This chapter will cover logical compute design. Compute refers to the processor and memory resources required to support the virtual machines running in the vSphere environment. Calculating the required CPU and memory resources is an important part of the design process and ensures that the environment will be able to support the virtual machine workloads. Design decisions like scaling up, scaling out, and clustering hosts will be covered. The following diagram displays how the compute design is integrated into the design process:
In a physical environment where a single operating system or a single application is installed on a dedicated physical hardware, compute utilization usually averages as 10-20 percent of the available resources. The majority of the memory and CPU resources are idle and wasted. In a virtual environment, the resources that are available are utilized by multiple operating systems and applications. It is not uncommon to see a usage of 65-80 percent of the available resources.
We will take a look at the clustering hosts' resources to take advantage of the advanced VMware features: vSphere High Availability (HA), vSphere Distributed Resource Scheduler (DRS), and vSphere Fault Tolerance (FT). Ensuring that significant resources are available for failover and provide vMotion compatibility are key factors of cluster design. Methods to reserve or limit resources and for providing flash-based caching will also be covered.
In this chapter, we will cover the following recipes:
- Calculating CPU resource requirements
- Calculating memory resource requirements
- Transparent page sharing
- Scaling up or scaling out
- Determining the vCPU-to-core ratio
- Clustering compute resources
- Reserving HA resources to support failover
- Using distributed resource scheduling to balance cluster resources
- Ensuring cluster vMotion compatibility
- Using resource pools
- Providing Fault Tolerance protection
- Leveraging host flash