INTERNET DRAFT Stephen N. Zilles Adobe Systems Inc. November 21, 1997 Expires: May 21, 1998 Rationale for the Structure of the Model and Protocol for the Internet Printing Protocol STATUS OF THIS MEMO This document is an Internet-Draft. 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.'' To learn the current status of any Internet-Draft, please check the ''1id-abstracts.txt'' listing contained in the Internet- Drafts Shadow Directories on ftp.is.co.za (Africa), nic.nordu.net (Europe), munnari.oz.au (Pacific Rim), ds.internic.net (US East Coast), or ftp.isi.edu (US West Coast). ABSTRACT This document is one of a set of documents which together describe all aspects of a new Internet Printing Protocol (IPP). IPP is an application level protocol that can be used for distributed printing on the Internet. There are multiple parts to IPP, but the primary architectural components are the Model, the Protocol and an interface to Directory Services. This document provides a high level overview of the architecture and provides the rationale for the decisions made in structuring the architecture. The full set of IPP documents include: Requirements for an Internet Printing Protocol Internet Printing Protocol/1.0: Model and Semantics Internet Printing Protocol/1.0: Protocol Specification Rationale for the Structure of the Model and Protocol for the Internet Printing Protocol Zilles draft-ietf-ipp-rat-02.txt [Page 1] Expires: May 21, 1998 INTERNET DRAFT Internet Printing Rationale November 21, 1997 1. ARCHITECTURAL OVERVIEW The Internet Printing Protocol (IPP) is an application level protocol that can be used for distributed printing on the Internet. This protocol defines interactions between a client and a server. The protocol allows a client to inquire about capabilities of a printer, to submit print jobs and to inquire about and cancel print jobs. The server for these requests is the Printer; the Printer is an abstraction of a generic document output device and/or a print service provider. Thus, the Printer could be a real printing device, such as a computer printer or fax output device, or it could be a service that interfaced with output devices. The architecture for IPP defines (in the Model document [IPP-MOD]) an abstract Model for the data which is used to control the printing process and to provide information about the process and the capabilities of the Printer. This abstract Model is hierarchical in nature and reflects the structure of the Printer and the Jobs that may be being processed by the Printer. The Internet provides a channel between the client and the server/Printer. Use of this channel requires flattening and sequencing the hierarchical Model data. Therefore, the IPP also defines (in the Protocol document [IPP-PRO]) an encoding of the data in the model for transfer between the client and server. This transfer of data may be either a request or the response to a request. Finally, the IPP defines (in the Protocol document [IPP-PRO]) a protocol for transfering the encoded request and response data between the client and the server/Printer. An example of a typical interaction would be a request from the client to create a print job. The client would assemble the Model data to be associated with that job, such as the name of the job, the media to use, the number of pages to place on each media instance, etc. This data would then be encoded according to the Protocol and would be transmitted according to the Protocol. The server/Printer would receive the encoded Model data, decode it into a form understood by the server/Printer and, based on that data, do one of two things: (1) accept the job or (2) reject the job. In either case, the server must construct a response in terms of the Model data, encode that response according to the Protocol and transmit that encoded Model data as the response to the request using the Protocol. Another part of the IPP architecture is the Directory Schema (described in the model document).. The role of the Zilles draft-ietf-ipp-rat-02.txt [Page 2] Expires: May 21, 1998 INTERNET DRAFT Internet Printing Rationale November 21, 1997 Directory Schema is to provide a standard set of attributes which might be used to query a directory service for the URI of a Printer that is likely to meet the needs of the client. The IPP architecture also addresses security issues such as control of access to server/Printers and secure transmissions of requests, response and the data to be printed. 2. THE PRINTER Because the (abstract) server/Printer encompasses a wide range of implementations, it is necessary to make some assumptions about a minimal implementation. The most likely minimal implementation is one that is embedded in an output device running a specialized real time operating system and with limited processing, memory and storage capabilities. This Printer will be connected to the Internet and will have at least a TCP/IP capability with (likely) SNMP [RFC1905, RFC1906] support for the Internet connection. In addition, it is likely the the Printer will be an HTML/HTTP server to allow direct user access to information about the printer. 