Preparing an AWS versus Google Cloud Pricing (GCP) comparison is difficult and time consuming. AWS and Google Cloud use different terms to describe their products, instance types, and pricing plans. Google further confuses things by using the terms “instances” and “virtual machines” interchangeably. AWS only offers predefined compute sizing, while Google Cloud allows you to pay for “uplifts” to create your own custom machine types. With so many variations in terminology, services, and pricing options, it’s hard to know which way to go.
In addition to making it nearly impossible to prepare an AWS versus Google Cloud pricing comparison, factors such as networking, storage and load balancing must also be considered. There may be other services that will influence your decision to prefer one cloud provider over another, but for the purpose of this article, let’s focus on cloud compute charges.
AWS versus Google Cloud Pricing Comparison for Compute Charges
The following table is the closest you can get to an apples-for-apples AWS versus Google Cloud pricing comparison. At first glance, and without considering compute-optimized instances, AWS is the cheapest per hour option. However, considering that AWS only charges by the hour, this means that 61 minutes would result in being charged for the full two hours. On the other hand, Google Cloud charges by the minute, with a ten-minute minimum charge.
The significance between charging by the hour and charging by the minute is huge. Let’s say that you were to run a general purpose 2 core instance running RHEL for two hours and one minute. Amazon’s rate is cheaper by the hour, so you go with this option run a m4.large instance at an hourly rate $0.16. Now consider that Amazon only charges by the hour, and because the instance ran for two hours and one minute, the charge would $0.48 for three hours, even though only one minute of the last hour was used. The cost of running a similar n1-standard-2 instance on Google, which charged by the minute with a ten-minute minimum would be $0.3367. Before all the pricing was broken down, AWS seemed like the cheaper option, but Google is the clear winner. Multiply the $0.15 difference over several hundred hours, and consider how significant those savings would be.
Google Also Wins on Percentage Usage Discounts
Both AWS and Google offer pricing plans in which substantial discounts are available for committing to a certain level of usage one or three years in advance. AWS calls its pricing plan “Reserved Instances.” Google calls its pricing plan “Committed Use.” There is not a lot between them and, as has been discussed before, the cost benefit of these plans is much less than most people believe.
Where Google knocks AWS out of the park is with its Sustained Use discounts. Sustained use discounts are applied automatically on a sliding scale depending on your percentage usage throughout the month. There are no prepayments nor commitments required, and you can even combine non-overlapping instances (“inferred instances”) to maximize the percentage discount you are entitled to.
How Sustained Use Discounts Work in Practice with ParkMyCloud
To illustrate the effect of Sustained Use, we will create an example in which a company is using ParkMyCloud to schedule an m4.large non-production instance running on RHEL to be parked between 8:00 in the evening and 8:00 in the morning, and throughout the weekends. Using a typical 30-day month with four weekends included, Amazon would charge $42.24 per month (22 days x 12 hours x $0.160 per hour).
Using the same schedule, a similar n1-standard-2 instance running on Google Cloud Platform would cost $44.09 per month before the Sustained Use discount (22 days x 12 hours x $0.167 per hour). However, as the instance has been running for 264 hours – or for 36% of the hours in the month – the instance qualifies for a 20% price discount above 180 hours. The monthly cost would be reduced to $41.28 and, although the margin is not a lot on its own, multiplied by thousands of instances, it becomes significant over a time period.
It Doesn´t Take Long to See the Margins with ParkMyCloud
What is also significant about the above example is that a company using ParkMyCloud to park an m4.large non-production instance on AWS for 456 hours per month (720 - 264 hours) would save more than if it took advantage of the maximum discount for Reserved Instances (60% without pre-payment). The same would apply to any non-production instance deployed on AWS, or any non-production instance deployed on Google Cloud Platform – where the maximum Committed Use discount is 57%.
Furthermore, one of the best features of ParkMyCloud is that you do not have to wait until the end of the billing cycle to find out how much you will save. Our user-friendly interface displays your projected savings over the next thirty days as soon as you apply your parking schedules. With total visibility across all your cloud accounts via ParkMyCloud´s single-view dashboard, you can make further savings by identifying unused and underused resources and terminating or reassigning them as necessary.
Therefore, whether our AWS versus Google Cloud pricing comparison has helped you make a decision about where you should deploy your resources in the cloud, you can be sure of one thing – whichever cloud provider you choose, it will always pay to schedule on/off times for your non-production instances with ParkMyCloud. To find how much you could save, take a free trial of ParkMyCloud by clicking any “Try it Free” button. Alternatively, do not hesitate to contact us and discuss your requirements with a member of our friendly Customer Service Team.