GCP Compute Engine Machine Types
What You'll Learn
You'll discover how Google Cloud Platform organizes virtual machine configurations into predefined, custom, and shared-core types to match different workload needs and budgets.
Understanding GCP's Machine Type Categories
Google Cloud offers three main approaches to sizing your compute resources:
Predefined Machine Types
These are pre-configured VM sizes organized into families similar to what you've seen in AWS and Azure. GCP provides general-purpose (N-series), compute-optimized (C-series), and memory-optimized (M-series) machines with fixed vCPU-to-memory ratios. Think of these like ordering from a restaurant menu—convenient, tested combinations ready to go.
Custom Machine Types
Here's where GCP stands out: you can create your own machine type by independently selecting the number of vCPUs and amount of memory. Want 6 vCPUs with 30 GB of RAM? You can build exactly that instead of settling for the nearest predefined size. It's like building a custom sandwich rather than choosing from the menu.
Shared-Core Machine Types
GCP offers f1-micro and g1-small instances that share physical CPU time with other VMs. These are incredibly cost-effective for low-traffic applications, test environments, or small workloads that don't need dedicated compute power. Think of this like a co-working space versus renting your own office—you pay less but share resources.
Choosing Your Machine Type
Start with predefined types for simplicity. Use custom types when your workload has specific requirements that don't match standard configurations. Reserve shared-core instances for development, testing, or truly minimal production workloads.
Key Takeaway: GCP offers predefined machine types for convenience, custom types for precise resource allocation, and shared-core types for cost-sensitive, low-demand workloads—giving you flexibility to optimize both performance and cost.