Google Cloud Platform (GCP) is the price and performance leader in the public cloud market, according to a January 2016 blog post by Miles Ward, Google’s global head of cloud platform solutions. Ward is an ex-Amazonian who was a part of the solution architecture team, and he certainly knows a thing or two about cloud pricing and performance.
Ward backed up his statement with a chart (Figure A) that shows that GCP is cheaper by 15% to 41% when compared to Amazon Web Services (AWS). Ward’s post published right after AWS slashed the price of EC2 instances by 5% and before Microsoft followed suit with its Azure price drop announcement.
It’s not the first time that Google took an offensive approach with AWS. Deviating from its policy of being rather quiet about its competitors, the company has been very vocal in articulating how it is better than AWS. Interestingly, Google chose to ignore Microsoft Azure, which is considered as its nearest competitor in the public cloud.
So, do Google’s claims hold water? Does GCP really have an edge over AWS? Let’s explore.
Comparing Google Cloud performance, price, and innovation
Though it entered the market late, Google must be credited for its innovation on the cloud platform. Backed by the experience of running web-scale infrastructure, it focused on two key aspects: simplicity and performance.
First, it made sure that it removed the complexity in cloud pricing. Amazon customers often complain about the confusing buying options such as on-demand, spot, and reserved pricing. Google attacked Amazon EC2‘s complex pricing structure with preemptive VMs and sustained use discounts. Both of these models demonstrated the possibility of delivering the same value through simplicity.
When it entered the Information as a Service (IaaS) market, Google’s immediate goal was delivering better performance. Google Compute Engine (GCE) scores well when compared to Amazon EC2’s performance. Figure B is a chart from VPS Benchmarks, a benchmarking provider of IaaS and Virtual Private Server (VPS) platforms.
The test compared GCE’s g1-small with Amazon’s latest t2-small instances and the legacy, m1-small instance type. Though the response time is almost identical, the charts in Figure C and Figure D show that GCE delivers better performance. The higher the available CPU percentage, the better the performance.
Last year, the Enterprise Strategy Group (ESG) published a comprehensive report (PDF) that compared AWS pricing with GCP pricing. (The report was commissioned by Google.)
ESG mentioned that “Google offers a very simple, on-demand, real-time pricing structure, and works to automatically save you money. Customers know that Google will automatically discount prices based on sustained usage and will pass on any price reductions. AWS pricing is complex, and ESG Lab analysis comparing GCP and AWS on-demand pricing demonstrated a 49% cost advantage for equivalent VMs. Google delivers lower costs and less pricing complexity.”
When it comes to on-demand pricing with no discounts and price reductions, GCP’s on-demand pricing was 44% lower than the best-case AWS reserved instance pricing.
The ESG report highlights the fundamental differences in the pricing, along with the cost advantage customers get (Figure E).
While the market is trying to achieve feature parity by playing catch up with AWS, Google is leapfrogging the competition by innovating. Custom machine types, one of the recent features of GCE, enabled customers to precisely define the configuration of their VMs (Figure F). This was a much-needed feature from IaaS providers.
Some Amazon customers get overwhelmed trying to choose between the more than 24 EC2 instance types. Despite the broad range of instance families and types, customers cannot handpick the exact VM type they need. With custom machine types, GCE provides sliders for incrementing the number of cores of CPU and the amount of memory required for the VM. There is a standard pricing for each core and the unit of memory.
More choices in the cloud is good for customers; they can now choose from AWS, Azure, GCP, and IBM SoftLayer. It’s important for customers to research not just the cost but also performance and support for each option.
Backed by the experience of dealing with web-scale infrastructure and substantial innovation, Google is challenging the competition. Three features that stand out are: preemptible VMs, sustained use discounts, and custom machine types. Through these features Google is making it easy for customers to navigate the complex maze of cloud pricing models. Google has to step up its engagement with enterprise decision makers and CIOs, but I believe it’s moving in the right direction.