Tiered Metadata Service Sample Clauses

Tiered Metadata Service. Broadly this is similar to a distributed service, but also somewhat follows the concepts of PaNData. Where communities have well defined metadata which is sufficiently homogenous, their catalogues, or some subset of the metadata contained therein could be aggregated within a ‘thematic’ metadata service, with a single (or distributed) central catalogue keeping a smaller subset of the metadata. With this model, ‘thematic’ groups could use something like ICAT (or possibly iRODS iCAT) to federate their metadata. Other communities within the thematic area could then make use of the thematic service, while non-thematic users would make use of the central information service. This again addresses some of the identified scalability issues and may help address some of the immediate needs of the EUDAT communities. In this case, the search and volume loads are distributed amongst the thematic service. Where heavy load is expected or high resiliency is required, it would be possible to distribute the metadata across different hosts/sites using database replication technologies to ensure consistency within the thematic area as is suggested in the Distributed Service mentioned previously. Similarly, the central service could be distributed for load balancing and resilience. Depending on the size of the communities, it may be possible to hold almost full metadata for each site within the thematic service, and only a smaller searchable subset in the central catalogue which would clearly be advantageous, and if a theme becomes too large (in terms of metadata entries) it could be further subdivided into sub-themes. While this approach scales quite well, it does not address issues where existing catalogues do not provide an interface for searching for changes or the needs of legacy data holders. In these cases, it may be possible to delegate queries directly to the service provider. The main problem with this approach will be finding communities within a thematic area which provide such homogeneous metadata. It has already been identified within the climate change community that the information provided by different sources is quite heterogeneous which makes interoperability a significant obstacle to cross disciplinary research.