The Internet Engineering Task Force (IETF) is an open international community that develops technical standards and operational guidance for the Internet. Its work covers the protocols that enable independently operated networks, devices, and applications to communicate. Rather than exercising regulatory authority, the IETF produces specifications whose influence depends on adoption by software developers, equipment manufacturers, network operators, and users. Its standards process emphasizes public discussion, technical review, rough consensus, and experience with working implementations.
Scope and purpose
The IETF’s central concern is the engineering of an interoperable Internet. It develops and maintains protocols for addressing, routing, transport, naming, security, and many application services. Its work encompasses major parts of the TCP/IP protocol suite and the mechanisms that allow networks under different ownership to function together.
The IETF is not an Internet government. It does not own Internet infrastructure, compel organizations to deploy particular technologies, or regulate online content. Nor does it develop every technology used on the Internet. Physical transmission technologies, hardware interfaces, and many aspects of the World Wide Web are standardized elsewhere.
IETF specifications seek to make independently developed implementations compatible. This makes interoperability a central engineering objective: a protocol must describe behavior precisely enough that different implementations can exchange information without relying on a single vendor’s software or infrastructure.
Historical development
The IETF first met in 1986, emerging from the technical coordination community surrounding the early Internet. Its initial focus was closely connected to the operational and engineering problems of research networks. As the Internet expanded beyond research institutions, participation broadened to include commercial network operators, equipment vendors, software companies, universities, and individual contributors.
The IETF inherited and developed a culture of open technical discussion and practical experimentation. The expression “rough consensus and running code” became a widely recognized description of that approach. It emphasizes resolving substantive technical concerns and learning from implementations rather than relying solely on formal voting or theoretical agreement.
The Internet Society, founded in 1992, became an institutional home supporting Internet standards work. Administrative arrangements subsequently evolved. Since 2018, the IETF Administration LLC has provided the legal and administrative structure for functions such as finances, contracts, meetings, and operational support. The LLC’s administrative role is distinct from the community’s authority over technical standards.
Participation and organization
An open community
The IETF has no conventional organizational membership system. People participate as individuals rather than as formally accredited representatives casting votes on behalf of companies or governments. In practice, many contributors are employed or funded by organizations with substantial interests in Internet technology.
Participation commonly takes place through public mailing lists, document reviews, online meetings, and in-person meetings. Most technical work continues between the IETF’s main meetings, which are normally held three times a year. Attendance at a physical meeting is not a prerequisite for contributing to a specification.
Open participation does not mean that every proposal is adopted. Proposals must attract sufficient interest, fit an appropriate scope, and withstand technical review.
Working groups and areas
Most standards work occurs in working groups. Each working group operates under a charter defining its scope, intended deliverables, and boundaries. Working-group chairs organize discussion and assess whether the group has reached consensus.
Working groups are organized into technical areas overseen by area directors. These areas cover broad subjects such as routing, security, transport, Internet-layer mechanisms, operations, and applications, although their names and boundaries can change.
The area directors, together with the IETF chair, form the Internet Engineering Steering Group (IESG). The IESG manages the standards process, oversees working-group activity, and reviews documents proposed for publication through the IETF stream.
The Internet Architecture Board (IAB) provides architectural oversight and handles liaison relationships and other responsibilities within the Internet standards system. It is not a substitute for working-group development or IESG standards review.
A community nominating committee selects candidates for many leadership positions, subject to the applicable confirmation procedures. The process is designed to renew leadership without turning standards development into a system of corporate membership votes.
How specifications are developed
Internet-Drafts
Work usually begins as an Internet-Draft, a publicly available working document. An Internet-Draft may be an individual proposal or a document adopted by a working group.
Internet-Drafts are provisional. They can change substantially, be replaced, or expire, and their existence does not imply IETF approval. Adoption by a working group means that the group has accepted a document as a basis for work, not that its contents are already agreed.
Drafts are revised in response to discussion, implementation results, interoperability testing, and specialist reviews. Review may address security, operational effects, compatibility, congestion behavior, internationalization, or interactions with other protocols.
Rough consensus
Rough consensus is neither unanimity nor a simple majority vote. Chairs assess whether the community has adequately considered significant technical objections and whether there is sufficient agreement to proceed.
The number of supporters can inform that assessment, but counting preferences is not the governing principle. A technically important objection cannot be dismissed merely because few people raise it. Conversely, an objection does not necessarily prevent progress once it has been examined and the group has a reasoned basis for proceeding.
