Abstract
Containerization is a part of a broader trend towards a fundamentally better approach to building and running software - sometimes called DevOps or Microservices or Cloud Native - these approaches are delivering massively better performance as measured by the agility of the organizations adopting them. HOWEVER, today containers do a bad job persisting data and as a result either stateful workloads are run on certain clouds or are run in a non DevOps friendly way. While all storage vendors are now claiming “support for containers” few if any have undertaken a refactoring of their storage so it may run on containers and can fit directly into fully automated Kubernetes or other environments. Why refactor storage services as micro services running on containers and what are the challenges one confronts?