IBM is leveraging the Kubernetes expertise of its Red Hat subsidiary to make it easier to deploy its blockchain platform anywhere and manage instances of the open source Apache CouchDB running on Kubernetes clusters.
The IBM Blockchain platform is based on the open source Hyperledger Fabric platform being developed under the auspices of the Linux Foundation. Its deployment on Red Hat OpenShift, which is based on a distribution of Kubernetes, not only advances portability but also makes it possible to containerize smart contracts, peers, certificate authorities and ordering services across a hybrid cloud computing environment.
At the same time, IBM is leveraging Operator software developed by CoreOS, which Red Hat acquired last year, to make it easier to deploy and manage CouchDB document databases. IBM also plans to provide support services for CouchDB databases alongside the cloud service instance of CouchDB it makes available in the form of the Cloudant platform, which IBM acquired in 2014.
Adam Kocoloski, vice president and CTO for IBM Cloud Databases, says Operator software will make it simple to provision instances of CouchDB on a Kubernetes cluster and automate over time a variety of ongoing management tasks. Despite that increased level of automation, however, Kocoloski doesn’t expect the role of the database administrator to be automated out of existence. Rather, as the number of stateful applications running on Kubernetes clusters increases, the number of database instances that need to be managed will increase. Automated databases these days are still more a statement of vision than reality, he notes.
IBM is already employing Kubernetes across its cloud services, and the Operator software it developed for CouchDB has already been vetted in those environments, he adds.
Kocoloski notes that most of the stateful applications showing up on Kubernetes clusters are not traditional systems of record applications. Rather, emerging systems of engagement applications account for the bulk of stateful applications being deployed on Kubernetes, says Kocoloski.
There is a definite trend of Kubernetes clusters running stateful applications, though it’s still unclear the degree to which this will occur. Containers frequently have been employed to build stateless applications, but many of the use cases for those applications are expected to shift to serverless computing frameworks. Kubernetes clusters, meanwhile, are still being vetted to determine what level of scale can be achieved when running stateful applications. However, it’s worth noting that blockchain databases IBM plans to deploy on Red Hat OpenShift might be considered the ultimate instance of an immutable system of record running on Kubernetes. In addition, IBM continues to roll out Cloud Paks that essentially replace legacy WebSphere middleware with Kubernetes-based microservices.
It may be early days still as far as Kubernetes’s adoption in the enterprise is concerned. Nevertheless, the debate seems to be shifting away from whether to adopt Kubernetes to a more nuanced argument over what types of containerized applications should be deployed on the platform and when.