Implementation experience provides evidence about whether a design can work and whether its specification is sufficiently clear. However, the slogan “running code” should not be interpreted as a universal requirement that every initial standards-track publication must already have multiple deployed implementations. Formal requirements depend on the document’s type and maturity level.
Review and approval
A working-group document ordinarily undergoes a working-group last call before being submitted for wider review. Documents proposed for standards-track publication generally receive an IETF Last Call, allowing the broader community to comment, followed by IESG evaluation.
Review can lead to revisions, additional discussion, or a decision not to approve publication. Some documents proceed as individual submissions sponsored by an area director rather than through a working group.
The process also includes procedures for appealing certain decisions and process failures. These distinguish disagreement with an outcome from claims that the required process was not followed.
RFCs and standards status
The IETF publishes its completed specifications and guidance in the Request for Comments (RFC) series. However, an RFC is not automatically an Internet Standard, and not every RFC comes from the IETF.
The RFC series contains several publication streams, including IETF, IAB, research, and independent-submission streams. These have different review and approval arrangements. Publication as an independent-submission RFC, for example, does not establish IETF consensus.
Important document categories include:
- Standards Track: specifications intended to become Internet standards. The principal maturity levels are Proposed Standard and Internet Standard. Draft Standard remains a historical designation associated with the earlier three-level process.
- Best Current Practice: documents describing accepted operational practices, procedures, or institutional processes.
- Informational: material published for information rather than as an Internet standard.
- Experimental: specifications or approaches intended for experimentation.
- Historic: material designated as obsolete or no longer recommended for current use.
Once published, an RFC’s substantive text is not continually rewritten. Later RFCs may update or obsolete it, while errata record reported problems. Understanding the current specification for a protocol can therefore require reading several related RFCs and checking their status.
Major technical contributions
The IETF’s work spans much of Internet communication:
- Transport: development and maintenance of TCP specifications and the standardization of QUIC, an encrypted transport protocol built over UDP.
- Web communication: HTTP semantics and major HTTP versions, including HTTP/2 and HTTP/3.
- Secure communication: Transport Layer Security (TLS), which provides security for many Internet applications.
- Naming: development and extension of the Domain Name System (DNS), including security and privacy-related mechanisms.
- Routing: protocols such as Border Gateway Protocol (BGP) and Open Shortest Path First (OSPF).
- Messaging: standards for email, including SMTP and IMAP.
- Internet-layer evolution: IPv6 and associated mechanisms for addressing, configuration, and packet handling.
Some foundational technologies originated before the IETF or outside its working groups. Its role often involves maintaining, extending, and clarifying existing technologies rather than inventing them from the beginning.
Relationships with other institutions
Internet standardization depends on several organizations with distinct responsibilities.
The Internet Assigned Numbers Authority (IANA) maintains protocol-parameter registries used by IETF specifications, such as assigned option values and identifiers. Standards documents commonly contain an “IANA Considerations” section explaining required registry actions.
The Internet Research Task Force (IRTF) supports longer-term research through research groups. Unlike IETF working groups, these groups are not primarily organized to produce Internet standards, although research results can influence later engineering work.
The World Wide Web Consortium (W3C) develops standards for many aspects of the web platform. The division of responsibility is not simply “Internet versus web”: the IETF standardizes HTTP and other mechanisms used by the web, while the W3C addresses technologies such as document formats, application interfaces, and accessibility. Coordination is necessary where their work intersects.
Security, intellectual property, and practical limits
Security analysis is an established part of IETF specification development. Documents are expected to address relevant threats and security properties, and privacy considerations have become increasingly prominent. Protocol design can involve trade-offs among encryption, network observability, performance, deployability, and resistance to interference.
IETF policies also address intellectual property. Contributors have disclosure obligations concerning relevant intellectual-property claims under the applicable rules. Publication of an IETF standard does not itself guarantee that every implementation is free of patent obligations.
The IETF’s open process has practical limits. Sustained participation requires time, technical expertise, and often financial support. Public access to discussions therefore does not eliminate differences in participants’ resources. Consensus-building can also be slow when proposals affect deployed infrastructure or create competing operational requirements.
Finally, standardization does not guarantee deployment. Adoption depends on implementation costs, compatibility, business incentives, operational needs, and coordination among independently controlled systems. A technically mature specification may remain little used, while a widely deployed protocol may require years of incremental revision.