As adoption of application containers continues to expand, more companies run up against the challenges involved in deploying the technology on a large scale. A startup called Kasten Inc. exited stealth mode today to tackle one of the biggest obstacles: data management.

The issue affects so-called stateful applications such as databases that must store information for an extended period of time. Containers provide an efficient means to run workloads and move them across different types of infrastructure, which is why the technology is so popular, but this flexibility doesn’t inherently extend to the data associated with a service. As a result, companies struggle to realize the operational benefits for many of their most important applications.

Kasten has built a data management platform called K10 that addresses the root of the challenge: the dynamic nature of container clusters. The individual instances in a deployment tend to come and go as the application running on top change. When developers need to patch a service, for example, they deploy new containers running the update version in place of the old ones, which are simply deleted along with their data.

The frequency with which such changes occur in a large-scale cluster can make it difficult for administrators to keep up. By way of a solution, Kasten said, K10 provides the ability to define rules and workflows for managing data that are executed automatically when the specified conditions are met. The platform provides control not only over containerized workloads’ persistent information but also configuration files.

According to Kasten, K10 thus enables companies to ensure that data policies are followed without hindering the pace of developers’ work. For added measure, the platform provides a monitoring capability to let administrators track that everything is indeed done in a compliant manner.  

Kasten has also mixed in data protection features to help recover stateful workloads after an outage along with several other features. These include “foundational mechanisms” aimed easing the task of migrating an application across different environments, which can be difficult even with containers when they have data inside.

Kasten offers K10 alongside an open-source tool known as Kanister that it describes as an “extensible framework for application-level data management.” The startup claims that the products have been adopted by several enterprises, including an unnamed leading global carrier.

Kasten’s value proposition is also endorsed by several prominent angel investors. The backer roster includes serial entrepreneur Amarjit Gill and Dan Dobberpuhl, an electrical engineer known for his early contributions to semiconductor design.

Image: Pixabay