This section describes the installation, configuration, and operation of the Hyrax Data server, a data server that integrates structured data with the world wide web. Hyrax is one example of a number of data servers that implement OPeNDAP’s Data Access Protocol (DAP).
For information on how to get Hyrax downloaded and running, please see the Hyrax Downloading and Installation guide that appears later in this manual.
Hyrax uses the Java servlet mechanism to hand off requests from a general web daemon to DAP format-specific software. This provides higher performance for small requests. The servlet front end, which we call the OPeNDAP Lightweight Front end Server (OLFS) looks at each request and formulates a query to a second server (which may or may not on the same machine as the OLFS) called the Back End Server (BES).
The BES is the high-performance server software from HAO. It reads data from the data stores and returns DAP-compliant responses to the OLFS. In turn, the OLFS may pass these response back to the requestor with little or no modification, or it may use them to build more complex responses. The nature of the Inter Process Communication (IPC) between the OLFS and BES is such that they should both be on the same machine or able to communicate over a very high-bandwidth channel.
The OLFS and the BES will run and serve test data immediately after a default installation. Additional configuration is required for them to serve site specific data.
-
THREDDS Catalog Support: Hyrax supports the THREDDS catalogs. It can serve user supplied static catalogs and it will dynamically generate THREDDS catalogs of it’s internal holdings.
-
Dataset Aggregation: Collections of related data resources can be collected into a single dataset using the aggregation features. Typically these are formed for geographic tiles, time series, etc.
-
Adding/modifying dataset content.: Datasets can be modified by the server without having to actually change the underlying files. These views are independently accessible from the original data. Both dataset metadata and data values may be added or changed.
-
Supports multiple source data formats: Server can ingest source data stored as HDF4, HDF4-EOS, HDF5, HDF5-EOS, NetCDF-3, NetCDF-4, CEDAR, FITS, Comma Separated Values, and raw ASCII and Binary formats. Because of Hyrax’s extensible design, it’s easy to add new source data formats.
-
Supports data retrieval in multiple return formats: Hyrax is able to return data in DAP, DAP4, NetCDF-3, NetCDF-4, JSON, CSV, and ASCII formats, Or, you can add your own response types.
-
Gateway: Hyrax supports a gateway feature that allows it to provide DAP (and other Hyrax) services for remotely held datasets that are stored in any of Hyrax’s source data formats.
-
RDF: Hyrax provides RDF descriptions of it’s data holdings. These can enable semantic web tools to operate upon the metadata content held in the server.
-
Server Side Functions: Hyrax supports a number of Server side functions out of the box including (but not limited to):
-
geogrid: Subset applicable DAP Grids using latitude and longitude values.
-
grid: Subset any DAP Grid object using the values of it’s map vectors.
-
linear_scale: Apply a linear equation to the data returned, including automatic use of CF attributes.
-
version: The version function provides a list of the server-side processing functions available.
-
New ones are easy to add.
-
-
Extensible WebStart functionality for data clients: Hyrax provides WebStart functionality for a number of Java based DAP clients. It’s simple to add new clients to the list that Hyrax supports.
-
Extensible/Configurable web interface: The web interface for both Hyrax and the administrator’s interface can be customized using CSS and XSL. You can add your organizations logo and specialize the colors and fonts in the presentation of data sets.
-
Administrator’s interface: Control and dynamically update Hyrax from a convenient web interface. See the Admin interface documentation.
-
WMS services: Hyrax now supports WMS services via integration with ncWMS.
-
JSON responses: Both metadata and data are now available in a JSON encoding.
-
w10n: Hyrax comes with a complete w10n service stack. W10n navigation is supported through the default catalog where all datasets and "structure" variables appear as graph nodes. Data can be acquired for atomic types or arrays of atomic types in a number of formats.
-
Feature Request: If there is a feature you would like to see but don’t, let us know: [email protected] or [email protected]. (You need to subscribe first.)
Hyrax has a number of modules that provide the actual functionality of the server: Reading data files, building different kinds of responses and performing different kinds of server processing operations. Most of these modules work with the BES but some are part of the front (web facing) part of the server.
-
Gateway handler (Interoperability between Hyrax and other web services)
-
SQL handler (Unsupported)
-
WMS - Web Mapping Service via integration with ncWMS.
If you would like to build Hyrax from source code, you can get signed source distributions from the download page referenced above. In addition, you can get the source code for the server from GitHub, either using the Hyrax project or by following the directions on our developer’s wiki.
-
How to use Eclipse with Hyrax Source Code. Note that this is a a work in progress, but it will help with some of the odd steps that Eclipse seems to require.
We maintain a wiki with a section deveoted to Developer Information specific to our software and development process. You can find information there about developing your own modules for Hyrax.
We hope you find this software useful, and we welcome your questions and comments.
Technical Support:
-
[email protected] (You need to subscribe first.)