Primary Domain
Internet Infrastructure
Within the Primary Domain facet, Internet Infrastructure intelligence groups reporting by primary domain so readers can follow a focused area of internet infrastructure, governance, connectivity markets, or digital capital. The page brings together related articles, public evidence, institutions, companies, people, regional exposure, operating dependencies, and market context that may otherwise sit across separate category pages. It explains the domain, the likely actor class, the market or governance context, and the source material readers should use when comparing signals. Operators, analysts, and governance readers can see how the same domain appears across events, profiles, market shifts, public-source evidence, regional dependencies, and longer-cycle infrastructure decisions over time.
Story
One Revoked Certificate, Three Browser Outcomes at APNIC 62
The certificate was on the revocation list. That was not the end of the story. Geoff Huston's APNIC 62 demonstration put the same adverse record in front of three browser policies and obtained different results—a useful warning about the distance between publishing a security…
History
The Record Was There. RWhois Could Not Find It: RFC 2167
A 1997 directory design exposed an uncomfortable difference between storing a record and putting it where a query can discover it. RWhois could refer a reader toward the responsible area; neither the referral nor an answer certified the underlying claim.
History
A Fast Bus Was Not an Ethernet: RFC 2143’s Missing Peer
The attraction was obvious in 1997: put nearby workstations on a fast SCSI cable and carry IP between them. The hard part was not fitting an IPv4 datagram into a command. It was making a peripheral bus accept an unsolicited message from another host—and finding a neighbor when…
IETF
A Batch Receipt Is Not a Security Action: The IETF’s Multi-SET Push Test
An IETF draft would send many security-event tokens in one HTTPS request. Its individual receipts may cut transport cost, but they leave a separate question for every receiving organisation: which event actually changed a local account or credential, and when?
History
The Connection Closed. Its Measurements Stayed: RFC 2140 and TCP's Host-Pair Memory
A short TCP exchange can finish before it has learned much about the route it used. In 1997, Joe Touch asked why the next exchange between the same hosts should have to forget everything the first had observed. The answer was not to preserve the entire connection. It was to…
IETF
Two “Not Ready” Reviews Put ACTN’s Packet–Optical Control Boundary on Trial
The IETF’s packet–optical ACTN draft has reached Last Call with an extensive map of controller interactions. Security and operations reviewers now ask what the map cannot settle: who trusts a topology update, and what survives when only one layer completes a change.
History
The Socket Call Survived; the Address Contract Did Not: RFC 2133
A program could keep calling `connect()` after IPv6 arrived and still misunderstand the address it was sending. RFC 2133 preserved the familiar socket verbs while moving compatibility into the structures, families and kernel state those verbs carried.
IETF
The Last Call Is for the Question, Not the Filter
The IESG has opened a final comment period on how to describe the limits of inter-domain source address validation. The October 1 deadline concerns an Informational problem statement and requirements, not approval of a universal filtering protocol. The distinction matters at the…
History
One DHCP Packet, Four Different Claims: What RFC 2132 Did Not Identify
A client could put a vendor class, a binding key and a shopping list into the same DHCP message. RFC 2132 gave each a different job. Reading them as one trustworthy portrait of a machine mistakes a configuration exchange for proof of identity or operation.
History
The Command Stayed ASCII. The Reader Needed More Than Bytes: RFC 2130
A mail server's command and the explanation a person reads after it fails can travel in the same conversation. The 1996 IAB character-set workshop insisted they were not the same design problem. Its report, RFC 2130, treated international text as a chain of explicit choices, not…
History
The Circuit Answered. The Flow Was Not Ready Yet: RFC 2129
In Toshiba’s 1997 cut-through proposal, a router could hear that its neighbour recognized a virtual circuit and still have no permission to send a particular IP flow over it. RFC 2129 put those answers in two different exchanges—and made both answers expire.
History
One Call Became Many Interfaces. Several Channels Became One Call: RFC 2127
RFC 2127 made an ISDN connection legible by separating its physical access, signaling, bearer channels and upper protocol into distinct managed interfaces. That clarity came with a warning still useful today: a topology row, a call counter, a channel allocation and a service…
History
Version 3 Stayed. The Service Changed: RFC 2126’s Compatibility Trap
An old peer could return a perfectly parseable Class 0 confirmation to a Class 2 request that offered no alternative. The TCP connection was open, the TPKT version was familiar, and the reply was still unacceptable. RFC 2126 made that difference explicit.
CASE FILE
RFC 9647 and the Missing Evidence Chain Between Babel State and Network Behaviour
A Babel route row can report that a path is selected while the operational proof behind that selection remains incomplete: the YANG datastore may show intent, the interface may appear enabled, the neighbour may exist, and the route may be marked feasible, yet the chain from…
History
Permission Came Before the Call: RFC 2125’s Two-Stage Bandwidth Decision
Two peers could agree that another line should be tried and still have no new bandwidth. RFC 2125 made that gap explicit: BAP first obtained a response to the proposed action, then a separate Call-Status message reported what happened when somebody actually placed the call.
IETF
WebAuthn’s 2026 Rejoin Handoff: The Governance Receipt Behind a New Participation Cycle
The Web Authentication Working Group’s 11 September 2026 recharter begins a new institutional phase for WebAuthn. The central challenge is not simply continuing technical discussion; it is preserving a reliable record of how an approved charter becomes active participation and…
History
RFC 2123: A Flow Record Was a Rule-Set Projection, Not the Wire
NeTraMet did not preserve traffic as such. Before a packet could become a flow row, a downloaded rule set decided which protocol mattered, which attributes survived, which direction counted as source, and how much detail the meter could afford. RFC 2123 made that editorial…
History
The Link Opened a Session, Not an Object: RFC 2122’s VEMMI Boundary
A `vemmi` hyperlink looked like an address, but its destination was not a file waiting to be fetched. RFC 2122 used the Web as a dispatch surface for a different system: a local client would open a continuing multimedia session, negotiate service and identity, and later decide…
History
The Reset Was in the Next Packet: RFC 2118’s MPPC Recovery
RFC 2118 gave Microsoft Point-to-Point Compression a continuous memory and a deliberately spare recovery signal. When packet loss broke that shared history, the receiver asked for a reset—but it was the sender’s next `FLUSHED` data packet, not a separate Reset-Ack, that…
History
The Browser Remembered the Session. It Did Not Prove the Person: RFC 2109
In RFC 2109’s model shop, a server called a browser-held value `Customer="WILE_E_COYOTE"`. Later requests returned that value while separate cookies accumulated for a rocket launcher and a shipping method. The exchange made stateless HTTP remember a sequence. It did not make the…