Jeffry Molanus

Container Attached Storage was introduced at FOSDEM last year, we will revisit the basic concept shortly and give an update how we have abstracted storage fully in user space, and are capable of breaking the million IO barrier with the latest data engine.

Last years talk was more a generic introduction about the concepts and ideas around openEBS. This time around we will go over more detail as to how we do it. We will go over the concept of front-ends and back-ends and how they allow for asymmetrical backends.

We will also be talking about, how we implement proxying and replication in openEBS and the ability to do rebuilds. Also, how can we make sure we are making good use of the latest hardware, e.g., NVMe devices and NVMe-OF (RDMA as well as TCP) and make sure we do not incur a penalty of being in user space?

The OpenEBS project has taken a different approach to storage when it comes to containers. Instead of using existing storage systems and making them work with containers; what if you were to redesign something from scratch using the same paradigms used in the container world? This resulted in the effort of containerizing the storage controller. Also, as applications that consume storage are changing over, do we need a scale-out distributed storage systems?

The OpenEBS project has taken a different approach to storage when it comes to containers. Instead of using existing storage systems and making them work with containers; what if you were to redesign something from scratch using the same paradigms used in the container world? This resulted in the effort of containerizing the storage controller. Also, as applications that consume storage are changing over, do we need a scale-out distributed storage systems?

Also, as we in effect have a storage instance per containerized applications, what other benefits could we get? Is there something we can do now that we have a one to one mapping between application and controller? We like to think so and will go over a couple.

Finally, how can we make enterprise-class like storage features (i.e., snapshots, clones, compression, replication, etc.) available in a typical container in such a way, that we do not depend on OS and or cloud provider specifics?