Skip to content

Planning & Economics

1 module is currently being reworked. Watch this section over the next few days.

Before buying a single server, you need to answer fundamental questions about your infrastructure strategy. The decision to run Kubernetes on your own hardware requires careful consideration of the trade-offs between cloud convenience and the control, performance, and long-term cost savings of an on-premises deployment. Should we go on-prem at all? How many servers do we need? How should we organize our clusters? What will it cost over three years?

In this section, we will explore the critical planning stages required for a successful on-premises Kubernetes initiative. We will evaluate the business case for repatriation, understand how to properly size servers for specific workloads, and design a cluster topology that ensures high availability and resilience across your physical data centers.

Finally, we will delve into the economics of running your own hardware. You will learn how to calculate Total Cost of Ownership (TCO), build a comprehensive budget encompassing CapEx and OpEx, and implement FinOps practices to track and charge back usage across your internal teams. By the end of these modules, you will have a solid foundation for designing and funding an enterprise-grade on-premises Kubernetes environment.

ModuleDescriptionTime
Module 1.1: The Case for On-Premises KubernetesCloud vs on-prem decision framework, five drivers, breakeven analysis45 min
Module 1.2: Server Sizing & Hardware SelectionCPU, RAM, NVMe, NUMA, GPU considerations60 min
Module 1.3: Cluster Topology PlanningShared vs dedicated control planes, etcd sizing, rack-aware scheduling60 min
Module 1.4: TCO & Budget PlanningCapEx vs OpEx, power/cooling, staffing, cloud breakeven45 min
Module 1.5: On-Prem FinOps & ChargebackResource tracking, internal billing models, optimizing on-prem efficiency45 min