Architecture - Cellar



Cellar: the global picture

Cellar receives publications and their associated metadata based on the Metadata Encoding Transmission Standard (METS). A METS package is a zip package containing digital publications, associated descriptive and technical metadata and description of the organization of the FRBR WEMI hierarchy (Work-Expression-Manifestation-Item, see metadata page for details). Each METS package is validated and then the digital content and metadata are stored. The digital publication is stored in a file system and the metadata is stored in a knowledge graph. Digital publications and metadata are disseminated via a RESTful interface. Metadata can also be query via the SPARQL interface. RSS and Atom feeds are also available

Cellar statistics

  • 44 million of files with a size of 40 TB are stored in the file system.
  • 1 billion of triples and 95 million of resources are stored on the knowledge graph.
  • The size increases every day.
  • Cellar treats in average 20 million of requests per day.

Cellar ecosystem

Cellar does not live alone. It is part of the Cellar ecosystem, which includes various systems and services.

A typical workflow of receiving publications, storing and disseminating them, includes the following steps.

On ingestion part:

  1. A component named IMMCbuilder creates an IMMC package, containing digital publications and metadata.
  2. CERES is the reception and validation service that receives IMMC packages, validates the digital publications and metadata, and finally generates the METS package for Cellar storage and dissemination.
  3. Digital publications and metadata are stored in Cellar on a file system, an RDBMS database and a knowledge graph.

On dissemination part:

  1. Dissemination is taking place via EUR-Lex and OP Portal, which send queries to Cellar via the RSS interface, the REST API and the knowledge graph SPARQL query interface.

On archiving part:

  1. Archiving is taking place using the Archivist component of Cellar. Archivist sends the publications and metadata to the long term archive, EUDOR.