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.
History
The Missing Segment Did Not Stop the Next One: How RDP Separated Reliability from Order
In 1984, the Reliable Data Protocol made a choice that still unsettles casual descriptions of transport: a later message could be safely received, individually acknowledged and even handed to an application while an earlier message was still missing. Reliability described whether…
History
The Name Was Not the Address: How RFC 814 Kept Identity Separate from Route
In 1982, a stale host table could send queued mail to an unexpected machine after the intended host had moved. RFC 814 treated that failure as more than bad data. It exposed four different references—name, address, route and port—and argued that an Internet host should translate…
CASE FILE
The Request Signature Survived. The Certificate Could Still Change: RFC 10002 and the Authority Inside CMC
RFC 10002 makes certificate enrolment look less like one approval and more like a chain of bounded decisions. An end entity can sign an intact request, an RA can attest to identity or key possession and request changes in an outer layer, and a CA can apply policy before issuing a…
CASE FILE
The Probe Took the Same Path. It Still Did Not Share the Queue: RFC 10014 and the OAM Evidence Boundary
An operations team can truthfully report that its probe crossed the same nodes and links as customer traffic while missing the congestion that harmed that traffic. RFC 10014 explains why: packet participation, topological path and forwarding treatment are different properties.…
History
The Acknowledgement That Stopped at the Link: How PPP Localized Reliability
In 1994, PPP acquired an optional way to number, acknowledge and retransmit frames across one link. RFC 1663 made that recovery loop precise—and kept its authority deliberately narrow. A returned acknowledgement could establish progress between two adjacent peers; it could not…
Story
ARIN Links ROAs to IRR Objects. Only the ROA Side Has a Visible Action Log
ARIN has made it possible for one routing-security action to leave two related records. That convenience is real. The evidentiary problem begins when the records later separate: the documented public history can explain the ROA, but not yet the whole operation that created…
History
The Hierarchy Was Really a Graph: How Gopher Put the Next Server Inside Every Menu Line
Gopher made a scattered collection of autonomous machines look like one calm hierarchy. The illusion did not come from a global catalogue or a persistent session. It came from a modest menu line that separated what a reader saw from what a client had to do next: interpret a type…
History
The Report That Could Not Declare the Link Bad: How PPP Kept Quality Policy Local
A PPP peer could say how many packets and octets it had sent, and return what it had seen from the other direction. It could not make the protocol pronounce the link acceptable. Link-Quality-Reports standardized a shared accounting mechanism while leaving the threshold, the…
CASE FILE
The Server Returned Four Batches. It Had Not Frozen the Mailbox: RFC 10022 and the UIDBATCHES Closeout Boundary
An IMAP server can answer `UIDBATCHES 2000` with four orderly ranges and a final `OK`. That answer is useful precisely because it is less than a snapshot: the endpoints may name no stored message, deletions may thin a range, new mail may rise above it, and no later FETCH, COPY or…
History
One Link Was Actually Several: How PPP Multilink Kept a Bundle in Sequence
A second line could join a PPP connection without beginning a second network conversation. Multilink PPP treated the lines as members of one bundle, cut a packet into independently framed fragments, and gave the receiver just enough shared order to put the packet back together.…
CASE FILE
The Card Was Removed. Its Configuration Stayed: RFC 10016 and the Authority of System State
RFC 10016 gives device-supplied configuration a standard place to be seen. It also exposes a harder operational truth: the device can withdraw a system node while client intent remains, and a client can withdraw an override only to uncover the system value beneath it. The four…
History
The Success That Ended Its Own Stream: How XMPP Restarted After TLS and SASL
In XMPP, a successful security negotiation did not authorize the old XML stream to continue. It made that stream obsolete. The parties kept the same TCP connection, discarded context that no longer deserved trust, exchanged new stream headers and learned again which features were…
History
The Characters the Checksum Never Saw: How PPP Filtered the Serial Path Before Trusting the Frame
PPP did not ask its checksum to remember every character that crossed a serial path. It first defined which bytes belonged to the frame, which were temporary disguises, and which low control characters intervening equipment might have inserted. Only after the receiver reversed…
CASE FILE
The Tree Was Active. One Leaf Still Had No Packet: RFC 10018 and the P2MP Closeout Boundary
The Tree Was Active. One Leaf Still Had No Packet: RFC 10018 and the P2MP Closeout Boundary intelligence summary explains the development, the public evidence available to readers, the organisations involved, the regional context, market exposure, and the infrastructure…
CASE FILE
The Endpoint Still Spoke TLS 1.2. Its Old Handshake Had to Go: RFC 10015 and the Retirement Boundary
RFC 10015 does not switch TLS 1.2 off. It makes a more exacting intervention: keep the protocol version where it is still needed, but remove the finite-field Diffie-Hellman and RSA key-exchange paths whose risk can no longer be justified. The hard part begins after that sentence…
History
The Header That Appeared Only When It Saved Bytes: How IPComp Made Compression Conditional
IPComp negotiated a common way to decompress traffic, then allowed each packet to decline the offer. If compression plus its four-byte header did not make that packet smaller, the correct wire format was the original packet with no IPComp header at all. Capability lived in the…
History
The Datagram Relay That Died with a Stream: How SOCKS5 Bound UDP to a TCP Association
A SOCKS5 relay could still be receiving perfectly formed UDP datagrams when its TCP control connection vanished. The specification nevertheless ended the association. UDP had no farewell to offer, so the protocol borrowed a lifetime from a different transport—and kept that…
History
The Key That Could Not Lock the Tunnel: How GRE Separated Flow Identity from Security
GRE called a four-byte field a Key before the specification could say what that key opened. Six years later, the standards-track repair gave it a smaller and more useful job: distinguish one logical flow inside a tunnel. The name survived. The claim of security did not.
Europe and Middle East Regional ISP Trends
SeaExpress’s AS31444 makes route visibility easier than route-authorisation accountability
SeaExpress’s AS31444 makes route visibility easier than route-authorisation accountability intelligence summary explains the development, the public evidence available to readers, the organisations involved, the regional context, market exposure, and the infrastructure…
Story
Four Counterparty Pairs Made 89% of RIPE NCC's July IPv4 Transfer Total
RIPE NCC's August member update says 5,935,616 IPv4 addresses were transferred in July, 4,159,744 more than in June. The public transfer tables reproduce both numbers exactly. They also change the meaning of the headline: four repeated from-to pairs account for 5,299,712…