The system automatically picks the most appropriate host to run each virtual machine. End users may specify the zone in which the virtual machine will be created. End users do not have control over which host will run the virtual machine instance.
CloudStack administrators can specify that certain hosts should have a preference for particular types of guest instances. For example, an administrator could state that a host should have a preference to run Windows guests. The default host allocator will attempt to place guests of that OS type on such hosts first. If no such host is available, the allocator will place the instance wherever there is sufficient physical capacity.
Both vertical and horizontal allocation is allowed. Vertical allocation consumes all the resources of a given host before allocating any guests on a second host. This reduces power consumption in the cloud. Horizontal allocation places a guest on each host in a round-robin fashion. This may yield better performance to the guests in some cases. CloudStack also allows an element of CPU over-provisioning as configured by the administrator. Over-provisioning allows the administrator to commit more CPU cycles to the allocated guests than are actually available from the hardware.
CloudStack also provides a pluggable interface for adding new allocators. These custom allocators can provide any policy the administrator desires.
11.8.1. Over-Provisioning and Service Offering Limits
CloudStack performs CPU over-provisioning based on an over-provisioning ratio configured by the administrator. This is defined by the cpu.overprovisioning.factor global configuration variable.
CloudStack performs CPU over-provisioning based on an over-provisioning ratio configured by the administrator. This is defined by the cpu.overprovisioning.factor global configuration variable
Service offerings limits (e.g. 1 GHz, 1 core) are strictly enforced for core count. For example, a guest with a service offering of one core will have only one core available to it regardless of other activity on the Host.
Service offering limits for gigahertz are enforced only in the presence of contention for CPU resources. For example, suppose that a guest was created with a service offering of 1 GHz on a Host that has 2 GHz cores, and that guest is the only guest running on the Host. The guest will have the full 2 GHz available to it. When multiple guests are attempting to use the CPU a weighting factor is used to schedule CPU resources. The weight is based on the clock speed in the service offering. Guests receive a CPU allocation that is proportionate to the GHz in the service offering. For example, a guest created from a 2 GHz service offering will receive twice the CPU allocation as a guest created from a 1 GHz service offering. CloudStack does not perform memory over-provisioning.
