Time Horizon
Multi-year
Within the Time Horizon facet, Multi-year time-horizon intelligence organises articles by the period over which a signal is expected to matter. The page helps readers distinguish immediate operational changes from longer-cycle governance, investment, standards, and infrastructure shifts that may unfold across quarters or years. It connects timing assumptions with public evidence, related actors, market context, customer exposure, policy pressure, and infrastructure planning so readers can judge whether a development is urgent, strategic, or still waiting on confirming evidence. The page also explains how time horizon changes the meaning of a signal, which organisations may be exposed, and which infrastructure decisions require short-term action or long-cycle monitoring.
History
The Relay Changed the Message. It Had to Call Itself a Gateway: RFC 1344
A carrier can claim neutrality while the entity in its custody remains untouched. The claim becomes harder after it changes an image, replaces a reference with a copy or divides one message into several. RFC 1344 gave that moment a name in 1992: the intermediary was no longer…
History
The Test Bed Could Demonstrate a Future. It Could Not Decide the Public Network: RFC 1259
RFC 1259 imagined the NREN as a new subdivision built before the surrounding city. Its mistakes could teach later builders, but its streets, prices and residents could not decide the whole metropolis. The crucial Internet-history lesson was not prediction. It was the design of a…
IETF
Eric Rescorla and the Encrypted Hello That Could Not Name the Server
An encrypted ClientHello can make a familiar public clue disappear from the first TLS flight. That does not turn the remaining handshake into an identity ledger. RFC 9849, co-authored by Eric Rescorla, specifies a carefully bounded privacy envelope: it can hide the private…
History
The Network Did Not Have to Keep the Beat. The Receiver Rebuilt It: RFC 1257
In 1991, a voice sample could arrive early, the next one late, and the listener could still hear a regular cadence—if the receiver knew the deadline, kept enough memory and woke at the right time. RFC 1257 used that construction to move a costly promise out of every subnetwork…
History
The Packet Got a New Lifetime at Every Border: RFC 1326
A tunnel is often described as a clean bargain: put one protocol inside another, cross a network that understands the outer wrapper, then recover the original packet at the exit. RFC 1326 asked what happens when two such bargains face each other and routing sends the packet back…
History
The Analysis Could Not Stand In for the Test Network: RFC 1245 and RFC 1246
In July 1991, OSPF arrived with more than a protocol specification. Its supporting case was deliberately divided into an analysis and an experience report. That division preserved a distinction Internet engineering still needs: a model can bound expected behaviour, but only…
History
The NIC Owned the Question. It Did Not Own the Network: RFC 1302
A help desk can keep faith with a user without pretending to control every system behind the problem. RFC 1302 made that distinction unusually clear in 1992. A Network Information Center was expected to take responsibility for an inquiry until it was resolved in some way. Yet…
IETF
Deborah Brungard and the Transport Profile That Did Not Choose a Network
A standards requirement can make a capability available without dictating the network that will use it. RFC 5654, edited by Deborah Brungard and four colleagues, describes requirements for an MPLS Transport Profile. Its most important restraint is plain: the requirements concern…
History
The MIB Had to Redraw Who Could Change AppleTalk: RFC 1243 and RFC 1742
A management model is also a map of responsibility. In 1991, the first AppleTalk MIB let an SNMP manager write route fields and zone mappings while reporting how port settings had been obtained. Four years later, its replacement moved several of those lines. The revision matters…
History
The Box Dropped No Frames. It Had Not Passed the Other Tests: RFC 1242
A clean throughput trial is seductive because it ends in one number. RFC 1242 gave that number a strict meaning, then surrounded it with other terms that refused to disappear. Zero loss at one offered rate did not answer latency, burst tolerance, overload, restart, filtering or…
History
The Counter Was Total. Subscription Discards Were Somewhere Else: RFC 1304
A number labelled total looks like the end of an investigation. In RFC 1304, it was the beginning of one. The 1992 MIB could count SIP packets discarded for protocol or bit errors, list other address-related failures and remember the latest disagreement of each syntactic type.…
History
The Tunnel Carried the Packet. The Error Lost the Question: RFC 1241
A tunnel can look transparent in the forward direction and become opaque the moment something fails. RFC 1241 drew that asymmetry with unusual precision. Its outer packet could carry an untouched IP datagram through a separate routing space. But an ICMP error from inside that…
NPNOG
Broken Dates, Empty Lists and Missing Slides: Can npNOG Preserve Its Own Memory?
npNOG’s public websites preserve a useful record of programmes, teaching materials and event continuity, but they also contain date conflicts, unfinished download promises and pages whose headings outlive their visible data. Those gaps do not prove that an event failed, a…
History
The MIB Arc Changed. Vendors Still Had to Revise: RFC 1239
Nothing in a management console announces that a number has become architecture. RFC 1239 did. Five MIBs kept describing familiar interfaces, yet their entity identifiers moved from an experimental branch to standard branches. The standards record changed in two pages; every…
North America Institutional Trends
AtaiBeckley’s Lilly Deal Has Two September Gates and a Seven-Year CVR Tail
Two timestamps, 47 hours apart, now organize Eli Lilly’s proposed purchase of AtaiBeckley. Stockholders are due to vote on 8 September. Australia’s merger waiting period is scheduled to expire on 10 September. Neither event is the closing, and neither turns the headline “US$6.75…
History
The Master Said Accepted. The Applications Had Not Reported an Outcome: RFC 1301
Twelve small verdicts travelled inside every Multicast Transport Protocol packet. One could say accepted, another pending, a third rejected. They looked like receipts, but they were not votes from twelve receivers. RFC 1301 gave one master the right to set those values after…
History
The Agent Spoke with One Voice. The Subtree Had Another Owner: RFC 1227
To a management station, an SNMP agent looked like one speaker. RFC 1227 placed a switchboard behind that voice: local processes registered branches of the MIB, priorities chose which process answered, and a broader branch could hide a narrower one. The reply was real, but its…
North America Datacenter Trends
Fervo’s 396MW Google PPA Comes in Four Tranches—and the 600MW Option Has a Reimbursement Tail
The nearly-one-gigawatt headline hides four different contractual states. Fervo Energy has signed a 396MW power purchase agreement with Google, divided into four 99MW deliveries. It must also offer roughly 600MW more, but that capacity becomes contracted only if Google accepts…
History
The Trap Carried 0.0.0.0. The Envelope Named the Agent: RFC 1298
An SNMP trap crossed an IPX network with four zeroes where its agent address might have been expected. That was not missing data by accident. RFC 1298 deliberately emptied the Trap-PDU field and told the receiving manager to take the source from the transport layer instead. The…
History
The Radio Frame Entered IP. Its Physical Edges Did Not: RFC 1226
A packet can cross a boundary without carrying every trace of the medium it came from. RFC 1226 put an AX.25 frame inside one IP datagram, discarded the HDLC marks that had framed it on a serial link, and retained the checksum. That choice made a useful tunnel format—and a…