Network Working Group D. Cridland Internet-Draft C. King Expires: October 27, 2007 Isode Limited April 25, 2007 Contexts for IMAP4 draft-cridland-imap-context-01 Status of this Memo By submitting this Internet-Draft, each author represents that any applicable patent or other IPR claims of which he or she is aware have been or will be disclosed, and any of which he or she becomes aware will be disclosed, in accordance with Section 6 of BCP 79. 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. The list of Internet-Draft Shadow Directories can be accessed at http://www.ietf.org/shadow.html. This Internet-Draft will expire on October 27, 2007. Copyright Notice Copyright (C) The IETF Trust (2007). Abstract The IMAP4rev1 protocol has powerful search facilities as part of the core protocol, but lacks the ability to create live, updated results which can be easily handled. This memo provides such an extension, and shows how it can be used to provide a facility similar to virtual mailboxes. Cridland & King Expires October 27, 2007 [Page 1] Internet-Draft IMAP CONTEXT April 2007 Table of Contents 1. Conventions used in this document . . . . . . . . . . . . . . 3 2. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 3 3. Protocol Changes . . . . . . . . . . . . . . . . . . . . . . . 3 3.1. Overview . . . . . . . . . . . . . . . . . . . . . . . . . 3 3.2. Context Hint . . . . . . . . . . . . . . . . . . . . . . . 4 3.3. Notifications of changes . . . . . . . . . . . . . . . . . 4 3.3.1. ADDTO Return Data Item . . . . . . . . . . . . . . . . 5 3.3.2. REMOVEFROM Return Data Item . . . . . . . . . . . . . 6 3.4. Partial results . . . . . . . . . . . . . . . . . . . . . 6 3.5. Caching results . . . . . . . . . . . . . . . . . . . . . 7 4. Formal Syntax . . . . . . . . . . . . . . . . . . . . . . . . 8 5. Security Considerations . . . . . . . . . . . . . . . . . . . 9 6. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 9 7. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 9 8. References . . . . . . . . . . . . . . . . . . . . . . . . . . 9 8.1. Normative References . . . . . . . . . . . . . . . . . . . 9 8.2. Informative References . . . . . . . . . . . . . . . . . . 10 Appendix A. Cookbook . . . . . . . . . . . . . . . . . . . . . . 10 A.1. Virtual Mailboxes . . . . . . . . . . . . . . . . . . . . 10 A.2. Trash Mailboxes . . . . . . . . . . . . . . . . . . . . . 10 A.3. Other uses . . . . . . . . . . . . . . . . . . . . . . . . 11 A.4. Resynchronizing Contexts . . . . . . . . . . . . . . . . . 11 Appendix B. Server Implementation Notes . . . . . . . . . . . . . 11 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 12 Intellectual Property and Copyright Statements . . . . . . . . . . 13 Cridland & King Expires October 27, 2007 [Page 2] Internet-Draft IMAP CONTEXT April 2007 1. Conventions used in this document In examples, "C:" and "S:" indicate lines sent by the client messaging user agent and IMAP4rev1 ([IMAP]) server respectively. The examples show a server which supports [ESEARCH] and [IDLE], neither extension is required for this specification. The IDLE command is used to denote an extended period of time during which any response may be sent to the client. 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 [KEYWORDS]. Other capitalised words are typically names of IMAP extensions or commands - these are uppercased for clarity only, and are case- insensitive. [[ Editorial comments are like this. XML2RFC working source is held at http://svn.dave.cridland.net/svn/ietf-drafts/ draft-cridland-imap-contexts.xml ]] 2. Introduction Although the basic SEARCH command defined in [IMAP], as enhanced by [ESEARCH], is relatively compact in its representation, this reduction only saves a certain amount of data, and huge mailboxes can overwhelm the storage available for results on even relatively high- end desktop machines. This memo borrows concepts from [ACAP], providing a windowed view onto search results, as well as bandwidth and round-trip efficient updates. It is intended that the protocol may be easily adapted onto the SORT command specified in [SORT]. 3. Protocol Changes 3.1. Overview This extension is present in any IMAP4rev1 server which includes the string "CONTEXT", or any string beginning "CONTEXT=", within its advertised capabilities. Such servers also accept three additional return options, and provide three new result data items, and no new responses. The first search Cridland & King Expires October 27, 2007 [Page 3] Internet-Draft IMAP CONTEXT April 2007 return option is CONTEXT, an optional hint that the criteria will be used repeatedly, and is defined in Section 3.2. The second is UPDATE, which causes the server to provide efficient notifications of changes to the results. This is defined in Section 3.3. Finally, the PARTIAL return specifier causes the server to return a subset of the results in set-syntax. This allows for "virtual scrollbars" and other UI conveniences to be achieved without having to preload the entire result set, and is described in Section 3.4. All of the return specifiers have no interaction with either each other or any return specifiers defined in [ESEARCH]. 