[mercanator]

In our organization we have seen tremendous value using docker for our CI server (bamboo) and for infrastructure deployment. On the CI server side of things.... Our CI server a.k.a the "host machine" in docker speak has a MySQL database that we did not want used by all of our builds and wanted to isolate the creation of our application from the ground up within using its own MySQL instance. Additionally there is a lot of other peripheral software that we did not want to install on our host machine but it was critical for our unit tests to run. This is where docker is valuable to us from a build perspective. It allows us to isolate the application for unit testing and not get caught up with the possibility of other running software on the same machine affecting the builds, therefore freeing us up from debugging, and therefore saving the company money because developers are not wasting time debugging CI server issues. It's easy to isolate CI server related issues from the docker container running the unit tests because a developer can just run the same tests using the container on their local machine, so it creates a consistent environment.

On the infrastructure deployment side of things.... Previous to our "dockerized" infrastructure we were managing about 7 different AMI's for all our servers and it was becoming a pain in the butt to manage the installation of new software if our application called for it, create a new AMI, then re-deploy said AMI. If you have experience with AWS and you have done this enough times, I'm sure you have faced at one point or another long wait times for your AMI to be created before you can re-deploy with that newly created AMI. This is time wasted on the application deployment side of things, but also on the personnel side of things while you wait for that damn thing to be created so you can re-deploy. Time is money and money waiting for resources to be available or for AMI's to be created is money taken away from the business. Additionally though in its infancy stage, we are using docker-compose (https://docs.docker.com/compose/) which offers some really nice ways of defining your container infrastructure within a single machine, I highly recommend looking into this for further efficiency.