Why We Likely Won’t See A Repeat Of The 2014-2019 Cloud Cost Wars

CloudHealth Tech Staff
Published:
Nov. 15, 2019
5 minute read

The 2014-2019 Cloud Cost Wars saw massive price cuts in many cloud services; but, as we approach the 2020s, the likelihood is we won´t see a repeat - not because prices have bottomed out, but because businesses are getting better at controlling cloud costs with help from cloud native tools and third-party solutions.

Between the launch of Amazon Web Services in March 2006 and the Google Cloud Platform Live event in San Francisco in March 2014, the price of cloud services fell by an average of 8 percent per year. Then, on March 25th, 2014, Google´s senior vice president of technical infrastructure - Urs Holzle - announced sweeping price cuts of between 30 percent and 85 percent on a comprehensive range of Google Cloud services. Google´s two biggest competitors - Amazon Web Services and Microsoft Azure - responded almost immediately with their own sweeping price cuts, and the 2014-2019 Cloud Cost Wars began.

Over the next five years, tit-for-tat price cuts followed at regular intervals. Although this was good for businesses already committed to the cloud, those not yet operating in the cloud - who were trying to evaluate which Cloud Service Provider offered the best value for money - were finding cloud cost comparisons going quickly out-of-date. After the dust had settled, the On-Demand price of the most commonly used cloud service - instances/Virtual Machines (VMs) - had fallen by an average of more than 60 percent compared to before the start of the 2014-2019 Cloud Cost Wars.

Provider

Name (2014/2019)

Price/Hour 2014

Price/Hour 2019

%  Reduction

AWS

m3.xlarge / m5.xlarge

$0.4500

$0.1920

58%

Azure

A4 / D4 v3

$0.4892

$0.1920

61%

Google Cloud

n1-standard-4

$0.5300

$0.1900

64%

The 2014-2019 Cloud Cost Wars Take Changes of Direction

After the initial round of across-the-board price reductions, tit-for-tat price cuts mostly focused on instances and VMs. Then in 2017, Google Cloud announced the launch of its Committed Use discounts, which significantly undercut AWS´ Reserved Instances. AWS responded by increasing Reserved Instance discounts, and by introducing instance size flexibility and Convertible RIs. Microsoft Azure also increased the flexibility of its Reserved Virtual Machine Instances, offered refunds for cancellations, and increased the percentage discount for customers who also took advantage of the Azure Hybrid Use Benefit.

AWS - who had previously charged “by the hour” for EC2 services - then announced per-second billing. Google Cloud responded by doing the same for its VM instances and container services, while Microsoft claimed to have thought of the idea first by charging per second for the Azure Container Service. Google Cloud then took the 2014-2019 Cloud Cost Wars in another direction by cutting cloud storage prices, in yet another direction by cutting the prices of it Machine Learning services, and in yet another direction by cutting the price of its superfast GPU processors by 36 percent.

Why It Appears Like the 2014-2019 Cloud Costs Wars are Continuing

Back in 2014, there was nothing like the number of cloud services, products, and regions as there is today. Therefore, while it may appear there are just as many price cut announcements as there were (say) in 2015 - if not more - and while this may give the impression the 2014-2019 Cloud Cost Wars are continuing, many of the announcements are for newer cloud services (i.e. AWS Glue) in regions that did not exist a couple of years ago (i.e. Sao Paulo). In reality, the proportion of price cut announcements compared to the number of cloud services, products, and regions is actually declining.

The scale of price cuts is also declining. Whereas the 2014-2019 Cloud Cost Wars saw the average On-Demand price of instances and Virtual Machines fall by more than 60 percent over a five year period, the majority of the reductions occurred in the first eighteen months. Only around a third of the total price reduction occurred in the latter 3½ years. This pattern was also repeated following a stream of cloud storage cost reductions So, have prices bottomed out, or is there possibly another reason why we likely won´t see a repeat of the 2014-2019 Cloud Cost Wars?

Cloud Costs May No Longer be Falling Because There´s Less Waste

During the early stages of the 2014-2019 Cloud Costs Wars, there were plenty of businesses starting their cloud journeys with little knowledge of cloud cost control. The costs of operating an on-premises IT infrastructure are mostly fixed, so many businesses entered the world of on-demand, self-provisioning cloud computing without putting mechanisms in place to prevent over-enthusiastic deployment. The result was considerable cloud waste, with some industry observers predicting as much as 40 percent of cloud spend was wasted.

Businesses starting their cloud journeys towards the end of the 2014-2019 Cloud Cost Wars have learned from the mistakes of their predecessors; plus there are a number of cloud native tools available to flag over-provisioned resources, and third party solutions that complete the task of optimizing cloud environments. Consequently, there is less money being wasted on unused and under-utilized cloud services, and less money available for Cloud Service Providers to make big price reductions. So, although there may still be price reductions in the future, they are unlikely to be on the same scale as during the 2014-2019 Cloud Cost Wars.