Internet Reliability
Internet Reliability is a slice of Internet Architecture. It is broader than network uptime: it includes whether people and tools can still find and use the thing they meant after names, hosts, and paths move.
People usually arrive with a rename that broke half their docs, or a URL that went cold. HCI Nerdz writes that pain and the diagnosis in ordinary language — start with Broken links after a URL rename. This hub zooms out: how the same mistake scales from a single machine to a developer ecosystem to the public internet.
What counts as reliability here
| Layer people mean | What Internet Reliability adds |
|---|---|
Packet / host uptime |
Necessary, not sufficient — a 200 on the wrong identity is still a broken reference |
Docs and API URLs |
Reference durability: citations, |
Developer experience |
Ecosystem reliability: renames are search-and-fix under known consumers, not org-wide archaeology |
Human usability of the web |
The internet remains usable as a shared memory, not only as a transport |
Map
-
Three altitudes — machine ↔ DevX/ecosystem ↔ internet
-
Reference reliability — URL rot as the familiar slice → HCI An alternative to URLs
-
Identity stack — CAS, mutable pointers, reverse consumers → HCI + connectome-fs
-
Near-term absorb — redirects, aliases, Cool URI,
CONSUMERS.md -
NDN and friends — delivery-name vs application-label
-
Provider search and mirror backups — search APIs + API-compatible consumer mirrors (sibling under Internet Architecture)
Related reading
| Piece | Where |
|---|---|
Feeling the breakage |
|
Diagnosis and treatment (HCI) |
|
Filesystem / graph substrate |
(If you came looking for menu-style install guides titled Navigating by content, that is a different HCI essay — not this reliability topic.)
Diagrams on this hub
Internet Reliability pages use build-time Mermaid and PlantUML (via Kroki → SVG) plus hand-authored SVG mockups. See Antora diagram formats on the docs hub.