Network Working Group M. Crispin Internet Draft: IMAP MULTIAPPEND University of Washington February 2000 Document: internet-drafts/draft-crispin-imap-multiappend-00.txt INTERNET MESSAGE ACCESS PROTOCOL - MULTIAPPEND EXTENSION Status of this Memo This document is an Internet-Draft and is in full conformance with all provisions of Section 10 of RFC 2026. Internet-Drafts are working documents of the Internet Engineering Task Force (IETF), its areas, and its working groups. Note that other groups may also distribute working documents as Internet- Drafts. 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." The list of current Internet-Drafts can be accessed at http://www.ietf.org/ietf/1id-abstracts.txt To view the list Internet-Draft Shadow Directories, see http://www.ietf.org/shadow.html. A revised version of this draft document will be submitted to the RFC editor as a Proposed Standard for the Internet Community. Discussion and suggestions for improvement are requested, and should be sent to ietf-imapext@IMC.ORG. This document will expire before 10 July 2000. Distribution of this memo is unlimited. Abstract This document describes the multiappending extension to the [IMAP] protocol. This extension provides substantial performance improvements for IMAP clients which upload multiple messages at a time to a mailbox on the server. A server which supports this extension indicates this with a capability name of "MULTIAPPEND". Crispin [Page 1] INTERNET DRAFT IMAP MULTIAPPEND EXPIRES 10 July 2000 Introduction The MULTIAPPEND extension permits uploading of multiple messages with a single command. When used in conjunction with the [LITERAL+] extension, the entire upload is accomplished in a single command/response round trip. A MULTIAPPEND APPEND operation is atomic; either all messages are successfully appended, or no messages are appended. In the base IMAP specification, each message must be appended in a separate command, and there is no mechanism to "unappend" messages if an error occurs while appending. Also, some mail stores may require an expensive "open/lock + sync/unlock/close" operation as part of appending; this can be quite expensive if it must be done on a per-message basis. If the server supports both LITERAL+ and pipelining but not MULTIAPPEND, it may be possible to get some of the performance advantages of MULTIAPPEND by doing a pipelined "batch" append. However, it will not work as well as MULTIAPPEND for the following reasons: 1) Multiple APPEND commands, even as part of a pipelined batch, are non-atomic by definition. There is no way to revert the mailbox to the state before the batch append in the event of an error. 2) It may not be feasible for the server to coalesce pipelined APPEND operations so as to avoid the "open/lock + sync/unlock/close" overhead described above. In any case, such coalescing would be timing dependent and thus potentially unreliable. In particular, with traditional UNIX mailbox files, it is assumed that a lock is held only for a single atomic operation, and many applications disregard any lock that is older than 5 minutes. 3) If an error occurs, depending upon the nature of the error, it is possible for additional messages to be appended after the error. For example, the user wants to append 5 messages, but a disk quota error occurs with the third message because of its size. However, the fourth and fifth messages have already been sent in the pipeline, so the mailbox ends up with the first, second, fourth, and fifth messages of the batch appended. Crispin [Page 2] INTERNET DRAFT IMAP MULTIAPPEND EXPIRES 10 July 2000 Extension to IMAP4rev1 Base Protocol Commands 6.3.11. APPEND Command Arguments: mailbox name one or more messages to upload, specified as: OPTIONAL flag parenthesized list OPTIONAL date/time string message literal Data: no specific responses for this command Result: OK - append completed NO - append error: can't append to that mailbox, error in flags or date/time or message text BAD - command unknown or arguments invalid The APPEND command appends the literal arguments as new messages to the end of the specified destination mailbox. This argument SHOULD be in the format of an [RFC-822] message. 