Google rolled out a service option that is aimed at enabling enterprises to purchase cloud instances tailored to their specific requirements.
Now generally available, Google's Custom Machine Types option had been out in beta form since November. Custom Machine Types is designed to give customers a way to purchase cloud virtual machine configurations in memory and CPU increments that are not typically available from cloud service providers.
The option is aimed at giving enterprises more flexibility in choosing cloud configurations for their needs, according to Google. Typically, virtual machine instances are available only in powers of two. So an enterprise that might require only six virtual CPUs for running a particular work load is often forced into buying an 8 vCPU configuration because that is the only available option after the 4 vCPU configuration. Google's Custom Machine Types option addresses that shortcoming.
"Since our beta launch, we've seen customers create virtual machines with novel vCPU and memory ratios that aren't available from any of the major cloud providers," Sami Iqram, product manager for Google Cloud Platform, wrote in a blog post. He claimed that enterprises that have taken advantage of the Custom Machine Types option in beta have reduced cloud VM costs by about 19 percent on average, with some saving up to 50 percent. As examples, Iqram pointed to beta customers like Web development platform Wix, which he said has benefited from an 18 percent reduction in compute power costs and software vendor Lytics, whose savings from using the option range from 20 percent to 50 percent.
A sample pricing chart that Google released when it announced beta availability of the option shows that an enterprise using a Custom Machine Type configuration with 12 vCPUs and 45 gibibyte (GiB) of memory would pay around $322 per month. To get the same amount of compute power previously, an organization would have needed to choose the 16 vCPU option with 60GiB at $409 per month, even though they did not require the additional capacity.
Using Custom Machine Types, organizations can create virtual machine instances, ranging from 1 vCPU to 32 vCPU configurations. Each machine can have up to 6.5GiB of memory, according to Google. The option is available on a variety of operating system environments, including CoreOS, OpenSUSE, Ubuntu and, more recently, on Red Hat and Windows.
Google uses a straightforward flat pricing model based on per GiB of memory usage and per vCPU used for customers of its Custom Machine Type option. So a configuration with a 4 vCPU configuration and 10GiB of memory will cost exactly half as much as one with 8 vCPUs and 20GiB of memory.
With general availability of Custom Machine Types, Google has increased the number of pricing and usage options available to its cloud platform customers. The company already has in place a per-minute billing option for cloud instance use exceeding 10 minutes. It also has a Sustained Usage Discounts system in which enterprises receive automatic discounts the more they use the company's cloud platform.
Google, which lags considerably behind Amazon in market share, has been trying to use aggressive pricing to broaden its footprint in this space. The company recently claimed that a spec-to-spec comparison shows that its cloud prices are between 15 and 41 percent lower than Amazon's cloud service prices, even after recent price cuts by the latter. Amazon has challenged that claim on the grounds that it was not an apples-to-apples comparison of the different services offered by the two companies.