3.2. Context Hint The return option CONTEXT SHOULD be used by a client to indicate that subsequent use of the criteria are likely. Servers MAY ignore this return option, or use it as a hint to maintain a full result set, or index. A client might choose to obtain a count of matching messages prior to obtaining actual results. Here, the client signifies its intention to fetch the results themselves: C: A01 SEARCH RETURN (CONTEXT COUNT) UNDELETED UNKEYWORD $Junk S: * ESEARCH (TAG "A01") COUNT 23765 S: A01 OK Search completed. 3.3. Notifications of changes The search return option UPDATE, if used by a client, causes the server to issue unsolicited notifications containing updates to the SEARCH results which would be returned by an unmodified SEARCH. These results are carried in ADDTO and REMOVEFROM data items in ESEARCH/ESORT responses. Both ADDTO and REMOVEFROM data items SHOULD be delivered to clients in a timely manner, as and when results changes, whether by new messages arriving in the mailbox, metadata such as flags being changed, or messages being expunged. Typically, this would occur at the same time as the FETCH, EXISTS or EXPUNGE responses carrying the source of the change. Unlike [ACAP], there is no requirement that a context need be created Cridland & King Expires October 27, 2007 [Page 4] Internet-Draft IMAP CONTEXT April 2007 with CONTEXT to use UPDATE, and in addition, the lack of UPDATE with a CONTEXT does not affect the results caused by later SEARCH commands - there is no snapshot facility. There is no interaction between UPDATE and any other return options, therefore use of RETURN (UPDATE MIN), for example, does not notify about the minimum UID or sequence number, but notifies instead about all changes to the set of matching messages. In particular, this means that a client using UPDATE and PARTIAL on the same search program MAY receive notifications about messages which do not interest it currently. This time, the client will require notifications of updates, and chooses to obtain a count: C: B01 UID SEARCH RETURN (UPDATE COUNT) DELETED KEYWORD $Junk S: * ESEARCH (TAG "B01") COUNT 74 S: B01 OK Search completed, will notify. 3.3.1. ADDTO Return Data Item The ADDTO return data item contains, as payload, a list containing pairs of a position and a set of results to be inserted at the position. For ESEARCH responses, the position MAY be zero, and MAY be ignored by clients. The results are specified as UIDs or message numbers, depending on how the UPDATE was specified. If the UPDATE was present in a SEARCH command, the results will be message numbers; in a UID SEARCH command, they will be UIDs. C: B02 IDLE S: + Idle [...] S: * 23762 FETCH (FLAGS (\Deleted \Seen)) S: * ESEARCH (TAG "B01") UID ADDTO (0 32768) C: DONE S: B02 OK Not Idle. Note that this example assumes message 23762 with UID 32768 previously had neither \Deleted nor $Junk set. Also note that only the ADDTO is included, and not the COUNT. Cridland & King Expires October 27, 2007 [Page 5] Internet-Draft IMAP CONTEXT April 2007 3.3.2. REMOVEFROM Return Data Item The REMOVEFROM return data item contains a set of results to be removed. The results to be removed are referenced by message number or UID, as appropriate, and need not be in the same order as the results. Servers SHOULD sort the results in order to use the sequence-set syntax as efficiently as possible. There is no requirement on servers to avoid issuing REMOVEFROM return data at any particular moment, in particular this is distinct from EXPUNGE responses. The results are specified as UIDs or message numbers, depending on how the UPDATE was specified. If the UPDATE was present in a SEARCH command, the results will be message numbers; in a UID SEARCH command, they will be UIDs. Command B03 here is purely an example of a command which prohibits EXPUNGE messages. The REMOVEFROM could have been sent without any command in progress. C: B03 SEARCH RETURN () 1:* ALL S: * ESEARCH (TAG "B03") ALL 1:49152 S: * ESEARCH (TAG "B01") UID REMOVEFROM 32768 S: B03 OK Search completed. C: B04 IDLE S: + Idle S: * EXPUNGE 23762 [...] C: DONE S: B04 OK Not Idle. 3.4. Partial results The PARTIAL search return option causes the server to provide in an ESEARCH response the range from the results denoted by the sequence range given as the mandatory argument. The first result is 1, thus the first 500 results would be obtained by a return option of "PARTIAL 1:500", and the second 500 by "PARTIAL 501:1000". This intentionally mirrors message sequence numbers. Where a PARTIAL search return option references results which do not exist, by using a range which starts or ends higher than the COUNT of results, then the server returns those results which are in the set. This yields a PARTIAL return data item which has, as payload, the original range and a potentially missing set of results which may be shorter than the extent of the range. Cridland & King Expires October 27, 2007 [Page 6] Internet-Draft IMAP CONTEXT April 2007 