Google Cloud Compute

Compute Engine

Virtual machines on Google's infrastructure, from small shared-core VMs to large GPU machines.

What is Compute Engine?

Compute Engine runs virtual machines on Google's infrastructure. You pick a machine family (general purpose E2/N2/N4, compute or memory optimised, or GPU), an operating system and disks, and can resize or replace VMs at any time.

Key features

  • Predefined and custom machine types
  • Sustained-use and committed-use discounts
  • Spot VMs for interruptible work
  • Live migration during maintenance
  • Managed instance groups with autoscaling

What it is used for

  • Hosting web applications
  • Running business software on Linux or Windows
  • GPU and high-memory workloads

When to use it

  • Full control over the server
  • Lifting existing servers to the cloud
  • Long-running workloads

When something else fits better

  • Containerised web apps (Cloud Run is simpler)
  • Small event-driven code (use functions)

How pricing works

Billed per second for vCPUs and memory, plus disks, IP addresses and network egress. Discounts apply automatically for sustained use or through commitments. Prices change and differ by region; we confirm current costs for your setup before you commit.

Equivalent on AWS

The closest AWS service is Amazon EC2: resizable virtual servers (instances) in the cloud, with a wide choice of CPU, memory, storage and networking.

How DomainHostly helps with Compute Engine

  • Plan: choose the right setup, size and region for your workload and budget.
  • Set up securely: configure Compute Engine with least-privilege access, encryption and backups where they apply.
  • Migrate: move from shared hosting, your own servers or the other cloud with minimal downtime.
  • Run and optimise: monitoring, updates and regular cost reviews so you only pay for what you use.

Compute: Virtual machines and servers to run your applications, from a single small server to large fleets.

What to consider: Compute

  • Right-size: start with the smallest machine that meets your load and resize after watching real usage.
  • Use a load balancer and at least two servers in different zones for anything customers depend on.
  • Commit (Savings Plans / committed-use discounts) only for capacity you run all month; use spot or preemptible capacity for interruptible jobs.

Related services: Compute