3.3 – A decentralised model for aid information
We suggest that the standard adopts a decentralised model for the distribution and storage of aid information – rather than one in which content and data is deposited in some central repository.
There are several advantages associated with a decentralised approach:
- The costs associated with maintaining and updating a central repository could potentially be high. Using a registry which users can edit can reduce ongoing costs.
- Multiple versions of the information can be stored in multiple locations, such as organisational websites, third party websites (individuals, other organisations, …), third party storage facilities (e.g. the Internet Archive, Amazon Web Services, Talis Connected Commons, …) and so on.
- Data publishers may wish to (continue to) primarily publish and syndicate their data on their own website, rather than in a third party repository. E.g. government departments and international organisations may have their own established systems and procedures for publishing data.
- This model allows users to build on datasets, and contribute improved versions back to the ‘ecosystem’ (such as converting data to particular formats such as RDF, aggregating datasets and so on).
It is suited to scenarios in which there are many actors, with different needs from the data.
- It is suited to the incremental, collaborative development of data 1.
Below is a diagram representing how information might move around between data publishers, users, websites and storage services.