Since moving a significant amount of IT infrastructure from IDC to the cloud, FinOps has become a buzzword. The concept is simple: adding financial management to IT operations—Fin+Ops=FinOps.
Use cost visualization tools like ice to generate a bill, showing where the money is spent, with proactive notifications.
With the bill in hand, the next steps become tedious and detailed, often including:
These methods aim to cut costs but often sacrifice usability, performance, or redundancy. For example, would you hand-wash clothes to save electricity? Sacrificing availability, performance, user experience, or redundancy for cost savings can hardly be called optimization.
This involves embedding cost-saving measures into daily operations, like putting “turn off the lights” into the employee handbook, which can feel trivial and ineffective.
These three steps often seem impressive but lack concrete execution details and quantifiable metrics. They boil down to continually monitoring your bill and shutting down unused resources. It feels like following frugal advice from parents, which, while sensible, is hard to measure and apply effectively in the cloud.
The real answer has been around for decades. The history of computer science is filled with research on preventing resource leaks, providing reliable technical solutions. For example, operating systems automatically reclaim resources when a process exits, and garbage collection technologies handle memory cleanup.
Cloud computing, through APIs, allows users to share resources on a massive scale, similar to working on a giant supercomputer. In traditional programming, do you manually manage all resources? No, the operating system handles it. So why should it be different in the cloud?
Most programs were moved from IDC to the cloud, leading to the belief that Ops engineers should manage cloud resources. This inertia has driven cloud computing for over a decade, which is why even advanced IaC tools like Terraform still rely on manual git operations. If you see the cloud as a cluster of virtual servers, this might seem fine. But if you view the cloud as a giant supercomputer, you’ll realize that even Terraform is a static resource allocation tool. It’s like managing all your malloc and free operations through git-controlled YAML files.
While Kubernetes allows horizontal scaling of computing power, many core components in the cloud are beyond just computing. Dynamically allocating resources based on workload and real-time prices aligns more closely with how an operating system should function.
Like all impractical slogans, FinOps will eventually lose its appeal. ClapDB, however, genuinely cares about your bill. It’s a cloud-first data warehouse that eliminates the need for FinOps by integrating finance and architecture from the ground up. This design optimizes cloud resources, allocating them on-demand and discarding them when done.
Welcome to the FinArch era and the new age of cloud computing.