Chapter 2. Working with Raster Data

In the first chapter, we looked at setting up one or more instances of GeoServer in a way that will make them perform well in a production environment. However, configuring instances so that they run well is only a part of the performance equation. Loading well-structured and optimized data for GeoServer to serve is another part.

In this chapter, we will focus on how raster data can be prepared and stored so that GeoServer can serve it efficiently, and in doing so, we will maximize the benefit of the effort that we put in setting up GeoServer. This chapter assumes a basic knowledge of how raster datasets are served by GeoServer and how to use the web administration interface, configure stores, and publish layers.

We will address a range of topics, from looking at how we can increase the types of raster data we can serve to how we can serve vast coverage of raster data efficiently. By the end of this chapter, you will have a better understanding of the following...

Increasing the raster formats supported by GeoServer

Out of the box, GeoServer supports a good range of raster formats, and for the majority of use cases, the standard supported formats are, most likely, all you will need. The standard formats supported by GeoServer Version 2.5.2 are:

• ArcGrid: The ArcGrid coverage format (http://en.wikipedia.org/wiki/Esri_grid)

• GeoTIFF: A tagged image file format with geographic information (http://en.wikipedia.org/wiki/GeoTIFF)

• Gtopo30: The gtopo30 coverage format (http://en.wikipedia.org/wiki/GTOPO30)

• ImageMosaic: An image-mosaicking plugin

• WorldImage: A raster file accompanied by a file that contains geo-reference information

It is likely that the vast majority of raster data that you will use with GeoServer will be in GeoTIFF and/or WorldImage (for example, TIFF files with an accompanying TIFF world file). However, there might come a time when you need to serve a nonstandard raster format through GeoServer, for example, MrSID. Luckily for us, these fringe...

How to optimize raster data for better performance

As talented as the GeoServer developers are, there is only so much they can do from a software perspective to help us run high-performing web mapping servers. As users of the software, we also have a part to play to get the best performance out of it. In the first instance, this is achieved through optimizing our deployment strategy and ensuring that our production environment is configured optimally, as we saw in Chapter 1, Installing GeoServer for Production. Secondly, and arguably more importantly, it's the way we prepare the data that we will deliver through GeoServer.

Unfortunately, it is the second element that most people overlook when setting up their web mapping servers—irrespective of the software they use. Well-structured and organized data can make a huge difference on the response times of a mapping server. While GeoServer is very flexible in the number of formats it can handle, and the many flavors of each format it can manipulate...

How to serve very large raster datasets

So far in this chapter, we looked at how we can increase the number of raster formats that GeoServer can serve, and considered how we can prepare our raster data for efficient serving. However, in these discussions, we only looked at relatively small numbers of raster images. We have not yet given any consideration to how we might approach the not-so-insignificant issue of serving large volumes of high resolution imagery. For example, how might we approach the challenge of serving raster data at a national scale such as large-scale aerial surveys or a national map product such as Ordnance Survey's OS StreetView?

We can stitch all the individual raster images together to create a single file containing the complete coverage. However, this will result in a considerably large file for processing (and, in fact, one that it might not be possible to store). The OS StreetView product, for example, requires approximately 10,592 files to cover Great Britain...

Summary

In this chapter, we looked at how we can increase the formats supported by GeoServer by implementing the GDAL plugin. To do this, we have to install the GDAL binaries and ensure that our environment is appropriately configured so that GeoServer is aware of it.

We looked at how important it is to consider the raster data that we want to serve through GeoServer, and learned how we can use gdalinfo to inspect our raster files to learn more about their structure. Having understood the structure of our raster images, we looked at strategies to manipulate them so that they are optimized; this included considering whether they should be merged to create larger single files. Other strategies to optimize our raster data included adding overviews and ensuring that the image data is organized internally as tiles. We saw how gdaladdo can be used to manage overviews and learned about the format creation options to control tiled-image data structures. We looked at how the ImageMosaic format can...