented BNF for Syntax Specifications: ABNF,” January 2008.). There definitions will be found for basic ABNF elements like ALPHA, DIGIT, SP, etc. To that, the following term is added for use in this document.

   TCHAR = VCHAR / SP / HTAB    ; visible plus white space

The VCHAR (from [RFC5234] (Crocker, D. and P. Overell, “Augmented BNF for Syntax Specifications: ABNF,” January 2008.)) and TCHAR rules give basic character types from varying sub-sets of the ASCII character set for use in various commands and responses.

Note that in ABNF, string literals are case insensitive. That convention is preserved in this document, and implies that FTP commands and parameters that are added by this specification have values that can be represented in any case. That is, "RANG" is the same as "rang", "Rang", "RaNg", etc., and "ftp.example.com" is the same as "Ftp.Example.Com", "fTp.eXample.cOm", etc.

2.2.  Server Replies

Section 4.2 of [RFC0959] (Postel, J. and J. Reynolds, “File Transfer Protocol,” October 1985.) defines the format and meaning of replies by the server-PI to FTP commands from the user-PI. Those reply conventions are used here without change.

   error-response = error-code SP *TCHAR CRLF
   error-code     = ("4" / "5") 2DIGIT

Implementers should note that the ABNF syntax (which was not used in [RFC0959] (Postel, J. and J. Reynolds, “File Transfer Protocol,” October 1985.)) used in this document, and other FTP related documents, sometimes shows replies using the one line format. Unless otherwise explicitly stated, that is not intended to imply that multi-line responses are not permitted. Implementers should assume that, unless stated to the contrary, any reply to any FTP command (including QUIT) can be of the multi-line format described in [RFC0959] (Postel, J. and J. Reynolds, “File Transfer Protocol,” October 1985.).

Throughout this document, replies will be identified by the three digit code that is their first element. Thus the term "500 reply" means a reply from the server-PI using the three digit code "500".

3.  Byte Ranges in STREAM Mode

In STREAM mode, the data connection contains just a stream of unformatted octets of data. Explicit restart markers thus cannot be inserted into the data stream, they would be indistinguishable from data. For this reason, the FTP specification [RFC0959] (Postel, J. and J. Reynolds, “File Transfer Protocol,” October 1985.) did not provide the ability to do restarts in stream mode. However, there is not really a need to have explicit restart markers in this case, as restart markers can be implied by the octet offset into the data stream.

Because the data stream defines the file in STREAM mode, a different data stream would represent a different file. Thus, an offset will always represent the same position within a file. On the other hand, in other modes than STREAM, the same file can be transferred using quite different octet sequences and yet be reconstructed into the one identical file. Thus an offset into the data stream in transfer modes other than STREAM would not give an unambiguous restart or end point.

If the data representation TYPE is IMAGE and the STRUcture is File, for many systems the file will be stored exactly in the same format as it is sent across the data connection. It is then usually very easy for the receiver to determine how much data was previously received, and notify the sender of the offset where the transfer should be restarted. In other representation types and structures more effort will be required, but it remains always possible to determine the offset with finite, but perhaps non-negligible, effort. In the worst case, an FTP process may need to open a data connection to itself, set the appropriate transfer type and structure, and actually transmit the file, counting the transmitted octets.

If the user-FTP process is intending to restart a retrieve, it will directly calculate the restart marker and send that information in the RESTart command. However, if the user-FTP process is intending to restart sending the file, it needs to be able to determine how much data was previously sent, and correctly received and saved. The purpose of the SIZE command, as documented in [RFC3659] (Hethmon, P., “Extensions to FTP,” March 2007.), is to get this information.

4.  The RANGe Command (RANG)

A new command "RANG" is added to the FTP command set to allow the client to specify both an start point byte range and an end point byte range of a file from a server-FTP process.

The syntax for the RANG command when the current transfer mode is STREAM is:

   range-command = "RANG" SP start-point  SP end-point CRLF
   start-point = 1*DIGIT
   end-point = 1*DIGIT

[NOTE: end-point is inclusive.]

start-point gives the number of octets of the immediately-following transfer to not actually send, effectively causing the transmission to be started at a later point. A value of zero effectively causes the transmission to be started at first byte. end-point gives the number of octets, counted from the beginning of the file, of the immediately-following transfer to stop sending at, effectively causing the transmission to be ended. (That is, the end point is relative to the start of the file and not relative to the start point). The server-PI will response to the RANG command with a 350 reply, indicating that RANG parameters has been saved, and that another command, which should be RETR, should then follow to complete the ranged request.

   range-response = range-ok / error-response
   range-ok       = "350" SP *TCHAR CRLF

The RANG command is intended to get partial data of a file. When the RANG command is used with RETR, client bears the responsibility of merging the retrieved data with partially retrieved file. It may choose to use the data obtained other than to complete an earlier transfer, or to re-retrieve data that had been retrieved before.