The subset of results are returned in sequence-set syntax, and servers SHOULD order results from a SEARCH for maximum efficiency. Clients need not request PARTIAL results in any particular order. // Recall from A01 that there are 23764 results. C: A02 UID SEARCH RETURN (PARTIAL 23500:24000) UNDELETED UNKEYWORD $Junk C: A03 UID SEARCH RETURN (PARTIAL 1:500) UNDELETED UNKEYWORD $Junk C: A04 UID SEARCH RETURN (PARTIAL 24000:24500) UNDELETED UNKEYWORD $Junk S: * ESEARCH (TAG "A02") UID PARTIAL (23500:24000 ...) // 264 results in set syntax elided, // this spans the end of the results. S: A02 OK Completed. S: * ESEARCH (TAG "A03") UID PARTIAL (1:500 ...) // 500 results in set syntax elided. S: A03 OK Completed. S: * ESEARCH (TAG "A04") UID PARTIAL (24000:24500 NIL) // No results are present, this is beyond the end of the results. S: A04 OK Completed. 3.5. Caching results Server implementations MAY cache results from a search or sort, whether or not hinted to by CONTEXT, in order to make subsequent searches more efficient, perhaps by recommencing a subsequent PARTIAL search where a previous search left off. However servers MUST behave identically whether or not internal caching is taking place, therefore any such cache is required to be updated as changes to the mailbox occur. An alternate strategy would be to discard results when any change occurs to the mailbox. Cridland & King Expires October 27, 2007 [Page 7] Internet-Draft IMAP CONTEXT April 2007 4. Formal Syntax The collected formal syntax. This includes definitions from [IMAP] and [IMAP-ABNF], and uses ABNF as defined in [ABNF]. capability =/ "CONTEXT" / "CONTEXT=" atom addto-position = number ;; Number may be 0 for SEARCH result additions. ;; from RFC3501 modifier-context = "CONTEXT" modifier-partial = "PARTIAL" SP seq-range ;; from [IMAP] modifier-update = "UPDATE" search-return-opt =/ modifier-context / modifier-partial / modifier-update ;; All conform to , from [IMAP-ABNF] ret-data-addto = "ADDTO" SP "(" addto-position SP sequence-set *(SP addto-position SP sequence-set) ")" ;; from [IMAP] ret-data-partial = "PARTIAL" SP "(" seq-range SP partial-results ")" ;; is the requested range. ;; from [IMAP] partial-results = sequence-set / "NIL" ;; from [IMAP] ;; NIL indicates no results correspond to the requested range. ret-data-removefrom = "REMOVEFROM" sequence-set ;; from [IMAP] search-return-data =/ ret-data-partial / ret-data-addto / ret-data-removefrom ;; All conform to , from [IMAP-ABNF] Cridland & King Expires October 27, 2007 [Page 8] Internet-Draft IMAP CONTEXT April 2007 5. Security Considerations It is believed that this specification introduces no serious new security considerations. However, implementors are advised to refer to [IMAP]. Creation of contexts, both for UPDATE and PARTIAL, can benefit from storing potentially large result sets on the server. Implementors are advised to take care not to provide a method for denial of service (DoS) attacks based on this; the notes in Appendix B may aid in implementation decisions. Note that a server avoiding storing the results will have much increased I/O, which may also be an avenue for DoS attacks. 6. IANA Considerations IMAP4 capabilities are registered by publishing a standards track or IESG approved experimental RFC. The registry is currently located at: http://www.iana.org/assignments/imap4-capabilities This document defines the CONTEXT IMAP capability. IANA is requested to add it to the registry accordingly. 7. Acknowledgements Much of the design of this extension can be found in ACAP. Valuable comments, both in agreement and in dissent, were received from Alexey Melnikov, Arnt Gulbrandsen, Randall Gellens, Cyrus Daboo, and others, and many of these comments have had significant influence on the design or the text. The authors are grateful to all those involved, including those not mentioned here. 8. References 8.1. Normative References [ABNF] Crocker, D., Ed. and P. Overell, "Augmented BNF for Syntax Specifications: ABNF", RFC 4234, October 2005. [IMAP] Crispin, M., "INTERNET MESSAGE ACCESS PROTOCOL - VERSION 4rev1", RFC 3501, March 2003. [IMAP-ABNF] Cridland & King Expires October 27, 2007 [Page 9] Internet-Draft IMAP CONTEXT April 2007 Melnikov, A. and C. Daboo, "Collected Extensions to IMAP4 ABNF", RFC 4466, April 2006. [KEYWORDS] Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, March 1997. 