Free Webinar: State of GeoServer, 2.22 Release
GeoServer keeps advancing, with great support from the community, and the trust that organizations are putting in one of the most popular geospatial data web publishing tools. We, at GeoSolutions, are very fortunate to have in our team key code contributors, as well as customers supporting the growth of GeoServer, for the benefit of the whole geospatial community. Version 2.22 was released some months ago (see official blog) and the new release, 2.23, is coming in the next months. You can register by clicking on the button below.
This blog highlights the new functionalities and provides the information for our next webinar on February 28th where you can learn more about these releases and interact with our lead GeoServer developer, Andre Aime. Here are the highlights of the 2.22.0 release, to whet your appetite. More information about the webinar at the end of the official GeoServer blog post.
The styles list provides a Format column indicating the format used.
Thanks to Mohammad Mohiuddin Ahmed for this change.
Significant improvements in raster rendering performance
Raster rendering performance has increased significantly for two specific use cases:
- GeoTIFF hyperspectral images, with hundreds of bands, and band interleaved structure
- Mosaicking hundreds of small images
Hyperspectral sensors collect information at a very high spectral resolution, producing images with hundreds of bands. The typical pixel interleaved layout, where all the bands of a single pixel are stored together, is particularly inefficient while rendering a false color image, where only three of them are used. A band interleaved, where each band is stored in a separate bank, is more efficient. GeoServer previously loaded band interleaved images in an inefficient way, but that has been handled, improving both memory usage and rendering performance, in proportion to the number of bands found in the GeoTIFF. For the typical hyperspectral image, that implies an improvement of a couple of orders of magnitude.
The second use case involves mosaicking hundreds of images, under the notion that each one has a significant number of overviews. Showing the entire mosaick involves opening all these files, fetching the smallest overview, and mosaicking the result: the process used to be slow and very memory intensive (going with the square of the output image size). The implementation has been improved so that the memory used in now linear with the output image size, and the amount of processing has been reduced as well, providing again a couple of orders or magnitude speed up when mosaicking several hundreds small images.
Thanks to Andrea Aime (GeoSolutions) for these improvements.
Welcome Page Updates
The welcome page description provides a summary of the workspaces and layers available to the current user.
The header includes a welcome message and a link to the organization providing the service.
Each web service is listed using the service title as a heading, followed by the service abstract as a description. The protocols provided by the service are displayed as blocks linking to the web service URL. These are the URLs used to access the service in a desktop or web application.
The services shown are based on the permissions of the current user. As an example when logged in as an Administrator the REST API service is shown with a link to the API endpoint.
For more information on the welcome page and an example of how to use service URLs in QGIS visit the user manual Welcome reference page.
Workspace and Layer Selection
Use the top-right corner of the welcome page to:
- Select workspace to browse workspace web services
- Select layer and layergroup for layer specific web services
You can book mark or share this page (which is great for providing a team or project with its own distinct web services and landing page).
Using a workspace virtual web service is great if you are setting up a GIS project, supporting a web application, or providing GIS services for a team. This is especially true as it is straight forward to manage security on a workspace basis.
Using a layer virtual web service is great when registering a layer with a catalogue service such as GeoNetwork. It provides a web service that can only be used to access a single layer.
For the technical background on this feature see Virtual web services in the user manual. This functionality has been present in GeoServer for a long time; but because it required hand editing URLs many users were not aware of capability.
Contact Information and Service Metadata
Contact information now includes a welcome message to be used as introduction on the welcome page for the global services. Editing the contact details for a workspace will override this introduction for visitors viewing the workspace services.
Disabling global services prevents any services from being accessable or listed on the initial welcome page.
All these fields, including the email address, make use of GeoServer internationalization allowing the welcome page to be customized for all your visitors.
For background information visit GSIP-202. Thanks to Jody Garnett and GeoCat for these improvements.
Community modules news
News about community modules improvements, and new community modules you’ll find in the 2.22.x series.
A reminder that GeoServer community modules are still being worked on and are not directly available for download. If you are interested in these topics please support their completion directly by compiling the source code and contributing; or financially by sponsoring or contracting the development team working on the activity.
COG reader support for Azure
The COG reader community module now supports COGs stored in Azure as well. The location of the COG can be provided as a HTTP(s) link, while eventual access credentials should be provided as system properties.
To support this activity contact Daniele (GeoSolutions).
STAC datastore and mosaicking
A new community module, STAC datastore, supports connecting to a STAC catalog implementing the STAC API, and serve collections as vector layers, and items as features in said layers, with full filtering and time dimension support, if the server implements a CQL2 search.
The store can also be used as an index for an image mosaic, if the STAC API assets points to accessible Cloud Optimized GeoTIFFs.
To support this activity contact Andrea (GeoSolutions).
Vector mosaicking datastore
The vector mosaic datastore allows indexing many smaller vector stores (e.g., shapefiles, FlatGeoBuf) and serving them as a single, seamless data source.
An index table is used to organize them, know their location on the file system (or blob storage) and their footprint, along with eventual variables that can be used for quick filtering (e.g., time, collecting organization, and so on).
This can have some advantages compared to the typical database storage:
- Cheaper, when dealing with very large amounts of data in the cloud, as blob storage costs a fraction of an equivalent database.
- Faster for specific use cases, e.g, when extracting a single file and rendering it fully is the typical use case (e.g. tractor tracks in a precision farming application). This happens because the file splitting de-facto imposes and efficient data partitioning, and shapefile access excels at returning the whole set of features (as opposed to a subset).
We invite you to join our free webinar on Tuesday, Feb 28th at 11:00 ET / 16:00 GMT / 17:00 CET (more time zones), to learn more about the 2.22 release and what is being prepared for 2.23 from our GeoServer Technical Lead, Andrea Aime. We are planning to do a 1 hour presentation and 30 minutes of AAA (Ask Anything to Andrea). You can register by clicking on the button below.
Hope to see you virtually on Feb 28th, meanwhile, stay safe and keep strong!
The GeoSolutions team