MapR Ties Kubernetes to Persistent Storage
MapR Technologies at the Strata Data Conference this week announced it is incorporating Kubernetes into the MapR Data Fabric developed to provide access to persistent storage for both monolithic and microservices-based applications.
Jack Norris, senior vice president for data and applications at MapR, says there’s been a sharp rise in interest in deploying microservices based on containers on top of Kubernetes clusters. MapR previously supported containers. But by adding support for Kubernetes the company is providing a layer of software that makes it easier for applications running on top of Kubernetes to access persistent storage via the MapR Data Fabric.
The MapR Data Fabric also supplements Kubernetes by providing the means through which organizations can dynamically connect containers to data sources. It also provides the high availability mechanisms need to make sure data is always available.
Norris says while organizations such as the Cloud Native Computing Foundation (CNCF) have made it possible to connect storage to Kubernetes, the overall data management process associated with providing containerized applications with access to persistent storage remains too manual and complex. Solving this issue has become a pressing matter because IT organizations are moving now to build stateful applications based on containers that require access to persistent storage.
At the same time, Norris notes IT organizations are dealing with a much wider variety of data sources, including streaming data. The MapR Data Fabric provides to mechanism through which IT organizations can manage an increasingly complex set of data pipelines that need to feed into microservices that are frequently updated, he says.
Those pipelines now increasingly span multiple clouds. To address that issue, MapR Data Fabric also includes connectors to container services found on public clouds such as Amazon Web Services, Microsoft Azure and Google Cloud. In addition, MapR Data Fabric is compatible with NFS, S3, HDFS, ODBC interfaces and storage formats. IT organizations are also provided tools to optimize application processing cost, performance or compliance requirements.
As microservices continue to evolve, it’s becoming increasingly clear that the worlds of DevOps and DataOps are starting to converge. Norris says enterprise IT architects are looking to drive that convergence to improve overall efficiency using platforms such as MapR Data Fabric. That convergence is then changing the way data pipelines are constructed and maintained as containerized applications are developed and upgraded, he says.
Because persistent storage is problematic it’s arguable that the development of stateful applications based on containers is being slowed. Stateless applications based on containers are already widespread, so there’s already a large pool of developers familiar with containers. The core issue is providing access to persistent storage in a way that makes it possible for stateful applications based on containers to truly scale.
Naturally, addressing these issues will require something of a meeting of the minds between developers, enterprise architects, storage administrators and database administrators (DBAs). But this juncture it may not be a question of whether stateful applications will be built using containers, but rather when and how.
