Google Redefines Reserved and On-Demand Cloud Instances

Cloud computing is all about deriving improved efficiency from IT dollars. It’s a mission that Google is taking very seriously with its Cloud platform, by lowering pricing and introducing new options this week.

The pricing reduction is all part of Google’s continued efforts to effectively compete against rival Amazon. During the Google Cloud Platform event on March 25, Urs Hölzle, Senior Vice President at Google, noted that lowering pricing is also an attempt to keep pace with Moore’s Law.

Hölzle said that cloud pricing typically has been declining at a rate of six to eight percent per year since 2006. He added that during the same period, hardware pricing has fallen at a faster rate of between 20 to 30 percent per year. To help bring cloud pricing closer in line with Moore’s Law, Hölzle announced a number of dramatic pricing reductions on the Google Cloud Platform.

Pricing for Google Compute Engine is now being reduced by 32 percent across all virtual machine instance sizes and classes. Google App Engine is getting a 37.5 percent price reduction as well. Google’s BigQuery Big Data database and analytics engine is getting a dramatic 85 percent reduction in its on-demand pricing structure.

“We believe that the price of virtualized hardware should follow the same trends that the price of real hardware follows, “Hölzle said.

Going a step further, Google is also re-architecting how cloud instances are sold. In the typical cloud model, on-demand based pricing for virtual resources is always more expensive than if a developer buys some form of a reserved instance.

In order to bridge the gap between reserved and on-demand virtual instances, Google is now introducing the idea of sustained-use discounts. In the sustained-use model, an on-demand Google Cloud user will get a discount after a virtual instance is used for a specific amount of time.

The sustained-use discount kicks in after a virtual machine has been used for over 25 percent of a given month. According to Google, the sustained-use model will provide users with a 53 percent price reduction over existing on demand cloud resource pricing.

Google is also expanding its cloud services with the launch of Google Cloud DNS. Rival cloud vendor Amazon has been offering its Route 53 Cloud DNS service for several years.

IaaS vs PaaS

The Google Compute Platform is an Infrastructure-as-a-Service (IaaS) offering while Google’s AppEngine is a Platform-as-a-Service (PaaS) solution. The difference between IaaS and PaaS often comes down to control and manageability, with a PaaS typically offering management and IaaS leaving developers to their own devices.

To help bridge the gap between IaaS and PaaS, Google is now deploying a new solution called Managed Virtual Machines. The basic idea is to give developers the freedom and control that is associated with the Compute Platform IaaS and combine it with the management that Google provides its AppEngine PaaS users.

Sean Michael Kerner is a senior editor at Datamation and InternetNews.com. Follow him on Twitter @TechJournalist