8-bit characters are permitted in the message. A server implementation that is unable to preserve 8-bit data properly MUST be able to reversibly convert 8-bit APPEND data to 7-bit using a [MIME-IMB] content transfer encoding. Note: There MAY be exceptions, e.g. draft messages, in which required [RFC-822] header lines are omitted in the message literal argument to APPEND. The full implications of doing so MUST be understood and carefully weighed. If a flag parenthesized list is specified, the flags SHOULD be set in the resulting message; otherwise, the flag list of the resulting message is set empty by default. If a date-time is specified, the internal date SHOULD be set in the resulting message; otherwise, the internal date of the resulting message is set to the current date and time by default. If the append is unsuccessful for any reason, the mailbox MUST be restored to its state before the APPEND attempt; no partial appending is permitted. The server MAY return an error before processing all the message arguments. If the destination mailbox does not exist, a server MUST return an error, and MUST NOT automatically create the mailbox. Unless it is certain that the destination mailbox can not be created, the Crispin [Page 3] INTERNET DRAFT IMAP MULTIAPPEND EXPIRES 10 July 2000 server MUST send the response code "[TRYCREATE]" as the prefix of the text of the tagged NO response. This gives a hint to the client that it can attempt a CREATE command and retry the APPEND if the CREATE is successful. If the mailbox is currently selected, the normal new message actions SHOULD occur. Specifically, the server SHOULD notify the client immediately via an untagged EXISTS response. If the server does not do so, the client MAY issue a NOOP command (or failing that, a CHECK command) after one or more APPEND commands. Example: C: A003 APPEND saved-messages (\Seen) {310} S: + Ready for literal data C: Date: Mon, 7 Feb 1994 21:52:25 -0800 (PST) C: From: Fred Foobar C: Subject: afternoon meeting C: To: mooch@owatagu.siam.edu C: Message-Id: C: MIME-Version: 1.0 C: Content-Type: TEXT/PLAIN; CHARSET=US-ASCII C: C: Hello Joe, do you think we can meet at 3:30 tomorrow? C: (\Seen) " 7-Feb-1994 22:43:04 -0800" {286} S: + Ready for literal data C: Date: Mon, 7 Feb 1994 22:43:04 -0800 (PST) C: From: Joe Mooch C: Subject: Re: afternoon meeting C: To: foobar@blurdybloop.com C: Message-Id: C: MIME-Version: 1.0 C: Content-Type: TEXT/PLAIN; CHARSET=US-ASCII C: C: 3:30 is fine with me. C: S: A003 OK APPEND completed C: A004 APPEND bogusname (\Flagged) {1023} S: A004 NO [TRYCREATE] No such mailbox as bogusname C: A005 APPEND test (\Flagged) {99} S: + Ready for literal data C: Date: Mon, 7 Feb 2000 22:43:04 -0800 (PST) C: From: Fred Foobar C: Subject: hmm... C: {35403} S: A005 NO APPEND failed: Disk quota exceeded Note: The APPEND command is not used for message delivery, because it does not provide a mechanism to transfer [SMTP] envelope information. Crispin [Page 4] INTERNET DRAFT IMAP MULTIAPPEND EXPIRES 10 July 2000 Modification to IMAP4rev1 Base Protocol Formal Syntax append = "APPEND" SPACE mailbox 1*append_message append-message = [SPACE flag_list] [SPACE date_time] SPACE literal MULTIAPPEND Interaction with UIDPLUS Extension Servers which support both MULTIAPPEND and [UIDPLUS] will have the "resp-code-apnd" rule modified as follows: resp-code-apnd = "APPENDUID" SPACE nz_number 1*(SPACE uniqueid) That is, the APPENDUID response code returns as many UIDs as there were messages appended in the multiple append. Security Considerations Security issues are not discussed in this memo. References [IMAP] Crispin, M., "Internet Message Access Protocol - Version 4rev1", RFC 2060, December 1996. [LITERAL+] Myers, J. "IMAP4 non-synchronizing literals", RFC 2088, January 1997. [UIDPLUS] Myers, J., "IMAP4 UIDPLUS extension", RFC 2359, June 1988. Author's Address Mark R. Crispin Networks and Distributed Computing University of Washington 4545 15th Aveneue NE Seattle, WA 98105-4527 Phone: (206) 543-5762 EMail: MRC@CAC.Washington.EDU Crispin [Page 5]