Did you know that Google Cloud has its own secure and reliable infrastructure as a service (IaaS) component that’s built on the same infrastructure as the tried-and-true Google search engine and other services?
You do now. And I’m here to tell you all about how Google Compute Engine works in this latest installment of GCP 101. So far in these series, we’ve examined Google Cloud from a high level and also have taken a look at building a data pipeline in GCP.
Now we’re going to dive into GCP’s various services, all of which you can mix and match to create your perfect cloud environment, starting first with Compute Engine.
What is Google Compute Engine?
We touched upon this service briefly in our first GCP 101 blog but didn’t get too deep into the specifics of what this means.
In short, this service allows you to run your apps on virtual machines (VMs), or instances, on physical hardware in Google’s global data center. Google Compute Engine makes it easy for you to have access to virtual machines that deliver large amounts of computing power in a cost-effective, secure cloud environment which spans 23 Google Cloud regions.
Compute Engine is great for when you need more control of the underlying infrastructure. For example, you might use Compute Engine when you:
- Are migrating existing applications through lift-and-shift or lift-and-modernize approaches to kick off your infrastructure modernization journey.
- Run Windows or other 3rd party applications where you are bringing your own license (BYOL) to CGP or use a license-included VM image. These include ones available in the GCP Marketplace.
- Have highly customized business logic or you want to run your own storage system.
This service offers the choice of preset or custom machine sizes that most closely resemble your on-premise structure to best support your workloads. A partner will do this by rightsizing your environment with recommendations for the machine sizes that work best with your instance types and managed instances groups.
Google Cloud offers a wide range of Compute Engine machine types. These include:
- General-purpose N1, N2, N2D and E2 machines that offer the best price-performance ratio.
- Memory-optimized VMs that offer higher memory per core, up to 12 TB.
- Compute-optimized machines that offer the highest performance per core for compute-intensive workloads.
- Shared-core machines for N1 and E2 VMs for a cost-effective way to run small, non-resource-intensive applications.
You also can use what are known as preemptible VMs. These are low-cost, short-term instances that are ideal for running batch jobs and fault-tolerant workloads. Google Cloud indicates that these VMs cut budgets up to 80% over traditional VMs but offer the same performance and capabilities for short-term use.