Q&A: MapR's Jack Norris on the Converged Data Platform for Docker
Today, MapR, one of the leaders in the Big Data space has announced a new environment that is optimized for Docker and container-based architectures, which the company bills as "critical for today’s modern architectures that require application development agility, fast time-to-value, and scale." The company claims it is "the industry’s first persistent storage for containers that offers complete state access to files, database tables, and message streams from any location." The MapR Converged Data Platform for Docker includes the MapR Persistent Client Container (PACC) that lets stateful applications and microservices access data for application agility.
MapR SVP of Data and Applications Jack Norris (seen here) caught up with us for an interview related to the news. Here are his thoughts.
What are some of the primary challenges that organizations working with container technology are experiencing?
The biggest challenge impeding the adoption of containers today, is the difficulty in supporting stateful applications. Container platforms have no out-of-box, scalable way to let applications persist state. In case of application or hardware failure, all data written by applications is lost. As a result, applications that are containerized today are simple and stateless. Companies have a huge amount of existing/legacy applications that could benefit from containers, but the lack of a complementary data platform that can eliminate the complexity of supporting these applications has limited their deployment.
You’ve just released a solution for persistent storage for containers. What are the advantages this provides?
The MapR Converged Data Platform for Docker provides a scalable, easily accessible data platform that complements the containerization of applications using Docker. Our solutions includes the MapR Persistent Application Client Container (PACC) which are pre-defined, extensible containers. You can containerize existing applications without any code changes because of the libraries we provide in the containers. You can deploy your containers on any node immediately without requiring any storage-specific provisioning on those nodes. In addition to existing applications, we enable the containerization of new, innovative applications by offering more than just file storage. Our platform supports a wide range of data types including files, database tables, streams, images, and more. The MapR platform scales to meet the needs of the largest and most complex organizations with a solution that supports the availability, integrity, and performance required for the containerization of all applications including stateful apps.
How does the underlying platform in your Converged Data Platform make using containers more streamlined?
Our platform provides the ability to access any data stored as files, databases, and streams is built into the container. The containerized application accesses and stores the directly in our converged data platform that can be accessed from anywhere the container moves. If the container is lost due to hardware or software issues, the container can be brought up in a new location and pick up where it left off. The data management tasks for administrators are also greatly streamlined. The MapR Converged Data Platform provides uniform security, data protection, and management. And we’ve proven our solution at scale at hundreds of companies including those with petabytes of data and that span on-premise and the cloud.
You’ve just received a patent for the underlying platform, as well. What is the scope of this patent and what will it mean for MapR?
MapR has been granted multiple patents for the underlying architecture of our Converged Data Platform. These include the ability to converge the processing, analytics and management for file, table and streams. These patents also are at the core of our unique ability to overcome the challenges of providing scalable, strong consistency and high performance. What this means for MapR is that we have the data platform required by customers to drive digital transformation and this announcement is further validation of our innovation. The ability to both transform customer experience and operational excellence is uniquely provided by our platform.
Storage in general is going through a renaissance right now. Costs are falling, software-defined networking is on the rise. What are some of the major storage trends you see playing out over the next 10 years?
The assumptions that drove the storage market for the past decades are quickly changing. In the past, application development required specialized stores of data and dictated that data must be specially prepared and loaded. We are now moving to a fast moving data fabric where the applications are being brought to the data directly. We’re seeing a convergence of operations and analytics, as well as data-in-motion and data-at-rest. This demands a converged data platform that can serve this environment and MapR is at the forefront of this change. Traditional storage will go through the same phases that we saw with tape. Originally tape was a primarily medium that gradually shifted to archival use and then quickly faded. We will continue to see storage costs sharply fall and the use shift to cold archival storage.
Is Docker going to be a kind of de fact standard for containers as far as the eye can see? Or will other solutions give it serious competition?
Docker is receiving a lot of attention and experiencing tremendous momentum. A rich ecosystem is growing around Docker with many management and development options springing up. So yes it is becoming the established standard. That said, we will continue to see innovation and new approaches emerge. So at this stage it’s important to focus on standards that are being established to provide openness and interoperability that allows a rich ecosystem to flourish.
Anything you would like to add?
There’s a lot more to the MapR platform and we have more interesting news that will be coming shortly.