A working group, not a company

The 802.11 standard is the product of a volunteer committee. Understanding who votes, and how, explains why Wi-Fi evolves the way it does.

What IEEE 802 actually is

The number 802 is not a product code. It is the designation of a committee — formally, IEEE Project 802, the Local and Metropolitan Area Networks Standards Committee — convened for the first time in February 1980, which is where the digits come from. Beneath it sit working groups, each assigned a further number: 802.3 for Ethernet, 802.15 for personal area networks, 802.11 for wireless LANs. The working group is the operative unit. It holds meetings, circulates drafts, runs ballots and publishes the document that chip designers, firmware writers and certification bodies eventually use.

IEEE itself — the Institute of Electrical and Electronics Engineers — is a professional society headquartered in Piscataway, New Jersey, with members in roughly 160 countries. It writes no products and sells no spectrum. What it sells, in the commercial sense, is access to the finished standards document, though for technical people the drafts circulate freely enough through working group membership. The organisation that actually certifies products as interoperable is the Wi-Fi Alliance, a separate industry body that formed in 1999 precisely because having a ratified standard was not the same as having gear that worked together. The two bodies are sometimes confused; they are constitutionally distinct.

The 802.11 Working Group sits inside the IEEE Standards Association (IEEE SA), which maintains the procedural rules. Membership is open to anyone willing to pay the participation fee and attend the meetings, which run three times a year, typically in January, May and September. Corporate engineers dominate because their employers cover travel and fees, but the structure is formally individual: each person votes their own ballot, not their employer's. In practice, a company with twenty engineers in the room has twenty votes, which shapes whose proposals survive.

How a draft becomes a standard

The lifecycle of an 802.11 amendment follows a defined sequence. A task group — a subset of the working group — is formed around a specific technical scope, assigned a letter suffix (hence 802.11ax, 802.11be, and so on) and begins producing a draft document. Drafts are numbered from D0.1 upward; by the time a document reaches D5 or D6 it is usually nearing ballot readiness, though some task groups have produced a dozen drafts before closing.

Ratification requires a ballot in which working-group members vote approve, disapprove or abstain, with written comments mandatory on any disapprove vote. The rules require that disapprove comments be addressed — either by changing the text or by formally resolving the comment as not-persuasive — before the ballot can close. This comment-resolution process is where most of the time goes. An amendment that clears its working-group ballot then moves to a sponsor ballot, open to a broader pool of IEEE SA members, then to RevCom (the Standards Review Committee), which recommends final approval to the IEEE SA Board of Governors. From task-group formation to final publication, the elapsed time is rarely under four years and often closer to six.

The ordering of amendments in the public record follows task-group formation, not completion. That is why 802.11ac and 802.11ad were developed in parallel from roughly 2008 through the early 2010s even though they address different bands entirely — ac in 5 GHz, ad in 60 GHz — and why 802.11ax took longer to ratify than its letter position might suggest, finally clearing in February 2021 after a task group that opened in 2014. The letters are sequence numbers in the queue, not a ranking of importance or a delivery schedule.

Periodically the working group publishes a rollup: a new base document that incorporates all amendments approved since the previous base. The 2007 edition folded in eight amendments; the 2012 and 2016 editions likewise absorbed the amendments approved since the previous base. This matters because an amendment document is written as changes to the base, and reading an amendment without the base is genuinely difficult. The rollup resets the baseline and the amendment clock runs again.

Who shows up and who decides

The working group's roster has included engineers from chipmakers, device manufacturers, network operators, test-equipment companies and academic institutions. Qualcomm, Intel, Broadcom, Cisco, Apple, Samsung and many others have maintained consistent delegations. Because participation is individual but funded by corporations, the working group's direction reflects the commercial priorities of the industry's largest players — which is not a scandal but a structural fact worth understanding. A feature that serves a chipmaker's roadmap is more likely to attract the engineering hours needed to draft, defend and ballot it.

This dynamic produced some of the standard's most consequential decisions. Orthogonal Frequency Division Multiple Access (OFDMA) — the resource-scheduling technique borrowed from cellular, which allows an access point to address multiple clients simultaneously within a single transmission window — was a major addition in 802.11ax. Getting it into the standard required years of technical debate about implementation complexity, backward compatibility and the overhead of the trigger frames that orchestrate it. The feature survived because operators and access-point vendors, who stood to gain most from the efficiency improvement, maintained strong delegations and persistent advocacy through the ballot process.

The 802.11be task group, which produced what the Wi-Fi Alliance brands as Wi-Fi 7, added multi-link operation — the ability for a device to maintain simultaneous associations on multiple bands and channels — and 320 MHz channels. Both features arrived late in the draft cycle after extended debate; the 320 MHz channel in particular depends on the 6 GHz band being available, and regulators in different ITU regions have moved at different speeds on that question. The standard can specify the channel width; it cannot compel a spectrum allocation.

Drafts are numbered from D0.1 upward; by the time a document reaches D5 or D6 it is usually nearing ballot readiness, though some task groups have produced a dozen drafts before closing.

What the document is, and what it is not

The output of all this process is a specification: a document that defines frame formats, management procedures, PHY layer modulation schemes, timing requirements and a great deal else. It does not name products. It does not benchmark performance. It does not guarantee that two devices labelled 802.11ax will work together at the headline speed — that is the Wi-Fi Alliance's certification job, and certification covers a subset of the standard's optional features. Much of 802.11 is optional, which is how the document can accommodate everything from a battery-powered IoT sensor to a 16-antenna enterprise access point.

The IEEE SA publishes the current approved standards on its own database, where the publication history of every amendment and rollup is accessible. What shows there is a record of consensus — imperfect, slow, commercially influenced, but built by people with enough skin in the game to read the drafts and argue the comments. That process, not any single company's engineering team, is why the 802.11 family has a common frame structure recognisable across nearly three decades of hardware.

Read next

The rest of who actually decides

• CommitteeReleases as a clock3GPP numbers releases rather than years, and a feature's real availability date is the release it landed in plus the silicon lag.
• CommitteeCertification is a separate businessThe Wi-Fi Alliance and Bluetooth SIG certify and name; they do not write the underlying standard, which is a distinction that gets lost constantly.