What is Google’s Cloud Platform?

Described without… help from marketing

Sadly, the Wikipedia entry for GCP is garbage, and while the official docs are pretty good, the marketing-dust sprinkled on them gives me a toothache.

As part of my re-self-introduction to Google’s Cloud Platform, I wrote an objective summary of GCP for my own reference. I have a conflict of interest, so I can’t fix the Wikipedia entry, but I can share what I wrote, so here it is.

Google’s Cloud Platform

Google Cloud Platform (GCP) is a collection of Google’s computing resources, made available via services to the general public as a public cloud offering.

The GCP resources consist of physical hardware infrastructure — computers, hard disk drives, solid state drives, and networking — contained within Google’s globally distributed data centers, where any of the components are custom designed using patterns similar to those available in the Open Compute Project.

This hardware is made available to customers in the form of virtualized resources, such as virtual machines (VMs), as an alternative to customers building and maintaining their own physical infrastructure.

As a public cloud offering, software and hardware products are provided as integrated services that provide access to the underlying resources. GCP offers over 50 services including Infrastructure as a Service (IaaS), Platform as a Service (PaaS), and Software as a Service (SaaS) offerings in the categories of Compute, Storage & Databases, Networking, Big Data, Machine Learning, Identity & Security, and Management & Developer tools.

These services can be used independently or in combination for developers and IT professionals to construct their own, custom cloud-based infrastructure.

GCP is hosted on the same underlying infrastructure that Google uses internally for end-user products including Google Search and YouTube.

Regional Service Architecture

Each of Google Cloud Platform’s services and resources can be zonal, regional, or managed by Google across multiple regions.

A zone is a deployment area for resources within a region. Zones are isolated from each other to prevent outages from spreading between them, so each zones is considered a single failure domain within a region.

Zonal resources operate within a single zone; if a zone becomes unavailable all of its resources are unavailable until service is restored. Regional resources are deployed with redundancy across zones within a region. Multi-regional services (Google App Engine, Google Datastore, Google Cloud Storage, Google BigQuery) are managed by Google to be redundant and distributed within and across regions.

All GCP service instances are configured such that maintenance events are transparent to applications and workloads via live migration. Live migration moves running virtual machine instances out of the way of maintenance that is being performed.

GCP services are available in 18 zones in 6 regions: Oregon, Iowa, South Carolina, Belgium, Taiwan, and Tokyo¹.

In 2016, Google announced plans to make 22 zones and 8 new regions available in 2017: Sydney, Sao Paulo, Frankfurt, Mumbai, Singapore, London, Finland, and Northern Virginia.

Within each region, traffic tends to have round-trip network latencies of under 5ms on the 95th percentile.

Note that like everything I publish on Medium, these are my views on GCP, and may not represent the opinions of my employer.