One of the biggest unmet challenges associated with building microservices based on containers is managing all the data attached to them. As microservices dynamically appear and disappear, keeping track of what is being used or needs to be made available becomes a significant IT operational headache.
To address that issue, Kasten has developed a K10 platform to dynamically apply policies to data being accessed by containerized applications deployed on top of a Kubernetes cluster. In addition, Kasten is making available Kanister, a series of extensible open source blueprints that can be employed to automate lower-level data management tasks.
Kasten CEO Niraj Tolia says the rise of containerized applications is driving the need for a more application-centric approach to tasks ranging from moving and copying data to implementing data protection policies. In contrast, traditional legacy data management tools were never intended to be support dynamic use cases typically found wherever cloud-native applications are deployed.
Kasten provides IT organizations with a platform for managing data being accessed by stateful applications that have been containerized. That platform can be invoked either by developers directly using an application programming interface (API) or by IT operations teams employing a user interface, says Tolia.
The K10 platform applies policies defined by the IT operations teams automatically regardless of how data is being accessed. Tolia says this is critical because one of the significant barriers to adoption of containerized applications is compliance. IT operations teams need to be able to demonstrate what amounts to a chain of custody for data to comply with any number of rules in highly regulated industries.
Most IT organizations were struggling with data management long before the arrival of containers and microservices. The volume of data that needs to be managed has already increased exponentially. Plus, the velocity at which data is being created and accessed has significantly increased. The rise of ephemeral microservices accessing data intermittently in unpredictable ways is about to aggravate an already-complex challenge, especially now that data is routinely being accessed in both on-premises IT environments and multiple public cloud computing services.
Kasten is betting that the rise of microservices and containers will motivate IT organizations to tackle these issues because of all the inherent dependencies between microservices trying to access the same data. Tolia says many IT organizations face the choice of building these data management capabilities themselves or buying a platform designed from the ground up to address data management issues in a containerized environment.
Historically, most IT organizations have opted to buy rather than build a data management platform, Tolia notes. Kasten’s challenge will be the degree to which incumbent providers of data management platforms can extend their reach into containerized environments.
It’s not clear to when—or even to what degree—microservices and containers will engender a data management crisis. But all the elements to spark one are now coming together in a way that is likely to catch many IT organizations by surprise.