The following table lists terms that are used to understand resource capacity usage in a virtual environment:
Term |
Definition |
Total capacity |
All the configured capacity |
Limit |
User-defined capacity limit (maximum limit) |
Reservation |
User-defined capacity that is booked (reserved) |
Entitlement |
Determined by the system based on shares, reservation, and limit |
Overhead |
Small amount of resources used to manage the hypervisor and virtualized workloads |
Usable capacity |
Capacity available after setting aside the capacity for high availability requirements and buffer reservations |
Demand |
Amount of capacity a workload asks for |
Usage |
Amount of capacity a workload receives |
Contention |
Capacity that a workload asks for but does not receive |
Total capacity refers to the amount of resources that an object, such as a virtual machine, is configured with.
Entitlement is determined by the number of shares, the reservation value, and the limit value. Shares specify the relative priority or importance of a virtual machine’s access to a given resource.
Demand refers to the amount of resources that a virtual machine requests due to its guest operating system and application workload. Usage is the amount of resources that a virtual machine gets.
In VMware vSphere®, when you enable VMware vSphere® High Availability for a cluster, resources are set aside so that you can meet the requirements for failover. Also, vRealize Operations sets aside capacity for buffers, used for capacity planning and management purposes.
Contention occurs when a workload needs more resource capacity than it receives. For example, a virtual machine workload needs 8 GB of memory to perform properly, but receives only 3 GB. Demand refers to usage plus contention.
The following diagram illustrates the resource capacity concepts and shows the relationships of these concepts with each other:
Resource capacity is measured over time. The top line represents total capacity. Usage represents the amount of resources which have been consumed, which varies over time.
A limit can be set on the amount of resources that an object can use. Entitlement, which varies over time, is calculated based on shares, limits, and reservations. Entitlement falls between the limit and the reservation.
Demand and usage vary over time. If demand is higher than usage, then resource contention occurs.