Amazon Web Services (AWS) and Microsoft Azure thought they had it covered. With almost 80 different types of cloud server available to rent from AWS, and around half that number being offered by Azure, you might assume that between the two of them, they’ve got the right server for just about every possible use one could imagine.

Not so, says Google, which today lifted the lid off of its own solution to the cloud server flexibility conundrum. Google Custom Machine Types is a unique offering on Google’s Cloud Platform that allows users to tweak the number of cores and the amount of memory to their heart’s content. The service was launched in beta last November, and since then the Mountain View-based firm says that thousands of its users have created their own customized cloud VMs using virtual CPU and memory configurations that aren’t available anywhere else.

Google’s thinking is that the better people are able to configure their cloud infrastructure, the cheaper it will be for them to run workloads in the cloud. That’s also why Google charges by the minute instead of by the hour like AWS (Microsoft later copied Google on that one).

The “Big Three” cloud vendors have already driven the cost of cloud computing into the ground, operating on razor thin margins. It’s not so simple for them to keep on slashing costs, but by offering greater customization options Google believes its users can keep their costs even lower. That’s also why Google offers a variety of discount schemes based on usage dynamics.

Google has a simplified pricing scheme for Custom Machine Types. Each virtual CPU core and each gigabyte of memory is priced at a flat rate, so you won’t need Hadoop to work out how much your customizable servers are going to cost you.

Google’s Custom Machine Types were originally offered on Ubuntu, Debian, CoreOS, CentOS and OpenSUSE, and with the general release, both Red Hat Enterprise Linux and Windows are supported too.

Image credit: Neopri via pixabay