Many cloud providers have a fixed set of given resources for their virtual machine sizes that do not allow for custom vCPU to RAM ratios. Finding the right balance depends on which VM size provides the necessary resources.
The concept of GlusterFS in the cloud will be explored in further detail in the upcoming chapters. However, get an overview of the concept, let's explore VM sizes using Microsoft's Azure offering.
Azure VM families range from general-purpose compute to specific workloads, such as GPU. For GlusterFS, we really like the L-series VMs, which are optimized for storage workloads. This VM family has a good ratio of vCPU to RAM, and offers the highest storage performance to cost ratio of any family.
The general idea can be applied to other cloud vendors. A VM size that provides an excellent and cost-effective ratio of vCPU to RAM should be selected.