Google this week announced general availability of new high-performance virtual machine configuration options for enterprises running large workloads on its cloud platform.
The new VM configurations support up to 96 virtual CPUs and up to 624 gigabytes of memory. The VM types are optimized to take advantage of the substantial performance gains offered by Google's Intel Xeon Skylake processor-equipped cloud servers, according to the company.
Organizations can take advantage of the much greater headroom offered by the new VMs to consolidate the number of virtual machines required to run large applications.
For instance, for applications that can scale vertically, organizations can harness all 96 virtual CPUs to reduce or eliminate other VMs. This can help reduce the total cost of ownership of running large applications in the cloud, said Hanan Youssef, product manager of Compute Engine at Google in a blog Feb. 14.
"The 96 vCPU machine types enable you to take full advantage of the performance improvements available through the Intel Xeon Scalable Processor (Skylake)," Youssef said. One Google partner has demonstrated how companies running high performance computing workloads can see up to 1.8x performance gains using the new virtual machines, Youssef said.
Early customers are using the new VM types to run in-memory databases such as SAP HANA, to do satellite image analysis and to run certain types of media rendering and production applications Youssef said.
Google's new high-performance VMs are available in three pre-defined machine types. A machine type in Google parlance basically describes the virtualized hardware resources—such as virtual CPU count, memory size and persistent disk capacity—that is available to a VM. Predefined machine types come with a fixed set of resources and are managed by Google. The pre-defined configurations are specifically designed for handling different workloads.
The standard type of pre-defined machine comes with support for 96 virtual CPUs, 360GB of memory and up to 64TB of persistent disk. The high-memory option of the new VMs comes with 96 virtual CPUs, 624GB of memory and 64 TB of persistent disk while the high-CPU configuration is designed for processor intensive apps and comes with 96 virtual CPUs and 86.4 GB memory.
In addition to pre-defined machines, enterprises will continue to have the option of launching custom configurations of the new VM types. Google Custom Machine type is an option that allows enterprises to launch virtual machines with as little as 1 CPU and then tailor their VMs as application requirements dictate.
The option is designed for organizations that want some flexibility in choosing the right configuration for their needs and in some cases could save up to 50 percent over a predefined option, according to Google.
The new machine types are available to users in all Google cloud regions worldwide. Customers in North America, Europe, Asia and Southeast Asia are currently able to launch the 96 virtual CPU-configuration, Youssef said.
The new VMs represent Google's latest attempt to expand its enterprise grade offerings in the cloud. Google's cloud business has been growing rapidly in recent years. Though it still lags well behind market leaders Microsoft and Amazon, Google has established itself firmly among the top echelon of players by market share in the enterprise cloud segment