Kubernetes 101 - Deploying Kubernetes in Multi Cloud Deployments

1. Kubernetes provides the control pane to manage cluster deployments and configurations via yaml/json.
2. However in Multi Cloud Deployments the configurations for the cluster deployment is not Cloud agnostic and moving the containers between the Cloud Providers AWS, Azure, Oracle Cloud poses a major challenge as the managed cloud services provided by each Cloud Vendor is tightly coupled with the flavor of Cloud.
3. Also in the Multi Cloud deployment models managing the DevSecOps, Devops, Change Control, Centralized Security, Centralized Monitoring, Centralized Auditing, Alerts, Repository Management, CI/CD, Centralized code scan, Centralized Testing, etc gets complex as the Automation needs to be built specific to each cloud vendor
4. Solution is to build federation between the various Containerization models in Multi Cloud, which could lead to a customized control pane for cluster deployments, and the rest of Automation services. It is also worthwhile to invest in a cloud agnostic cluster management solution, which are a few available in the marketplace.
5. Bottom line is Kubernetes cluster farms are the next version of SOA, which enables to convert the Monolithic applications to Microservices based architecture. Though containers are not fully serverless and still needs a underlying host to run clusters (unlike aws lamda or azure functions to run app workloads) it helps for easy lift and shift of the workloads from on-prem to Cloud / Multi Cloud Deployments. In that, when we go for Multi Cloud Deployments, we need to make sure we do not end up building a Cloud vendor dependent Kube Cluster, which would end up being a Monolith at the end.
6. To preserve the true nature of the Microservices and Multi Cloud Native Kube Cluster, extreme care needs to be taken to Strategically define a Cloud agnostic Containerization Control Pane!!!

Cheers,

Vishwa