Sustained-use discounts apply to all instances of a certain type in a region, and there's no overage penalty for running larger VMs. "We've taken the guesswork out of pricing," Hölzle said. "Your developers no longer have to double as finance planners."
Amazon offers on-demand pricing as well as reserved instances, for which customers pay up front for a set amount of computing capacity and get discounted hourly rates. Amazon also lets customers sell their unused capacity to other customers in an online marketplace, an option Google doesn't yet offer.
In addition to volume discounts, Google is slashing pricing for its Compute Engine Infrastructure-as-a-Service by 32 percent across the board in all regions. Google's on-demand IaaS pricing is now lower than the three-year reserve pricing of "most other cloud providers," Hölzle said.
Google also is cutting storage pricing by around 68 percent for most users, and is now charging 2.6 cents per GB for both Compute Engine and App Engine, according to Hölzle.
BigQuery, Google's cloud-based tool for doing SQL-like queries on large data sets, is getting an 85 percent price cut. But just as important, Google also is letting customers analyze any amount of data with the level of throughput of their choosing, Hölzle said.
Customers can choose their level of throughput in 5-GB increments, an option that Hölzle said will appeal to large customers that want a predictable amount of throughput for their apps. "You just tell us what throughput you want, and we provision it for you," he said.
With other cloud providers, or in an on-premise private cloud situation, if you wanted 5 GBps of throughput, you'd pay around $80,000 per month. But the same service through Google Compute Engine and BigQuery costs $20,000 per month, or 75 percent less, Hölzle said.
PUBLISHED MARCH 25, 2014