The solution is simple but its creation was enormously complex: Hamburg-based Kubermatic (formerly Loodse) says organizations can now use its Kubermatic Kubernetes Platform to automate the provisioning and lifecycle management of services and applications across Kubernetes clusters in multicloud environments and geographic regions — all through a single API.
The additional capability comes through Kubermatic KubeCarrier, new software licensed under Apache 2.0 that the company claims offers an abstraction layer to Kubernetes Operators to provision applications across different clusters and clouds.
As a top-10 Kubernetes Project contributor (ahead of Alibaba, Amazon and Intel and just behind Google, VMware, Microsoft and Red Hat), Kubermatic has long offered “support for numerous clusters and their deployments,” said Scheele. “Now, we are supporting the services side, and helping customers manage applications and the day-to-day to operations.”
A big part of Kubermatic’s Kubernetes Platform is to help operations automate and manage “Day 2” Kubernetes operations. “We are going the next step by allowing customers to automate the provisioning of their services throughout their entire Kubernetes deployments, typically spread out among multicloud and on-premises infrastructures.”
Under the KubeCarrier umbrella, DevOps teams can manage databases, storage and other stateless applications and apply monitoring, service meshes, and other Kubernetes tools through a single interface to manage all of the deployed clusters,” Kubermatic CEO and co-founder Sebastian Scheele told The New Stack.
“If your internal customers are spinning up hundreds of databases, for example, how do you manage all of them by providing updates, etc?” said Scheele. “We created KubeCarrier to provide a growth framework for operations to leverage numerous clusters and the applications and tools running beneath them.”
An organization’s database team might, for example, support internal customers that offer CRM and other enterprise applications for use by the organization’s users while another DevOps team might rely on a time series data service on other clusters from a different cloud service. Instead of having to manage the applications separately, KubeCarrier allows the database operations team to manage the databases running across different clusters and separate providers with a single API. Management of applications across infrastructure providers is no longer a siloed task requiring a specialized skill set. Different teams also can have access to multiple applications and databases deployed in various cloud and on-premises environments without having to worry about their lifecycle management.
“You can layer this together, depending on the requirement and what the different teams need to talk to each other,” said Scheele. “They can individually provide this as a service to each other.”
Kubermatic is targeting customers such as telcos that will typically deploy and manage 10 or more clusters.
“If you say ‘okay, for the long term, I only plan to have a single cluster,’ then I don’t think you need this, but if you say ‘we’re currently already at five or 10 clusters and we need to scale up and scale more’, then that’s definitely the way to go,” said Scheele.
Kubermatic communicated the following features that it says KubeCarrier offers:
- Automated provisioning and entire lifecycle management of services, applications and API-accessible hardware devices.
- Registration of services, applications, and devices independent of cloud, region, or datacenter.
- Application service store.
- Multitenancy and user management with access right controls, permissions and policies to define quotas.
- Automated configuration and adjustable application services.
- Native support for all Kubernetes Operators.
Red Hat and VMware are sponsors of The New Stack.