Topic
Network-resource Evidence
Within the Topic facet, Network-resource Evidence topic intelligence connects articles that share a specific subject, signal focus, or monitoring theme. The page gives readers a richer path through related reporting, source evidence, market actors, and infrastructure implications, with enough context to understand why the topic matters across company movements, governance decisions, regional exposure, and operational risk. Readers can compare recurring signals, affected organisations, public evidence, market context, service continuity, procurement, competition, compliance, and strategic planning questions behind the subject instead of stopping at a thin list of matching articles. It explains what the topic covers, which infrastructure actors or policies are involved, what evidence supports the coverage, and why the subject may matter for operators, customers, investors, and policy readers.
IETF
The Scheduler Was Fair to Flows. One Customer Had Opened Eight.
FQ-PIE can give competing packet flows separate queue state and scheduled service while still answering the wrong fairness question for an operator. The mechanism sees five-tuples and finite buckets; leadership may care about customers, applications, tunnels, networks or service…
IETF
The Reply Found Its Port; the Next Request Could Still Vanish
RFC 3581 repaired a precise SIP failure: a response could know the right translated IP address and still be sent to the wrong port. Its `rport` parameter lets the server answer through the opening made by the request. That success belongs to one transaction. It does not turn the…
History
The Counter Reached Three. Collation Decided What Three Meant: RFC 3381
A progress number looked objective until the printer changed the order of sheets, copies and documents beneath it.
History
One SIP Reply Could Hide More Than One Message Delivery: RFC 3428
In December 2002, the SIP `MESSAGE` method made a quiet distinction between what traveled downstream and what came back upstream: one final response did not guarantee that only one terminal had received the message.
IETF
The Audio Played. The Extension Had Never Taken Effect.
An Opus packet can survive an optional extension being unknown, malformed or deliberately ignored. That continuity is the point of the design—and precisely why audible base output cannot certify the enhancement around it.
IETF
The Line Went Quiet; the Session Had Not Ended
RFC 3573 solved a narrow dial-up problem with a durable systems lesson: a temporary condition, the policy chosen in response, and the eventual service result are three different records. Its modem-on-hold message tells a remote session server what the access edge observed. It…
History
The Ethernet Frame Crossed the IP Network. The LAN Did Not Come With It: RFC 3378
Two distant Ethernet segments could exchange one familiar frame, but the tunnel did not make their local assumptions true at a distance.
History
One Unsupported Attribute Could Cancel the Whole Job Change: RFC 3380
The printer was allowed to say no to the entire edit. In 2002, RFC 3380 made a multi-attribute change a single decision: accept a valid resulting job state, or leave the job as it was.
History
The Registrar Could Manage the Domain. The Registry Still Ruled the Namespace: RFC 3375
The customer saw one company at the counter, but the durable authority sat in a different system behind it.
IETF
The Best Surrogate Was Chosen for the Resolver, Not the Client
A request-routing system can make a perfectly reproducible decision about the wrong subject. RFC 3568 exposes the substitution: DNS commonly reveals the address of a client-site resolver, and recursion can replace even that address with another name server. The system may then…
IETF
The Server Said Nothing Had Changed. Another Authority Was Already Serving Something Else.
The resolver received a tiny “nothing new” signal and kept the large signed RRset already in its cache. That saved a retrieval. It did not establish that every authoritative server held the same generation—or that the cached entity still represented the operator's current intent.
History
A Link Could Name a Record Without Naming It: RFC 3368
RFC 3368 tried to make a common name travel as a link. Its two URI forms quietly assigned the service question to different places: the address could choose one resolver, or leave the choice to the client.
History
The Call Had Two Truths: What SS7 Said and What SIP Could Route: RFC 3372
A gateway could preserve every legacy signaling byte and still send the call the wrong way. It could also route the call correctly while losing the feature that made the call meaningful.
IETF
One Packet Vanished. The Damage Arrived in Instalments.
An interleaver can turn one lost datagram into several separated 20-millisecond holes instead of one continuous wound in speech. That sounds like recovery. It is not. RFC 3558 buys a different loss shape by making the receiver hold more state and wait longer before it knows what…
History
The Protocol Standardized the Question. It Left Who Gets to Answer Outside: RFC 3367
Two browsers receive the same phrase from a user. Both produce a valid CNRP query. Each sends it to a different configured service, and each receives a different resource. Nothing in the exchange is malformed. The disagreement entered before the protocol began.
IETF
The Update Was Accepted. The Parent Had Not Published It Yet.
The child received `NOERROR` from the delegation update receiver and closed its change ticket. Minutes later, an authoritative parent server still returned the previous delegation. Nothing in that constructed sequence contradicts the draft: acceptance means the change should…
History
The Query Asked for English. The Service Heard a Hint: RFC 3367
In 2002, the IETF standardized a way to ask a deceptively human question: which record matches this name, in this language, for this place or category? RFC 3367 made those preferences legible across services—but left the meaning of a good answer to each service.
IETF
The Packet Saved Three Headers. It Lost Eighty Milliseconds at Once.
Four 20-millisecond feature pairs shared one RTP packet. The packetizer paid for one header instead of four. When that datagram disappeared, the speech engine did not lose a header-efficiency statistic; it lost one continuous 80-millisecond stretch of evidence. RFC 3557 makes…
History
The Address Could Be Assembled in Pieces. Every Piece Added a Place to Fail: RFC 3363
The resolver had three parts of the address and no answer. A6 had made each prefix independently replaceable; it had also made the last missing organization part of the critical path. RFC 3363 records the moment the IETF treated that trade as an experiment rather than a…
History
The Label Named a Fax Stream. It Did Not Make the Fax Interoperable: RFC 3362
A decoder can recognize `image/t38` perfectly and still know almost nothing about whether the next page will arrive. RFC 3362 mattered because it supplied a common name at one boundary—and because its text did not pretend that the name operated the rest of the path.