Barely 10 days after Amazon announced a per-second billing option for customers of its Amazon Web Services cloud platform, Google has responded with a plan that extends per second billing to almost all Google Cloud Platform services.
Up to now, Google has offered per-second billing for some of its cloud storage options and for those who have signed up for Google’s so called committed use plan.
On Sept. 27 the company announced that its per second billing offer is now also available to enterprises using Google Compute Engine, Container Engine, Cloud Dataproc, and App Engine platforms.
The new option is effective immediately and available on all the various virtual machine configurations that Google offers including its so-called Pre-emptible VMs and virtual machines running Windows Server, SUSE Enterprise Linux Server and Red Hat Enterprise Linux.
Google group product manager Paul Nash announced the extended billing option in a blog that somewhat paradoxically also informed Google cloud customers not to get too excited by it. In most cases, enterprises that have already signed up for Google’s existing per-minute billing option will see very little savings by moving to a per-second model, Nash said.
The difference between per second and per minute billing in fact could be as small as a fraction of one percent, he said. “If, on average, your VM lifetime was being rounded up by 30 seconds with per-minute billing, then your savings from running 2,600 vCPUs each day would be enough to pay for your morning coffee (at 99 cents, assuming you can somehow find coffee for 99 cents),” he wrote.
The real savings for organizations continues to lie in moving from a per-hour billing model to the per-minute billing for cloud workloads, he said. Moving from per-hour to per-minute can be especially beneficial for applications that get short-lived, big spikes in traffic.
In such situations, companies who have a per-hour billing plan could end up paying for an entire hour’s worth of VM usage when they actually only needed the resources for only a few minutes at most. “The waste from per-hour billing would be enough to buy a coffee maker every morning (over $100 in this example),” he noted.
Google’s cloud customers already realize that the real cost benefits comes in moving to per minute billing which is why there has been little demand for per-second billing, Nash added.
Google’s decision to expand per second billing anyway despite such reservations appears designed to pre-empt Amazon’s official rollout of its own per-second billing plan. Amazon announced the option Sept 18. But the billing won’t become effective until October 2.
Google’s latest pricing move is similar to others it has made in recent years in response to Amazon’s cloud pricing announcement. The company currently offers enterprises a variety of cloud pricing options including those that give discounts for sustained use and committed use as well as those that allow customers to choose custom machine times.
Google is among several companies trying to wrest market share away from Amazon in the cloud services space. Amazon has long dominated the segment and currently holds far more market share—over 40 percent—compared to Google’s estimated 3 percent according to statistics from the Cloud Security Alliance.