The Centre de Données Astronomiques de Strasbourg occupies the second floor of a turn-of-the-century building on the rue de l'Université in central Strasbourg, France. The building also houses the Strasbourg Observatory, which celebrated its centenary in 1981.
On the afternoon of April 22, 2026, the Centre's director, an astronomer named Mireille Lavergne, walked a visiting reporter through a series of small offices and explained what a data centre actually is.
It is, in the case of the CDS, a collection of about forty staff, three software platforms, and the patient accumulation of fifty-four years of astronomical catalogues. The Centre was founded in 1972 by a small group of French astronomers who recognised that the world's stellar catalogues were scattered across dozens of institutions and were already becoming difficult to find.
The Centre's three main services are SIMBAD, a database of identifications and references for individual astronomical objects; VizieR, a service that holds catalogues in queryable form; and Aladin, an interactive sky atlas.
Of the three, VizieR is the closest in spirit to what an older astronomer would call a catalogue. It currently holds more than twenty-three thousand catalogues, from the Bonner Durchmusterung of 1859 to last week's data releases from the various ongoing surveys. It is queried, on average, eleven million times per month.
Each catalogue is held in a standard form. The columns are documented. The units are explicit. The astrometry is referred to a standard reference frame, currently the ICRS realised by Gaia DR3. The bibliography is linked to the Astrophysics Data System maintained at Harvard.
Lavergne, who has been with the Centre since 2003, says the work is mostly invisible. Astronomers use VizieR the way they use electricity. They turn it on; the catalogue appears; they use it; they do not, generally, think about who maintains it.
The maintenance is, in fact, considerable. Each new catalogue submitted to VizieR is reviewed by a curator. The curator checks the units, the column descriptions, the cross-identifications with previous catalogues. The review takes, on average, about two weeks per catalogue. The Centre receives roughly four hundred new catalogues per year.
Some are small. A catalogue of binary star orbits with twelve entries from a single observatory in Brazil takes two days to ingest. Some are very large. The full Gaia DR3 catalogue, when it was loaded in June 2022, took three months to ingest, validate, and cross-match against the existing archive.
The cross-matching is one of the Centre's quieter contributions. SIMBAD, in particular, maintains a list of every identifier that has ever been used for every astronomical object. The star known to amateurs as Vega is also Alpha Lyrae, HR 7001, HD 172167, HIP 91262, BD+38 3238, and approximately three hundred and twenty other designations.
SIMBAD knows them all. It also knows which of them refer to the same physical object and which, due to coordinate errors or transcription mistakes, were once thought to refer to the same object but actually do not.
The Centre's archives include some material that is no longer scientifically current but is preserved for historical reasons. The 1862 catalogue of Friedrich Wilhelm Argelander, the Bonner Durchmusterung, is still queryable in its original form. So is the Cape Photographic Durchmusterung of 1896. So is the proper-motion catalogue compiled by Heinrich Louis d'Arrest in the 1860s.
These are consulted, on average, about two hundred times per year. They are not high-traffic data. They are kept available because the Centre's policy, formalised in 1996, is that no catalogue once ingested will be removed from the archive.
The policy has consequences. The archive grows. The hardware required to serve it grows. The energy bill grows. The Centre's annual budget, paid jointly by the Centre National de la Recherche Scientifique and the University of Strasbourg, has approximately doubled in the last fifteen years.
It remains, by any reasonable measure, a bargain. The Centre's services support, according to a recent internal estimate, roughly thirty percent of all professional astronomy papers published worldwide. The cost per supported paper is somewhere in the low double digits in euros.
Amateur astronomers use the Centre's services freely. The most common use, by some distance, is the identification of an object photographed by an amateur with a small telescope or a modest CCD. The amateur uploads coordinates; Aladin returns an image overlaid with catalogued objects within the field; the amateur knows what they have photographed.
Lavergne is asked, occasionally, what would happen if the Centre stopped operating. She says, mildly, that it would be inconvenient.
The inconvenience would be considerable. The Centre's competitors, of a sort, are the NASA/IPAC Extragalactic Database in Pasadena and the Astronomical Data Centre archives at various national observatories. None of them, by themselves, hold the breadth of what is at Strasbourg. Replicating the work would take, by Lavergne's estimate, twenty years and several hundred million euros.
The Centre does not plan to stop operating. Its current strategic plan, published in 2025, runs to 2030 and includes the ingestion of the Vera C. Rubin Observatory's first three data releases, the Euclid mission's wide and deep surveys, and a refresh of the SIMBAD identification rules to handle the volume increase.
It also includes, in a small paragraph in the appendix, a commitment to maintain a paper-based finding aid to the historical catalogue archive. The paper aid, kept in a metal cabinet on the second floor, is the only way to locate certain materials that predate digital cataloguing.
It is updated, by hand, by a librarian named Élise Mounier, who has worked at the Centre since 1991. She uses a fountain pen with iron-gall ink. She has, in thirty-five years, made very few errors. The corrections, when they occur, are crossed through with a single line, the way Furness corrected her measurements at Vassar in 1909.
Some things, in catalogue-making, do not change.
