An archive is judged on whether a citation still works a decade later. That is a hosting promise more than an editorial one.
Archives are quiet sites with enormous datasets, which is the opposite of what most hosting advice assumes.
A parish register or a photograph collection is hundreds of thousands of items with structured fields. Those are indexed for the fields researchers filter on rather than treated as ordinary posts.
Researchers go deep into result sets in a way ordinary visitors never do. Deep pagination is handled with cursors instead of offsets, which is why page four thousand does not time out.
Archive images are enormous because they have to be. Derivatives are generated on ingest so a researcher can browse thumbnails and still zoom into the original.
A record URL appears in published research and stays there. Permanent URLs and redirect history are treated as part of the archive rather than as site configuration.
The collection is often the only digital copy. Backups are the product here, not an insurance policy attached to it.
Archives get modest visitor numbers and hold more rows than most ecommerce stores. Sizing on traffic alone gets this exactly backwards.
Nobody tests this and researchers do it daily. Offset pagination degrades linearly with depth, which is why archives feel fine until somebody actually uses them.
The defining constraint is a dataset far larger than the traffic, and a promise that URLs outlive the site's redesigns.
| Setting | What we do | Why |
|---|---|---|
| Record indexing | Indexed on the fields researchers actually filter | Collections are searched by date, place, name and type in combination, which is nothing like how a content site is queried. |
| Deep pagination | Cursor-based rather than offset-based | Offset paging gets linearly slower with depth, and researchers routinely go thousands of pages in — the exact case nobody tests. |
| Scan derivatives | Generated at ingest, including reading-resolution sizes | Archive masters are deliberately enormous, so producing browse sizes on request makes the first researcher to open an item wait for all of it. |
| URL permanence | Redirect history preserved as part of the collection | Record URLs are cited in published work, and a redesign that changes them silently breaks references nobody can fix afterwards. |
| Backup role | Treated as preservation rather than as insurance | For most archives the site holds the only digital copy of the material, which makes the backup the artefact rather than a safeguard for it. |
Every plan includes free migration, daily backups, SSL and 24/7 support.
Move the archive across — migration is free, every record included.