Data Storage for Cloud-Native Applications

1. Summary

Containers have become very popular for new enterprise applications, and Kubernetes is the de facto standard for managing them in production. The reasons for the success of containers and Kubernetes are many but, most importantly, they promise real application portability across on-premises and public cloud infrastructures. Unfortunately, these promises fall short without the right data storage infrastructure to manage data correctly. In fact, though it’s extremely easy to recreate the same application environment across multiple Kubernetes clusters, data storage still presents a number of challenges, including data persistency and mobility.

In the last few years, several initiatives to support stateful applications and persistent data volumes have appeared. Most of them failed, and though the latest attempt, Container Storage Interface (CSI), looks promising, it’s still immature. Moreover, it’s designed to make traditional storage compatible with Kubernetes, which isn’t enough, especially now with all the different types of data we create and consume. And it poses additional risks, including:

• Performance and scalability: Most traditional storage systems aren’t designed to scale in the control plane and manage the large number of backend operations required by many modern applications.
• Portability: The storage array may simply not be available as a service or a virtual machine in the cloud.

MinIO offers a thoughtful alternative to traditional storage for Kubernetes environments, and it’s aligned with the needs of most applications developers:

• It’s compatible with Amazon S3, a familiar interface for developers.
• It’s flexible and can be deployed within the application or as a standalone storage system.
• It overcomes all the limitations imposed by traditional data volume provisioning mechanisms.
• It has the demonstrated scalability and performance features to serve a wide range of workloads.

In short, MinIO provides the right combination of simplicity, flexibility, scalability, and performance to build a scalable and persistent data storage layer, deployed in containers, that works across clouds and the edge to serve a vast range of applications.