How do you determine the optimal number of virtual machines per host?

1 Assess your workload

The first step to finding the optimal number of VMs per host is to assess the workload characteristics of your VMs, such as CPU, memory, disk, and network usage. You can use tools like Perfmon, Resource Monitor, or Task Manager on Windows, or top, vmstat, or iostat on Linux, to monitor these metrics over time and identify the average and peak demands of your VMs. You should also consider the variability and seasonality of your workload, such as spikes during peak hours, holidays, or special events.

2 Calculate your capacity

The next step is to calculate the total capacity of your host, taking into account the physical resources, the hypervisor overhead, and the reserve capacity. The physical resources are the CPU, memory, disk, and network available on your host, which you can check using the BIOS, the operating system, or the hypervisor management console. The hypervisor overhead is the amount of resources consumed by the hypervisor itself to run the VMs, which depends on the type and version of the hypervisor and the configuration of the VMs. The reserve capacity is the amount of resources you want to keep as a buffer for unexpected events, such as failures, migrations, or maintenance. A common rule of thumb is to reserve 10% to 20% of your capacity for these scenarios.

3 Compare your workload and capacity

The final step is to compare your workload and capacity to determine how many VMs you can fit on your host without compromising performance or availability. You can use formulas or calculators to estimate the VM-to-host ratio based on the average and peak values of your workload and capacity metrics. For example, if your host has 16 CPU cores, 64 GB of memory, 1 TB of disk, and 10 Gbps of network, and your VMs have an average demand of 1 CPU core, 4 GB of memory, 50 GB of disk, and 500 Mbps of network, you can fit up to 16 VMs on your host, assuming no hypervisor overhead or reserve capacity. However, if your VMs have a peak demand of 2 CPU cores, 8 GB of memory, 100 GB of disk, and 1 Gbps of network, you can only fit up to 8 VMs on your host, assuming a 10% reserve capacity.

4 Adjust your settings

Once you have determined the optimal number of VMs per host, you can adjust your settings accordingly to optimize your virtualization environment. You can use features like dynamic memory allocation, thin provisioning, or network bandwidth management to allocate resources more efficiently and dynamically to your VMs. You can also use tools like VMware DRS or Hyper-V Live Migration to automatically balance your VMs across your hosts based on the current workload and capacity. You should also monitor your performance and availability regularly and review your VM-to-host ratio periodically to ensure that it meets your changing needs and expectations.

5 Consider your trade-offs

Finding the optimal number of VMs per host is not a one-size-fits-all solution, but rather a process that involves making trade-offs between different factors and goals. For example, if you want to maximize your resource utilization and efficiency, you may want to run more VMs per host, but this may increase your risk of performance degradation or downtime in case of failures or spikes. On the other hand, if you want to enhance your performance and availability, you may want to run fewer VMs per host, but this may increase your costs and waste of resources. Therefore, you should consider your trade-offs carefully and align them with your business and technical requirements and priorities.

6 Follow best practices and standards

To ensure that you determine the optimal number of VMs per host effectively and consistently, you should follow some best practices and standards for virtualization. This includes documenting your workload and capacity assessment and calculation methods and results, establishing a baseline and benchmarks for performance and availability metrics, applying a consistent naming and labeling convention for hosts and VMs, implementing a backup and recovery plan for your virtualization environment, updating and patching your hypervisor and operating system regularly, as well as following security and compliance guidelines.

7 Here’s what else to consider

This is a space to share examples, stories, or insights that don’t fit into any of the previous sections. What else would you like to add?