3. RATIONALE FOR THE MODEL The Model [IPP-MOD] is defined independently of any encoding of the Model data both to support the likely uses of IPP and to be robust with respect to the possibility of alternate encodings. It is expected that a client or server/Printer would represent the Model data in some data structure within the applications/servers that support IPP. Therefore, the Model was designed to make that representation straightforward. Typically, a parser or formatter would be used to convert from or to the encoded data format. Once in an internal form suitable to a product, the data can be manipulated by the product. For example, the data sent with a Print Job can be used to control the processing of that Print Job. The semantics of IPP are attached to the (abstract) Model. Therefore, the application/server is not dependent on the encoding of the Model data, and it is possible to consider alternative mechanisms and formats by which the data could be transmitted from a client to a server; for example, a server could have a direct, client-less GUI interface that might be used to accept some kinds of Print Jobs. This independence would also allow a different encoding and/or transmission Zilles draft-ietf-ipp-rat-02.txt [Page 3] Expires: May 21, 1998 INTERNET DRAFT Internet Printing Rationale November 21, 1997 mechanism to be used if the ones adopted here were shown to be overly limiting in the future. Such a change could be migrated into new products as an alternate protocol stack/parser for the Model data. Having an abstract Model also allows the Model data to be aligned with the (abstract) model used in the Printer [RFC1759], Job and Host Resources MIBs. This provides consistency in interpretation of the data obtained independently of how the data is accessed, whether via IPP or via SNMP [RFC1905, RFC1906] and the Printer/Job MIBs. There is one aspect of the Model that deserves some extra explanation. There are two ways for identifying a Job object: (a) with a Job URI and (b) using a combination of the Printer URI and a Job ID (a 32 bit positive integer). Allowing Job objects to have URIs allows for flexibility and scalability. For example a job could be moved from a printer with a large backlog to one with a smaller load and the job identification, the Job object URI, need not change. However, many existing printing systems have local models or interface constraints that force Job objects to be identified using only a 32-bit positive integer rather than a URI. This numeric Job ID is only unique within the context of the Printer object to which the create request was originally submitted. In order to allow both types of client access to Jobs (either by Job URI or bynumeric Job ID), when the Printer object successfully processes a create request and creates a new Job, the Printer object SHALL generate both a Job URI and a Job ID for the new Job object. This requirement allows all clients to access Printer objects and Job objects no matter the local constraints imposed on the client implementation. 4. RATIONALE FOR THE PROTOCOL There are two parts to the Protocol: (1) the encoding of the Model data and (2) the mechanism for transmitting the model data between client and server. 4.1 The Encoding To make it simpler to develop embedded printers, a very simple binary encoding has been chosen. This encoding is adequate to represent the kinds of data that occur within the Model. It has a simple structure consisting of sequences of attributes. Each attribute has a name, prefixed by a name length, and a value.. The names are strings constrained to characters from a subset of ASCII. The values are either Zilles draft-ietf-ipp-rat-02.txt [Page 4] Expires: May 21, 1998 INTERNET DRAFT Internet Printing Rationale November 21, 1997 scalars or a sequence of scalars. Each scalar value has a length specification and a value tag which indicates the type of the value. The value type has two parts: a major class part, such as integer or string, and a minor class part which distinguishes the usage of the major class, such as dateTime string. Tagging of the values with type information allows for introducing new value types at some future time. A fully encoded request/response has a version number, an operation (for a request) or a status and optionally a status message (for a response), associated parameters and attributes which are encoded Model data and, optionally (for a request), print data following the Model data. 4.2 The Transmission Mechanism The chosen mechanism for transmitting the encoded Model data is HTTP 1.1 Post (and associated response). No modifications to HTTP 1.1 are proposed or required. The sole role of the Transmission Mechanism is to provide a transfer of encoded Model data from/to the client to/from the server. This could be done using any data delivery mechanism. The key reasons why HTTP 1.1 Post is used are given below. The most important of these is the first. With perhaps this exception, these reasons could be satisfied by other mechanisms. There is no claim that this list uniquely determines a choice of mechanism. 1. HTTP 1.0 is already widely deployed and, based on the recent evidence, HTTP 1.1 is being widely deployed as the manufacturers release new products. The performance benefits of HTTP 1.1 have been shown and manufactures are reacting postively. Wide deployment has meant that many of the problems of making a protocol work in a wide range of environments from local net to intranet to internet have been solved and will stay solved with HTTP 1.1 deployment. 2. HTTP 1.1 solves most of the problems that might have required a new protocol to be developed. HTTP 1.1 allows persistent connections that make a multi-message protocol be more efficient; for example it is practical to have a separate CreatJob and SendJob messages. Chunking allows the transmission of large print files without having to prescan the file to determine the file length. The accept headers allow the Zilles draft-ietf-ipp-rat-02.txt [Page 5] Expires: May 21, 1998 INTERNET DRAFT Internet Printing Rationale November 21, 1997 client's protocol and localization desires to be transmitted with the IPP operations and data. If the Model were to provide for the redirection of Job requests, such as Cancel-Job, when a Job is moved, the HTTP redirect response allows a client to be informed when a Job he is interested in is moved to another server/Printer for any reason. 3. Most network Printers will be implementing HTTP servers for reasons other than IPP. These network attached Printers want to provide information on how to use the printer, its current state, HELP information, etc. in HTML. This requires having an HTTP server which would be available to do IPP functions as well. 4. Most of the complexity of HTTP 1.1 is concerned with the implementation of HTTP proxies and not the implementation of HTTP clients and/or servers. Work is proceding in the HTTP Working Group to help identify what must be done by a server. As the Protocol document shows, that is not very much. 5. HTTP implementations provide support for handling URLs that would have to be provided if a new protocol were defined. 