Spatial Data Processing with Docker

by Sophia Parafina

One of the major hurdles of using FOSS4G is that compiled software may not be available for your operating system or that the software has dependencies that the OS may lack. In the past, the solution has been to:

1. go through the process of building and installing the software locally using a package manager
2. using an OS specific distribution such as Portable GIS or OSGeo4W
3. running the software in a virtual machine environment such as VirtualBox

Furthermore, if OS native packages are available, they may use different versions of the same library which can cause conflicts.

Containers are a complete runtime environment containing all the dependencies that an application needs. Unlike virtual machines, containers are purpose built for an application and do not include all the parts of an operating system that a virtual machine may include. Because containers are compact, they are more akin to processes than virtual machines and are amenable to the UNIX pipeline model of processing where data streams can be fed to different processes. This makes a processing workflow scriptable and repeatable regardless of the computing environment

This presentation will present an overview of containers, containers with FOSS4G and processing workflows for raster data.