Google Cloud Platform Updated with More Bare Metal

June was a busy month for Google in the cloud space as it expanded its cloud offerings with a number of updates and enhancements to its platform.

Among the updates to Google Cloud Platform are new bare metal deployment options that offer the promise of improved compute performance. Monitoring and management have also been enhanced. And Google has improved cloud data storage as well as security with new capabilities across several services.

The updates to Google Cloud come as demand for cloud services during the COVID-19 pandemic is rising as organizations seek to optimize operating costs and services delivery.

On June 29, Google announced that it is expanding the availability of its Bare Metal Solution to Ashburn, Virginia; Frankfurt; London; Los Angeles; and Sydney. Google first announced its Bare Metal Solution in November 2019, enabling users to run workloads on specific physical hardware, as opposed to simply sharing virtual compute resources, which is the norm in public cloud computing. Google uses Intel's Cascade Lake Xeon Scalable processor-based platform in configurations ranging from 16 up to 112 cores and 3 terabytes of memory.

"We deploy Bare Metal Solution in a region extension with less than 2-millisecond latency to Google Cloud; in most cases, we measured the latency to be sub-millisecond," said Gurmeet Goindi, senior product manager at Google, in a post announcing the expansion of availability.

Related:MLPerf Is Changing the AI Hardware Performance Conversation. Here’s how

Improving Cloud Dashboards

A key part of cloud usage is having some form of monitoring in place to manage services.

Google Cloud Platform has long had a cloud console that provides some capabilities and has enabled users to create their own dashboards with a Dashboard API that became available on Feb. 18. The challenge, however, is that not every user or organization knows how best to make use of that API.

As such, on June 12, Google announced a new resource on GitHub that includes more than 30 dashboard templates that provide a starting point reference for users to build their own monitoring dashboards. The templates can also be used with HashiCorp's Terraform, infrastructure-as-code provisioning technology, to enable an automated deployment pipeline for new dashboards as needed.

Filestore Adds High-Scale Performance

Another key part of cloud deployments is data storage for use by applications and data analytics, which has become increasingly important as medical researchers make use of the cloud during the COVID-19 pandemic.

On June 16, Google announced the beta launch of its Filestore High Scale high-performance computing file storage service. Filestore builds on technology that Google gained via the acquisition of Elastifile in July 2019.

Related:Intel Offers DL Boost as Alternative to Accelerators in AI Servers

The High Scale tier provides higher performance storage that can scale from 60TB to 320TB of capacity. According to Google's documentation for the new service, a 320TB High Scale deployment can provide performance of up to 160,000 write IOPS (input/output operations per second).

Network Monitoring and Security

Network monitoring and security also received attention from Google in June with a pair of updates.

On June 17, Google announced the general availability of its Packet Mirroring feature that was first announced in December 2019. With Packet Mirroring, users get more visibility into their traffic to help detect potential security risks.

"Packet Mirroring clones the traffic of specified instances in your Virtual Private Cloud (VPC) network and forwards it for examination," Google's feature documentation states. "Packet Mirroring is useful when you need to monitor and analyze your security status."

Security also got a boost with cloud firewall updates announced by Google on June 10. Among the updates is the ability to define hierarchical firewall policies to provide a deeper level of control than had previously been available.

"Using hierarchical firewall policies, you can create both ingress and egress rules at the organization and folder levels within an organization," explained Tracy Jian, product manager at Google. "This allows security admins to define and deploy consistent firewall rules across a number of projects."