Understanding and Designing for Docker I/O

Designing a Docker Storage Infrastructure

Container technology and Docker specifically places unique demands on the storage infrastructure that most legacy storage architectures are ill-prepared to handle. The initial concern is developing a storage infrastructure that will support a container based dev/ops environment. The second concern is that these infrastructures will have to support containers in production as the value of something more granular than a virtual machine (VM) is understood.

The State of Docker Storage Today

Docker storage today direct attaches to specific servers and supports an agile dev/ops environment. Gradually, early Docker adopters are moving to shared storage as the need to create stateful containers and the need to share data between containers increases. The need to share data between containers or even have stateful access to data becomes more obvious as the traditional enterprise looks to expand the Docker use case beyond test/dev to production and true dev/ops. The infrastructures that support these shared Docker environments are often initially modeled after virtualized server environments. While container technology has some similarity with virtualized servers, the I/O pattern is not one of them.

Download the full paper to read Storage Switzerlands’s advice on how to design a Docker storage architecture.

Docker I/O can be difficult. Learn how best to design a storage infrastructure for Docker containers.