Skip to content
All regions operational
NorthmereData Commons
Home/About

About the consortium

Northmere Data Commons is a non-profit foundation owned by its member institutions. It keeps large open research datasets available, citable and fast to download.

Founded 201438 member institutions19 staff in Utrecht and Tarn Valley

What we do

Funders increasingly require that data behind published research stay open. The difficulty is that the data got large: a single reanalysis is 79 TB and a soil survey is 63 TB, beyond what a journal supplement or a laboratory server can hold. Northmere exists to hold such collections for the long term and to deliver them at the speed that lets people actually use them.

We are not a cloud provider, a publisher or an analysis platform. We store files, describe them well, assign identifiers, check them for corruption every quarter, and send them to anyone who asks within the policy. Analysis happens where you have compute.

Principles

Open by default
Collections are public unless an embargo or a legal restriction applies. Metadata is always CC0.
Versioned and permanent
Published versions are immutable. Corrections are new versions. Every version is kept for at least ten years, funded from a ring-fenced reserve of 7.2 million EUR.
Plain formats
We only host formats with a public specification and a free reader.
Honest cost
Egress is metered so the load is visible. Fees go to operations, never to profit.

A short history

  • Six institutions pool servers

    The first shared archive held 190 TB of genome assemblies on disks in two basements.

  • First petabyte

    Climate and imagery programmes join; the archive passes 1 PB and moves to colocation.

  • Signed manifests and transfer endpoints

    Resumable bulk download becomes the main way data leaves the commons.

  • Foundation established

    The consortium becomes a Dutch foundation with a supervisory board elected by members.

  • NMR-5 published

    The reanalysis becomes the most downloaded collection in our history.

  • 11.6 PB and four regions

    The Singapore region opens in June and the catalogue passes 2,000 collections.

Governance

A supervisory board of seven oversees a management team of four. Full members elect four board seats; associate members elect one; two seats are held by funders. The technical committee meets monthly and decides formats, limits and release schedules. Domain curators, one team each for genomics, climate and imagery, decide what the commons admits.

Annual accounts, board minutes and the preservation plan are published each spring. Our technical report NM-TR-2026-02 describes the storage design: erasure-coded object stores in four regions, quarterly scrubbing and a tape copy of every collection held at a separate site.

Some of our members

  • Skarholm Institute for Earth Systems
  • Lindqvist Genome Centre
  • Tarn Valley Observatory
  • Halvard Polar Research Station
  • Brennick University Climate Unit
  • Aurelian Space Applications Laboratory
  • Ketterdam Plant Sciences Institute
  • Sorrel Coast Oceanographic Trust
  • Marrow Fen Agricultural Institute
  • Orsted Heath Soil Laboratory
  • Vessel Bay Marine Biology Station
  • Ilmari Geodesy Centre

The full list of 38 members is in the annual report.

Operations in numbers

Storage

95 PB raw usable across four regions. Each collection is held in at least three regions plus tape. Annual measured bit-rot rate: 2 files per 10 billion.

Network

Combined egress capacity of 340 Gbit/s, donated in part by national research networks. Peak sustained load recorded: 212 Gbit/s on 3 October 2026.

People

Nineteen staff: eight engineers, six curators, two support specialists, two in finance and membership, one director. Domain curators are seconded from members.

Want to talk? Contact the office or read about how we are funded.