6. An HTTP based solution fits well with the Internet security mechanism that are currently deployed or being deployed. HTTP will run over TLS/SSL. The digest authentication mechanism of HTTP 1.1 provides an adequate level of access control (at least for intranet usage). These solutions are deployed and in practical use; a new solution would require extensive use to have the same degree of confidence in its security. 7. HTTP provides an extensibility model that a new protocol would have to develop independently. In particular, the headers, content-types (via Internet Media Types) and error codes have wide acceptance and a useful set of definitions and methods for extension. 8. Although not strictly a reason why IPP should use HTTP as the transmission protocol, it is extremely helpful that there are many prototyping tools that work with HTTP and that CGI scripts can be used to test and debug parts of the protocol. Zilles draft-ietf-ipp-rat-02.txt [Page 6] Expires: May 21, 1998 INTERNET DRAFT Internet Printing Rationale November 21, 1997 9. Finally, the POST method was chosen to carry the print data because its usage for data transmission has been established, it works and the results are available via CGI scripts where that is relevant. Creating a new method would have better identified the intended use of the POSTed data, but a new method would be more difficult to deploy. Because there were no strong arguments for or against using a new method, the POST method was used. 5. RATIONALE FOR THE DIRECTORY SCHEMA Successful use of IPP depends on the client finding a suitable IPP enabled Printer to which to send a IPP requests, such as print a job. This task is simplified if there is a Directory Service which can be queried for a suitable Printer. The purpose of the Directory Schema is to have a standard description of Printer attributes that can be associated the the URI for the printer. These attributes are a subset of the Model attributes and can be encoded in the appropriate query syntax for the Directory Service being used by the client. 6. RATIONALE FOR SECURITY Security is an areas of active work on the Internet. Complete solutions to a wide range of security concerns are not yet available. Therefore, in the design of IPP, the focus has been on identifying a set of security protocols/features that are implemented (or currently implementable) and solve real problems with distributed printing. The two areas that seem appropriate to support are: (1) authorization to use a Printer and (2) secure interaction with a printer. The chosen mechanisms are the digest authentication mechanism of HTTP 1.1 and the TLS/SSL [TLS-PRO] secure communication mechanism. To provide access to both HTTP security and TLS security, IPP Printers provide two URLs for accessing the Printer object: one URL for which the scheme is HTTP for normal HTTP 1.1 access and one URL for which the scheme is HTTPS for access to the Printer using TLS/SSL protected communication. Two URLs are necessary because all allowed modes of TLS require some level of security beyond HTTP security so a TLS capable URL cannot be used to negotiate down to HTTP security. 7. COPYRIGHT This document and translations of it may be copied and furnished to others, and derivative works that comment on or Zilles draft-ietf-ipp-rat-02.txt [Page 7] Expires: May 21, 1998 INTERNET DRAFT Internet Printing Rationale November 21, 1997 otherwise explain it or assist in its implementation may be prepared, copied, published and distributed, in whole or in part, without restriction of any kind, provided that the above copyright notice and this paragraph are included on all such copies and derivative works. However, this document itself may not be modified in any way, such as by removing the copyright notice or references to the Internet Society or other Internet organizations, except as needed for the purpose of developing Internet standards in which case the procedures for copyrights defined in the Internet Standards process must be followed, or as required to translate it into languages other than English. The limited permissions granted above are perpetual and will not be revoked by the Internet Society or its successors or assigns. This document and the information contained herein is provided on an "AS IS" basis and THE INTERNET SOCIETY AND THE INTERNET ENGINEERING TASK FORCE DISCLAIMS 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. 8. REFERENCES [RFC1759] Smith, R., Wright, F., Hastings, T., Zilles, S., and Gyllenskog, J., "Printer MIB", RFC 1759, March 1995. [RFC1905] J. Case, et al. "Protocol Operations for Version 2 of the Simple Network Management Protocol (SNMPv2)", RFC 1905, January 1996. [RFC1906] J. Case, et al. "Transport Mappings for Version 2 of the Simple Network Management Protocol (SNMPv2)", RFC 1906, January 1996. [IPP-MOD] Isaacson, S, deBry, R, Hastings, T, Herriot, R, Powell, P, "Internet Printing Protocol/1.0: Model and Semantics" draft-ipp-mod-07.txt, November, 1997. Zilles draft-ietf-ipp-rat-02.txt [Page 8] Expires: May 21, 1998 INTERNET DRAFT Internet Printing Rationale November 21, 1997 [IPP-PRO] Herriot, R., Butler, S., Moore, P., Tuner, R., " Internet Printing Protocol/1.0: Protocol Specifications", draft-ipp-pro- 03.txt, November, 1997. [IPP-REQ] Wright, D., "Requirements for an Internet Printing Protocol", draft-ipp-req-01.txt, November, 1997. [TLS-PRO] Dierks, T., Allen, C., "The TLS Protocol Version 1.0", , November, 1997 9. AUTHORS ADDRESS Stephen Zilles Adobe Systems Incorporated 345 Park Avenue MailStop W14 San Jose, CA 95110-2704 Phone: +1 408 536-4766 Fax: +1 408 537-4042 Email: szilles@adobe.com Zilles draft-ietf-ipp-rat-02.txt [Page 9] Expires: May 21, 1998