Products Digital products Opendap

Distribution of Mercator products using Opendap

Opendap (Open source Project for a Network Data Access Protocol, formerly known as DODS) is a protocol enabling users to access data via HTTP (Internet). It is based on a client-server architecture, in which the client (i.e.: the user's computer, connected to the Internet) sends requests to the Opendap server (installed at Mercator Ocean) to enable it to download data. From a computing perspective, Opendap is both:

• a data request and Internet transfer protocol which uses HTTP to submit requests and send answers,
• a set of server applications (Java servlets or CGI applications) enabling the user to extract geo-referenced data.

on the client side, analysis and display software is required for accessing Opendap data. The easiest solution is to use a web browser, which gives direct access to the data on the server.

Opendap clients : MATLAB, Ferret, GrADS, NCO tools, CDAT and Excel.

Opendap enables the user to extract the data he is interested in, without having to download complete files containing all the information. In particular, he can extract data concerning: a geographical area, a time period, an ocean variable, a depth, the output format. The version of Opendap which is installed at Mercator is the DODS Catalog/Aggregation Server (AS). Its main feature is the ability to create virtual datasets by aggregating time series data from several different files. Opendap can be used directly via the user's own client (Ferret, Matlab, IDL, etc), by consulting the Opendap catalogue to extract the required information.

Home page of Mercator's Opendap server

The principle is the same (downloading data from a server to a client) but Opendap has several advantages over FTP:

• Access to metadata (a description of the data) separately from the data itself,
• Access to data sub-assemblies,
• Automatic format conversion,
• Data aggregation. For example, if the required data was distributed between several hundred different files, aggregation enables it to be gathered into a single virtual dataset, which can be displayed by the client.

See description of standard grids products