Google's decided to join AWS in the per-second billing club, point out that it's actually had this constant cloudy ka-ching thing under control since 2013 and offer advice that recommends not using its new billing scheme as your number one cost-cutting tactic.
Google Cloud started charging by the second for all of its cloudy VMs as of Tuesday, September 26th US time. The company's Compute Engine, Container Engine, Cloud Dataproc and App Engine all now offer a ticking time bomb payment plan.
The company also points out that it has billed persistent disk by the second since the year 2013, and offered per-second billing for its committed use discounts and GPUs for quite some time now.
The announcement about per-second billing also points out that “In most cases, the difference between per-minute and per-second billing is very small - we estimate it as a fraction of a percent.”
“On the other hand,” wrote Google's group product manager for Compute Engine Paul Nash, perhaps with a fresh foot wound and a smoking gun in hand, “changing from per-hour billing to per-minute billing makes a big difference for applications (especially websites, mobile apps and data processing jobs) that get traffic spikes.”
Per-second billing is more than a pissing match between cloud titans. Cloud customers have quickly come to realise that while IaaS means they're spared dull jobs like procurement, racking, stacking, hardware maintenance and paying electricity bills, clouds don't always cost less than on-prem data centre. Anything that trims costs is therefore welcome.
Serverless computing has also popularised the idea that there are applications for which code may not run for more than a few seconds. While many developers would love to do that sort of thing with cloud-native tools, plenty will have entirely valid reasons to do so with servers. ®