Portworx Briefing Note
Docker is changing the application development process, but most data centers want to leverage the agility and efficiency of containers from development all the way through production. The move to production requires data persistence that keeps the benefits of container portability, without the complexity of legacy systems.The challenge facing IT planners is that Docker was originally designed to use direct attached storage. IT professionals have two options. The first is to look for a solution that provides compatibility back to their legacy storage solutions. The second option takes a fresh approach, which leverages the Docker extensibility model to create Container Defined Storage. Portworx is developing a solution that integrates traditional storage features into Docker instead of forcing Docker into a legacy storage architecture.
The Docker Storage Challenge
Virtual machines (VM) recreate the entire physical server in software. The operating system redundancy between instances is a significant manageability and efficiency tax. Container technology, like Docker, is more granular, eliminating redundant parts of the operating system and sometimes even the overlap among application instances. The efficiency of container technology allows for many more containers per physical server. The additional containers can be created and eliminated almost instantly. While container technology is more efficient with server hardware the number of containers and the speed at which they are created or removed puts a significant drain on I/O.
Moreover, legacy storage systems are still playing catch-up to the VM revolution from the last decade, safely hiding in trimming overhead of VMs that we no longer need anyway. Containerized applications are arriving at the IT professionals doorstep, leaving them with a quandary. How do they scale and manage containers with state, without all this legacy complexity and cost?
The Portworx Solution
Portworx, instead of trying to fit legacy storage architectures into Docker, builds on Docker to deploy and manage storage. The Portwork’s storage software installs within a cluster of servers as a container. It then carves the storage resources within the cluster, negating the need for an external storage system. The Portworx system then manages the storage for all the Docker containers in the cluster, enabling highly-available persistent storage. Persistence means applications get to consume storage as an elastic resource — as they would in a cloud– easily writing and saving state.
Portworx also provides complete data services that enterprises expect from production storage solutions like snapshots and replication. However these features are not LUN specific, as they would be on a legacy storage system. Instead, they are container aware. Without it, the storage system re-introduces inefficiencies since every container on that LUN is impacted instead of just the container.
One of the largest hurdles to moving containers from a Dev/Ops utility to an enterprise-wide workflow is the lack of features in both protecting and delivering sufficient performance to those containers once they are in production. Features like snapshots, replication, container mobility and storage infrastructure scalability are potentially more important to a containerized environment than a virtual one. Portworx empowers enterprises to take a big step forward in creating a production ready container environment.