In computer network engineering, a Request for Comments (RFC) is a memorandum published by the Internet Engineering Task Force (IETF) describing methods, behaviors, research, or innovations applicable to the working of the Internet and Internet-connected systems.
Through the Internet Society, engineers and computer scientists may publish discourse in the form of an RFC, either for peer review or simply to convey new concepts, information, or (occasionally) engineering humor. The IETF adopts some of the proposals published as RFCs as Internet standards.
The inception of the RFC format occurred in 1969 as part of the seminal ARPANET project. Today, it is the official publication channel for the Internet Engineering Task Force (IETF), the Internet Architecture Board (IAB), and—to some extent—the global community of computer network researchers in general.
The authors of the first RFCs typewrote their work and circulated hard copies among the ARPA researchers. Unlike the modern RFCs, many of the early RFCs were requests for comments. The RFC leaves questions open and is written in a less formal style. This less formal style is now typical of Internet Draft documents, the precursor step before being approved as an RFC.
In December 1969, researchers began distributing new RFCs via the newly operational ARPANET. RFC 1, entitled "Host Software", was written by Steve Crocker of the University of California, Los Angeles (UCLA), and published on April 7, 1969. Although written by Steve Crocker, the RFC emerged from an early working group discussion between Steve Crocker, Steve Carr and Jeff Rulifson. (The document lists Bill Duvall as having attended only the final working group meeting prior to publication.)
In RFC 3, which first defined the RFC series, Crocker started attributing the RFC series to the "Network Working Group". This group seems never to have had a formal existence, being rather defined as "this group of people", but the attribution remains on RFCs to this day.
Many of the subsequent RFCs of the 1970s also came from UCLA, not only because of the quality of the scholarship, but also because UCLA was one of the first Interface Message Processors (IMPs) on ARPANET.
Douglas Engelbart's Augmentation Research Center (ARC) at Stanford Research Institute was another of the four first ARPANET nodes, as well as the first Network Information Centre, and (as noted by the sociologist Thierry Bardini) the source of a large number of early RFCs.
From 1969 until 1998, Jon Postel served as the RFC editor. (On his death in 1998, his obituary was published as RFC 2468.) Following the expiration of the original ARPANET contract with the U.S. federal government, the Internet Society (acting on behalf of the IETF) contracted with the Networking Division of the USC Information Sciences Institute to assume the editorship and publishing responsibilities (under the direction of the IAB). Later, Bob Braden took over the role of project lead, while Joyce K. Reynolds continued to be part of the team until October 13, 2006.
The RFC Editor assigns each RFC a unique serial number. Once assigned a number and published, an RFC is never rescinded or modified; if the document requires amendments, the authors publish a revised document. Therefore, some RFCs supersede others; the superseded RFCs are said to be deprecated, obsolete, or even obsoleted (sic). Together, the serialized RFCs compose a continuous historical record of the evolution of Internet standards and practices.
Note that the term RFC is not unique to this series. Several other organizations have published documents using the term RFC. However, the IETF RFCs are by far the best-known RFC series on the Internet.
The RFC production process differs from the standardization process of formal standards organizations such as ISO. Internet technology experts may submit an Internet Draft without support from an external institution. Standards-track RFCs are published with approval from the IETF, and are usually produced by experts participating in working groups, which first publish an Internet Draft. This approach facilitates initial rounds of peer review before documents mature into RFCs.
The RFC tradition of pragmatic, experience-driven, after-the-fact standards authorship accomplished by individuals or small working groups has important advantages over the more formal, committee-driven process typical of ISO and national standards bodies.
Most RFCs use a common set of terms such as "MUST" and "NOT RECOMMENDED" (as defined by RFC 2119), Augmented Backus–Naur Form (ABNF) (as defined by RFC 5234) as a metalanguage, and simple text-based formatting, in order to keep the RFCs consistent and easy to understand.
For more details about RFCs and the RFC process, see RFC 2026, "The Internet Standards Process, Revision 3".
