Introduction

Oceantron is a web server dedicated to the dissemination of ocean in-situ observation data collections.

The aggregated datasets can be requested for view or download with subsetting criteria (x,y,z,t,phenomenon) through standardized services (OGC/WMS, OGC/WFS, OGC/SOS, OPeNDAP).

The server reads locally the observation data (from netCDF collections, ODV binary database, RDBMS tables).

Datasets available on this server

DatasetsSOSWFSWMSDUMMYOPENDAP
INS-HCMR-MED-ARGO_NRT-OBS_PROFILE_LATEST
INS-HCMR-MED-TS_NRT-OBS_POINTSERIES_ARCHIVE
INS-HCMR-MED-TS_NRT-OBS_POINTSERIES_LATEST
INS-HCMR-MED-TS_NRT-OBS_PROFILE_LATEST
INS-HCMR-MED-TS_NRT-OBS_TRAJECTORIES_ARCHIVE
INS-HCMR-MED-TS_NRT-OBS_TRAJECTORIES_LATEST

Software architecture

The software architecture relies on a standard data model of the ocean in-situ observations as sampling features(Point, Profile, PointSeries, Trajectory , FeatureCollection...). This model, although derived from CSML work (http://csml.badc.rl.ac.uk/), is independant from any technical protocol or format (OGC, OPeNDAP, netCDF). Together with the observations, the requests and observation collections are described in this model. The whole is called valueObject in the system. Besides, plugins (businessUnit and frontdesks) are connected to the system to :

  • storageUnits : read sampling features from persistence layer (DBMS, files, ...)
  • transformationUnits : transform the sampling features (change parameters names, interpolation, split Trajectory into Points observation, ...)
  • frontDesks : parse data request and format the result into specific protocols (OPeNDAP/Dapper, OGC/WMS, OGC/WFS, OGC/SOS, ...)

Each of these components should be considered as a specific workshop on a production line. Each minds a specific business, repectively : get observation from a specific repository, transform them according to a known procedure, pack and send the observations as requested by a user.

The core system synchronizes the transmission of the sampling features through the production lines for each user request. The manager handles multi-session requests (in case a user can not get its whole pack of observations in a single request).