8.2. Informative References [ACAP] Newman, C. and J. Myers, "ACAP -- Application Configuration Access Protocol", RFC 2244, November 1997. [ESEARCH] Melnikov, A. and D. Cridland, "IMAP4 extension to SEARCH command for controlling what kind of information is returned", draft-melnikov-imap-search-ret-03 (work in progress), June 2006. [IDLE] Leiba, B., "IMAP4 IDLE command", RFC 2177, June 1997. [SORT] Crispin, M. and K. Murchison, "INTERNET MESSAGE ACCESS PROTOCOL - SORT AND THREAD EXTENSIONS", draft-ietf-imapext-sort-18 (work in progress), November 2006. Appendix A. Cookbook A.1. Virtual Mailboxes It is possible to use the facilities described within this memo to create a facility largely similar to a virtual mailbox, but handled on the client side. Initially, the client SELECTs the real "backing" mailbox. Next, it can switch to a filtered view at any time by issuing a SEARCH RETURN (COUNT UPDATE CONTEXT), and using SEARCH RETURN (PARTIAL x:y) as the user scrolls, feeding the results into a FETCH as required to populate summary views. A.2. Trash Mailboxes Certain contexts are particularly useful for client developers wishing to present something similar to the common trash mailbox metaphor in limited bandwidth. The simple criteria of UNDELETED only matches undeleted messages, and the corresponding DELETED search key can be used to display a per-mailbox trash-like virtual mailbox. Cridland & King Expires October 27, 2007 [Page 10] Internet-Draft IMAP CONTEXT April 2007 A.3. Other uses It is entirely possible to simultaneously have two or more UPDATE searches in operation. This can be used to build a grouped message display in some cases, and also allows for monitoring counts of messages matching certain complex criteria. A.4. Resynchronizing Contexts The creation of a context, and immediate access to it, can all be accomplished in a single round-trip. Therefore, whilst it is possible to elide resynchronization if no changes have occurred, it is simpler in most cases to resynchronize by simply recreating the context. Appendix B. Server Implementation Notes Although a server may cache the results, this is not mandated nor required. UPDATE processing, for example, can be achieved by comparison of the old flag state (if any) and the new, and PARTIAL can be achieved by re-running the search until the suitable window is required. This is a result of there being no snapshot facility. For example, on a new message, the server can simply test for matches against all current UPDATE context search programs, and for any that match, send the ADDTO return data. Similarly, for a flag change on an existing message, the server can check whether the message matched with its old flags, whether it matches with new flags, and provide ADDTO or REMOVEFROM return data accordingly if these results differ. For PARTIAL requests, the server can perform a full search, discarding results until the lower bound is hit, and stopping the search when sufficient results have been obtained. With some additional state, it is possible to restart PARTIAL searches, thus avoiding performing the initial discard phase. For the best performance, however, caching the full search results is needed, which can allow for faster responses at the expense of memory. One reasonable strategy would be to balance this trade-off at run-time, discarding search results after a suitable timeout, and regenerating them as required. This yields state requirements of storing the search program for any UPDATE contexts, and optionally storing both search program and Cridland & King Expires October 27, 2007 [Page 11] Internet-Draft IMAP CONTEXT April 2007 (updated) results for further contexts as required. Authors' Addresses Dave Cridland Isode Limited 5 Castle Business Village 36, Station Road Hampton, Middlesex TW12 2BX GB Email: dave.cridland@isode.com Curtis King Isode Limited 5 Castle Business Village 36, Station Road Hampton, Middlesex TW12 2BX GB Email: cking@mumbo.ca Cridland & King Expires October 27, 2007 [Page 12] Internet-Draft IMAP CONTEXT April 2007 Full Copyright Statement Copyright (C) The IETF Trust (2007). This document is subject to the rights, licenses and restrictions contained in BCP 78, and except as set forth therein, the authors retain all their rights. This document and the information contained herein are provided on an "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY, THE IETF TRUST AND THE INTERNET ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Intellectual Property The IETF takes no position regarding the validity or scope of any Intellectual Property Rights or other rights that might be claimed to pertain to the implementation or use of the technology described in this document or the extent to which any license under such rights might or might not be available; nor does it represent that it has made any independent effort to identify any such rights. Information on the procedures with respect to rights in RFC documents can be found in BCP 78 and BCP 79. Copies of IPR disclosures made to the IETF Secretariat and any assurances of licenses to be made available, or the result of an attempt made to obtain a general license or permission for the use of such proprietary rights by implementers or users of this specification can be obtained from the IETF on-line IPR repository at http://www.ietf.org/ipr. The IETF invites any interested party to bring to its attention any copyrights, patents or patent applications, or other proprietary rights that may cover technology that may be required to implement this standard. Please address the information to the IETF at ietf-ipr@ietf.org. Acknowledgment Funding for the RFC Editor function is provided by the IETF Administrative Support Activity (IASA). Cridland & King Expires October 27, 2007 [Page 13]