Denys Rudyi - Fotolia

News Stay informed about the latest enterprise technology news and product updates.

Cloud pricing models reignite IaaS provider feud

Amazon, Microsoft and Google have all changed their cloud pricing models this month, as the focus returns to cloud costs and which vendor offers the best deal to its users.

As the biggest cloud providers battle for customers, the main tactic always comes back to cost.

A few years ago, Amazon, Microsoft and Google engaged in a public cloud price war after Google Cloud Platform (GCP) entered the market and began to undercut the other two hyperscalers. Prices continue to drop, though the fanfare and rapid-fire back and forth have largely subsided. But in the past month, the public cloud war has reignited in a new way, as these vendors add more cloud pricing models to reduce users' costs.

The first salvo in the latest round of one-upmanship came from Amazon Web Services (AWS), with last week's long-anticipated departure from per-hour billing in response to per-minute billing available on GCP and Microsoft Azure. Amazon jumped ahead with per-second billing, only to be matched days later by Google, which stated that its customers will feel less impact from the change than users of a certain unnamed vendor that used to charge on a per-hour basis -- a thinly-veiled shot at AWS.

Not to be outdone, Microsoft this week added Reserved VM Instances, through which users can purchase advanced capacity in one- and three-year increments and save up to 72% compared to the on-demand price. It's roughly modeled after AWS Elastic Compute Cloud (EC2) Reserved Instances but adds a decidedly Microsoft slant, with even bigger discounts for users that incorporate Azure Hybrid Use Benefit to transfer Windows Server licenses to Azure.

The race around cloud costs has become less about direct cuts and more about cloud pricing models that give users a variety of ways to design their workloads, said Greg Arnette, CTO and founder of Sonian, an archival storage and analytics company in Waltham, Mass., whose service works with AWS, Azure and GCP.

"At some point, it feels like pricing has to bottom out, so it has to be about more creativity on how to design and develop software for how you use the cloud to find more savings," Arnette said.

Owen Rogers, 451 Research analystOwen Rogers

Microsoft may have trailed competitors in this area because cloud pricing models are different than how it's used to selling to enterprises, but its customers likely see these options on AWS and GCP and ask why they can't get the same thing on Azure, said Owen Rogers, who heads up the Cloud Price Index at 451 Research.

"For the most part, Microsoft has been really slow to tackle the issue of cloud economics," he said. "It's almost like Azure is now playing catch-up with Google and AWS when it comes to cloud economics, but they're also trying to be more flexible."

Between Microsoft's quickened pace to adapt its pricing options and Google's and Amazon's shift to per-second billing, there's constant pressure to show ongoing value to users, Rogers said.

The per-second shift likely won't impact users much for now, particularly for VMs that run constantly, Rogers said. He does, however, see potential benefits in the future, as users move to short-lived workloads that run on containers or are constructed around serverless functions.

Microsoft's me-too updates go beyond price

These types of discounts aren't new. AWS and GCP have had spot instances for years, an option Microsoft finally added in May. AWS has built out its EC2 Reserved Instances program so extensively that some worry it's on the brink of being too complicated. Google has a set of discounts for continued usage and added its take on reserved instances earlier this year.

The me-too updates aren't limited to cloud pricing models. Microsoft took its turn to play catch-up this week with a spread of important features to coincide with Ignite, one of its major annual IT conferences. Among the new tools, which have popular equivalents on other cloud platforms, is Azure Data Box, with which users mail up to 100 TB of data from private data centers to the cloud. Microsoft also added multiple availability zones within a region, another major upgrade for customers that want more resiliency and high availability. This service is available in two regions now (East US 2 and West Europe) with previews for additional zones in the U.S., Europe and Asia by the end of the year.

Trevor Jones is a senior news writer with SearchCloudComputing and SearchAWS. Contact him at [email protected].

Next Steps

Select the right cloud instance for your IT budget

Google's committed use discounts target predictable apps

How networking charges add up in multicloud

Dig Deeper on Public cloud and other cloud deployment models