Google argues that its cloud infrastructure is already design’ed for enterprise grade scalability (the pitch being “You can run on Google“), reliability and security and that it has the best fiber network in the industry. Google argues that if engineers were making the choice of cloud infrastructure, most would choose Google Cloud Platform – GCP given capabilities such as lightning-fast VM loading times, flexible per-minute pricing, simple and fast configuration, superior container-centric services and second-to-none machine learning and analytics functions such as BigQuery (which Google argues is superior to AWS). Google argues that its Google Compute Engine infrastructure can technically do anything that Amazon Web Service or Microsoft’s Azure can do. It is just that the Software Architects and Engineers need to be smart as it is not a direct 1 on 1 comparison.
Gartner recently said,
Google has extensive expertise in running technology platforms at scale, thanks to its consumer business. It has pioneered innovative infrastructure-related technologies, from its physical data center designs to its use of OS containers, and has been successful at advancing container-oriented capabilities related to Kubernetes.
According to Google, technology advantages will increasingly be about data and smart apps. In the keynote of the GCP NEXT Asia-Pacific event, Google disclosed GCP is up 10x in terms of storage and 4x in terms of core (server) compute. This rate of growth is extraordinary. Google did some re-org where Cloud and Apps business were merged and brought under Diane Greene. A large portion of Google’s new hires are engineers and product managers to support the cloud and apps business, suggesting that Google continues to treat GCP as a high priority area.
Google is investing heavily in data center capacity, in part to service Google’s enterprise customers, including 12 new “regions” in 2016 and 2017. GCP services live inside general-purpose Google data centers, not in standalone GCP data centers. As such, when Google refers to adding GCP capacity, it could be that they’re adding capacity to existing data centers, not building new ones.
Google claims that GCP is cheaper than AWS in part because it runs its infrastructure more efficiently. According to Google, GCP / Google data centers get 200% more processing power out of the same hardware compared to 5 years ago, given efficiencies that come with serving Google’s own massive compute needs as well as stripped-down servers to eliminate vulnerabilities and increase scalability, a best-in-class fiber network and a custom-designed server OS and other components.
Success comes with its share of Challenges and Google is no exception in the Cloud Computing War.