The DataONE infrastructure exposes application programming interfaces (APIs) that support interaction between Member Node and Coordinating Node, and to support development of applications that interact with the DataONE infrastructure (elements of the Investigator Toolkit).


Figure 1. Overview of the system components and the methods being called by each component. An arrow to a target indicates the origin is calling the method on the target.

All services in DataONE are implemented using a RESTful approach with HTTP as the transport protocol and XML for encoding messages.

A Member Node provides content to the federation, and there are two main technical aspects to Member Node participation in DataONE - the service interfaces to be implemented (i.e. the Member Node APIs), and the content that is to be served. At a fundamental level, all content in DataONE is treated as discrete, immutable objects, each of which has a unique identifier. A Member Node would be considered functionally complete if it were able to support the required services interfaces for Tier 1 participation (i.e. public access, read only content), and so enable the discovery of all objects available on the Member Node (through listObjects), low level description of each object (through getSystemMetadata), and retrieval of the object given it’s identifier (get).

There are basically three types of object being processed by DataONE: data objects, science metadata objects, and resource map documents. Each of these are uniquely identifiable by their unique identifier (PID), and each has associated System Metadata which describes the type, size, and so forth of the object.

Data objects are treated as opaque blobs. The object availability is registered by the coordinating nodes, and the blob is retrievable via the get method given its identifier.

Science metadata objects are metadata documents such as EML, FGDC, ISO19115 and so forth that provide metadata describing some data object(s). These are represented in XML according to their respective schema.

Resource Maps describe the relationships between data and metadata - they are basically RDF documents that conform to the OAI-ORE specifications.

From an object or class inheritance perspective, science metadata and resource maps might be considered specializations of the data object type in that more functionality is provided to DataONE by those types of object.

From a purely technical perspective, a Member Node may be completely functional and not provide any metadata or resource map documents. It will mean that the content provided will not be discoverable through the search interfaces.

Member Nodes can expose different levels of functionality that equate to four different tiers while participating in the DataONE network. The tiers and Member Node APIs necessary to support are summarized in the table below. Also indicated are the corresponding Coordinating Node APIs that are necessary for functionality expressed for each tier.

Tier Description MN APIs CN APIs
1 Public read, no authentication or access control of content. No content can be created through the DataONE service interfaces. The node can not act as a replication target. MNCore, MNRead CNCore, CNRead
2 Read and resolve with access control support. Tier 1 + MNAuthorization CNAuthorization, CNRegister, CNIdentity
3 Write (create, update, delete), possibly limited support for data types Tier 2 + MNStorage  
4 Replication target. The types of content supported by the node may be limited to a subset of object formats. Tier 3 + MNReplication CNReplication

The specific methods that must be implemented for each tier are listed in the respective APIs.


Related Topics