Link copied to clipboard!

Group 1: Core Compute

Virtual machine infrastructure: Compute Engine (managed VMs), Preemptible VMs (cost-optimized), Sole-tenant nodes (dedicated hardware), Managed Instance Groups (auto-scaling), Instance Templates (configuration blueprints). Core differentiators: performance isolation, cost models, autoscaling patterns, and compliance requirements.

Selection Principle: Start with standard Compute Engine; use preemptible for fault-tolerant workloads; sole-tenant for compliance/performance isolation; MIGs for stateless scaling; templates for consistent deployments.

Services & Core Identity

Compute Engine

Standard virtual machines with flexible machine types, custom configurations, and persistent storage options.

Preemptible VMs

Short-lived, cost-effective instances (up to 80% savings) that can be terminated with 30-second notice.

Sole-tenant Nodes

Dedicated physical servers for workloads requiring hardware isolation for compliance or performance.

Managed Instance Groups

Auto-scaling groups of identical VMs with health checking, load distribution, and rolling updates.

Instance Templates

Reusable VM configurations defining machine type, image, startup scripts, and network settings.

Key Differences

Dimension	Compute Engine	Preemptible VMs	Sole-tenant	MIGs	Templates
Use Case	General compute	Batch/fault-tolerant	Compliance/isolation	Auto-scaling apps	Standardized deployment
Cost Model	Standard pricing	Up to 80% discount	Premium for dedication	Standard + scaling	Template creation free
Availability	High availability	Can be terminated	Dedicated hardware	Multi-zone support	Configuration only
Scaling	Manual/scripted	Manual/scripted	Manual	Automatic	Used by MIGs
Management	Individual instances	Individual instances	Node-level	Group-level	Template-based

Mathematical Selection Model

Criteria [0..10]. Higher scores indicate better fit. Adjust sliders based on workload requirements.

Cost Sensitivity Fault Tolerance Compliance Need Auto-scaling Need Standard Deploy Performance Isolation Multi-zone Batch Workload

Score_ComputeEngine = 0.25*C_standard + 0.20*(10 - C_costSensitive) + 0.15*C_reliability + 0.15*C_flexibility + 0.15*C_availability + 0.10*C_simplicity Score_Preemptible = 0.35*C_costSensitive + 0.25*C_faultTolerant + 0.20*C_batchWorkload + 0.10*C_shortTerm + 0.10*(10 - C_complianceNeed) Score_SoleTenant = 0.30*C_complianceNeed + 0.25*C_perfIsolation + 0.20*C_security + 0.15*C_predictability + 0.10*(10 - C_costSensitive) Score_MIGs = 0.30*C_autoScaleNeed + 0.25*C_multiZone + 0.20*C_loadVariance + 0.15*C_healthManagement + 0.10*C_rollingUpdates Score_Templates = 0.30*C_standardDeploy + 0.25*C_consistency + 0.20*C_reusability + 0.15*C_automation + 0.10*C_governance

{{s.name}}: {{s.val | number:2}}

Interpretation Rules

Compute Engine: Standard workloads requiring consistent uptime and flexible configuration
Preemptible VMs: Batch processing, CI/CD, rendering, scientific computing where interruption is acceptable
Sole-tenant: Regulated workloads (healthcare, finance) requiring hardware isolation
MIGs: Web applications, APIs, microservices needing horizontal scaling
Templates: Organizations requiring standardized, repeatable VM deployments

When NOT to Use Core Compute

Fully managed application hosting (consider App Engine, Cloud Run)
Serverless event processing (consider Cloud Functions)
Container orchestration without VM management (consider GKE Autopilot)
Simple static websites (consider Firebase Hosting)

Summary

GCP Core Compute provides flexible VM infrastructure from cost-optimized preemptible instances to dedicated hardware isolation. Choose based on cost constraints, availability requirements, scaling patterns, and compliance needs. Combine services: use templates for standardization, MIGs for scaling, and appropriate VM types for workload characteristics.

Scan with Phone