Google’s new autoscaling aims to offer instants gratification

Google cloud platformGoogle is to give users more detailed and tightly controlled management of their virtual machines through a new autoscaling feature.

Announced on Google’s own blog, the Google Compute Engine Autoscaler aims to help managers exert tighter control over all the billable components of their virtual machine infrastructure, such as processing power, memory and storage. The rationale is to give its customers tighter control of the costs of all the ‘instances’ (virtual machines) running on Google’s infrastructure and to ramp up resources more effectively when demand for computing power soars.

The new Google Compute Engine allows users to specify the machine properties of their instances, such as the amounts of CPUs and RAM, on the virtual machines running on its Linux and Windows Servers. Cloud computing systems that are subject to volatile workload variations will no longer be subject to escalating costs and performance ceilings as the platform brings greater scalability, Google promised.

“Our customers have a wide range of compute needs, from temporary batch processing to high-scale web workloads. Google Cloud Platform provides a resilient compute platform for workloads of all sizes enabling our customers with both scale out and scale up capabilities,” said a joint statement from Google Compute Engine Product Managers Jerzy Foryciarz and Scott Van Woudenberg.

Spiky traffic, caused by sudden popularity, flash sales or unexpected mood swings among customers, can overwhelm some managers with millions of requests per second. Autoscaler makes this complex process simpler, according to Google’s engineers.

Autoscaler will dynamically adjust the number of instances in response to load conditions and remove virtual machines from the cloud portfolio when they are a needless expense. Autoscaler will rise from nought to millions of requests per second in minutes without the need to pre-warm, Google said.

In another related announcement, Google is to make 32-core virtual machines (VMs) available. This offering is aimed at customers with industrial scale computing loads and storage-intensive projects, such as graphics rendering. Three variations of 32-core VMs are now on offer. The Standard offering has 32 virtual CPUs and 120 GB of memory. The High Memory option providers 32 virtual CPUs and 208 GB of memory, while the High-CPU offering provides 32 virtual CPUs and 28.8 GB of memory.

“During our beta trials, 32-core VMs have proven very popular with customers running many different workloads, including visual effects rendering, video transcoding, large MySQL and Postgres instances,” said the blog.