One may retrieve almost any individual, published RFC, like RFC 5000, via a URL in the form of the following example: http://www.rfc-editor.org/rfc/rfc5000.txt
Every RFC is submitted as plain ASCII text and is published in that form, but may also be available in other formats. However, the definitive version of any standards-track specification is the ASCII version.
For easy access to the metadata of an RFC, including abstract, keywords, author(s), publication date, errata, status,
and especially later updates, the RFC Editor site offers a search form with many
features. A redirection sets some efficient parameters, example:
Not all RFCs are standards. Each RFC is assigned a designation with regard to status within the Internet standardization process. This status is one of the following: Informational, Experimental, Best Current Practice (BCP), Standards Track, or Historic (sic). Standards-track documents are further divided into Proposed Standard, Draft Standard, and Internet Standard documents. The term Historic is applied to deprecated standards-track documents or obsolete RFCs that were published before the standards track was established. Only the IETF, represented by the Internet Engineering Steering Group (IESG), can approve standards-track RFCs. Each RFC is static; if the document is changed, it is submitted again and assigned a new RFC number. If an RFC becomes an Internet Standard (STD), it is assigned an STD number but retains its RFC number; however, when an Internet Standard is updated, its number stays the same and it simply refers to a different RFC or set of RFCs. A given Internet Standard, STD n, may be RFCs x and y at a given time, but later the same standard may be updated to be RFC z instead. For example, in 2007 RFC 3700 was an Internet Standard—STD 1—and in May 2008 it was replaced with RFC 5000, so RFC 3700 changed to Historic, RFC 5000 became an Internet Standard, and STD 1 is RFC 5000. When STD 1 is updated again, it will simply refer to a newer RFC that will have completed the standards track, but it will still be STD 1. Best Current Practices work in a similar fashion; BCP n refers to a certain RFC or set of RFCs, but which RFC or RFCs may change over time.
The definitive list of Internet Standards is itself an Internet Standard, STD 1: Internet Official Protocol Standards.
An informational RFC can be nearly anything from April 1 jokes over proprietary protocols up to widely recognized essential RFCs like Domain Name System Structure and Delegation (RFC 1591). Some informational RFCs form the subseries for your information (FYI). While rarely added to today, some old FYIs are still interesting, for example, FYI 18 (RFC 1983), the Internet User's Glossary. FYI 17, The Tao of IETF, is now RFC 4677, published in 2006.
An experimental RFC can be an IETF document or an individual submission to the 'RFC Editor'. In theory it is indeed experimental; in practice some documents are not promoted on standards track because there are no volunteers for the procedural details.
The best current practice (BCP) subseries collects administrative documents and other texts which are considered as official rules and not only informational, but which do not affect over the wire data. The border between standards track and BCP is often unclear. If a document only affects the Internet Standards Process, like BCP 9, or IETF administration, it is clearly a BCP. If it only defines rules and regulations for Internet Assigned Numbers Authority (IANA) registries it is less clear; most of these documents are BCPs, but some are on the standards track.
The BCP series also covers technical recommendations for how to practice Internet standards; for instance the recommendation to use source filtering to make DoS attacks more difficult (RFC 2827: "Network Ingress Filtering: Defeating Denial of Service Attacks which employ IP Source Address Spoofing") is BCP 38.
A historic RFC is one that has been made obsolete by a newer version, documents a protocol that is not considered interesting in the current Internet, or has been removed from the standards track for other reasons. Some obsolete RFCs are not classified as historic, because the Internet standards process generally does not allow normative references from a standards track RFC to another RFC with lower status. Also, few are interested in working through the required procedural details to get RFCs classified as historic and update all RFCs normatively depending on it.
Status unknown is used for some very old RFCs, where it is unclear which status the document would get if it were published today. Some of these RFCs wouldn't be published at all today; an early RFC was often just that: a simple request for comments, not intended to specify a protocol, administrative procedure, or anything else for which the RFC series is used today.