Genomics in the AWS Cloud

CHAPTER 10Containerizing Your Workflow on the Desktop

This chapter will teach you about containerization, which is a way of stripping your applications down to their essentials and running only those key parts as an easily managed module—a container on a host computer. When you use flexible and scalable AWS services to run that host computer, you can have a very powerful genomics working environment to use at minimal cost.

Introducing Containerization

At one time, long ago, your only option for running applications was a physical server of some kind. Whether it was on your premises or hosted in a data center or some other way, there had to be a physical box with one or more physical processors, some memory, and some storage. This was an inefficient approach to the problem of providing applications. It was devilishly hard to strike a balance between making maximum use of expensive hardware (depreciating in value all the time) and having enough application capacity to handle surges in demand. One was either wasting capacity or annoying application users with shortfalls in performance and availability.

Virtual machines (VMs) provided an intermediate solution. With virtual machines, you could have images of applications stored as large files. These images contained all the information about the server that was required to run an application. They included an operating system, all the application software, and all of its dependencies. The image files could then be loaded by a ...