The RANG command must be the last command issued before the data transfer command that is to cause a partial data transfer. The effect of issuing a RANG command at any other time is undefined. The server-PI may react to a badly positioned RANG command by issuing an error response to the following command, not being a restartable data transfer command, or it may save the start-point and/or end-point byte range value and apply it to the next data transfer command, or it may silently ignore the inappropriate restart attempt. Because of this, a user-PI that has issued a RANG command, but that has not successfully transmitted the following data transfer command for any reason, should send another RANG command before the next data transfer command. If that transfer is not to be restarted, then "REST 0" should be issued.

An error response will follow a RANG command only when the server does not implement the command, or when command syntax is invalid. Any other errors, including such problems as start-point and/or end-point byte range out of range, should be reported when the following transfer command is issued. Such errors will cause that transfer request to be rejected with an error indicating the invalid restart attempt.

4.1.  FEAT Command Response for RANG Command

When replying to the FEAT command [RFC2389] (Hethmon, P. and R. Elz, “Feature negotiation mechanism for the File Transfer Protocol,” August 1998.), a server-FTP process that supports the RANG command, as specified here, MUST include, a line containing exactly the string "RANG STREAM". This string is case insensitive, and MAY be sent in any mixture of upper or lower case, however it SHOULD be sent in upper case. That is, the response SHOULD be:

   C> FEAT
   S> 211-Extensions supported:
   S>  ...
   S>  RANG STREAM
   S>  ...
   S> 211 END

The ellipses indicate place holders where other features may be included, and are not required. The one-space indentation of the feature lines is mandatory [RFC2389] (Hethmon, P. and R. Elz, “Feature negotiation mechanism for the File Transfer Protocol,” August 1998.).

   range-feat = SP "RANG" SP "STREAM"

4.2.  User-PI usage of RANG

The user-PI issues the FEAT command to query the server-PI if it supports the RANG command. In this example, the server-PI also supports REST.

   C> FEAT
   S> 211-Extensions supported:
   S>  ...
   S>  REST STREAM
   S>  RANG STREAM
   S>  ...
   S> 211 END

Assume that the transfer of a largish file has previously been interrupted after 802816 octets had been received, that the tranfer should stop at octet 1000000 of the file, that the previous transfer was with TYPE=I, and that it has been verified that the file on the server has not since changed.

   C> TYPE I
   S> 200 Type set to I.
   C> PORT 127,0,0,1,15,107
   S> 200 PORT command successful.
   C> RANG 802816 1000000
   S> 350 Restarting at 802816. End Byte range at 1000000. Send RETRIEVE
   C> RETR cap60.pl198.tar
   S> 150 Opening BINARY mode data connection
   [...]
   S> 226 Transfer complete.

In the above example, data is sent from offset 802816 to, and including, offset 1000000.

4.3.  RANG Command Errors

The server-PI SHOULD reply with a 500 reply if the RANG command is unrecognized or unimplemented.

The server-PI SHOULD reply with a 550 reply if the RANG command is used on a file that can not be found.

The server-PI SHOULD reply with a 551 reply if the server-PI is unable to deliver the file with TYPE I and MODE S.

The server-PI SHOULD reply with a 552 reply if the user is not allowed to use the RANG command.

The server-PI SHOULD reply with a 553 reply if the user requests the RANG of a directory, which is not allowed.

The server-PI SHOULD reply with a 5yz reply if the specified end point is larger than the actual file or the end point is before the start point.

5.  IANA Considerations

This new command is added to the "FTP Commands and Extensions" registry created by [RFC5797] (Klensin, J. and A. Hoenes, “FTP Command and Extension Registry,” March 2010.).

Command Name: RANG

Description: End point byte range (for STREAM mode).

FEAT String: RANG

Command Type: Service execution/parameter setting

Conformance Requirements: Optional

Reference: This specification

6.  Security Considerations

This memo does not directly concern security. It is not believed that any of the mechanisms documented here impact in any particular way upon the security of FTP.

7.  References

7.1. Normative References

7.2. Informative References

Appendix A.  Acknowledgements and Contributors

Thanks to ...

Portions of [RFC3659] (Hethmon, P., “Extensions to FTP,” March 2007.) were wholly reused in this document.

Appendix B.  Document History

[[ to be removed by the RFC editor before publication as an RFC. ]]

Known issues concerning this draft:

• Expand current alternatives (ABOR, close data connection) and why they are not optimal.

draft-bryan-ftp-range-00 : December 13, 2010.

• Initial draft.

Authors' Addresses