Glossary

Abandoned — A probe down long enough that nobody expects it back. The end of the probe status sequence.

Anchor — A larger, better-connected probe that also serves as a measurement target. See Anchors.

Anchoring mesh — The standing set of anchor-to-anchor measurements; a near-all-pairs latency matrix. See The anchoring mesh.

AnyBest — A 0–100 per-protocol confidence score that a prefix is anycast, from the upstream anycast feed. See Anycast.

Anycast — A prefix announced from several places at once, so one address answers from different locations depending on where you ask.

AS / ASN — Autonomous system, and its number. The unit of internet routing policy.

Catchment — The set of locations where an anycast prefix was observed answering.

Change-point — A record written only when something meaningful changed, rather than on every sample. How uptime is stored.

Connection event — A probe connecting to or disconnecting from RIPE's infrastructure. Measurement 7000.

Credits — RIPE Atlas's currency: earned by hosting probes, spent running measurements. Only your dashboard deals in them.

Dark — A country with no connected probe. See Deployments.

Data delay — RIPE's own ingestion lag, distinct from Atlas's archive lag. See Data delay and freshness.

Derived check — A monitoring check that reads data Atlas already has. Costs nothing; contrast UDM check.

DNSMON — A standing anchor-based measurement set watching root and TLD name servers.

Flag — An issue raised against a specific record. See Issues.

Flapper — A probe that connects and disconnects constantly, distorting aggregate statistics. Excludable in missed disconnects.

Fragile — A country down to a single connected probe: one unplug from dark.

Geofeed — An operator-published CSV mapping their netblocks to locations (RFC 8805). The strongest geolocation source Atlas has. See Geofeeds and geolocation.

GeoNames — The reference database behind Atlas's city names and coordinates.

Great-circle dispersion (GCD) — How geographically spread an anycast prefix's observed catchments are.

IXP — Internet exchange point, where networks interconnect. Sourced from PeeringDB.

Measurement 7000 / 7001 — RIPE built-in measurements for connection events and probe uptime respectively. The two series most of Atlas's health analysis rests on.

Missed disconnect — A disconnect event RIPE dropped, detectable as two consecutive connects for one probe. Inflates naive uptime.

Partial prefix — An anycast record where only part of the prefix is actually anycast; it names a backing prefix as the covering announcement.

PeeringDB — A self-reported database of networks, IXPs and facilities. Refreshed daily; says where operators claim presence, not where traffic goes.

Probe — A small device a volunteer hosts, from which measurements run. See Probes.

Region — A user-defined group of countries, usable as a filter across the site. See Locations.

RIS / RIS Live — RIPE's routing information service, and its live BGP update feed.

Roll-up status — A monitoring target's status: the worst of its checks, ignoring paused and disabled ones.

RPKI — Cryptographic validation of route origin; a per-prefix status in IP lookup and a monitoring check.

Scope — On an Atlas API key, read or write. Every key has read. See API keys.

Terminal status — A measurement that has stopped, failed, been forced to stop, been archived, or found no suitable probes. Terminal results are cached locally; ongoing ones are read live.

UDM — User-defined measurement: one a user created, costing them credits. Contrast RIPE's built-in measurements.

UN/LOCODE — UNECE codes identifying locations unambiguously, where a city name would not.

Viewer / staff / admin — The three Atlas roles, in increasing order. See Accounts and roles.

Well-known measurement — One RIPE tags system-well-known-target: all-probes runs against significant destinations, refreshed hourly and readable by everyone.