This is a discussion that comes up regularly, going way back to vSphere 4.0 when we first introduced some serious features around thin provisioning storage for Virtual Machines. The objective of this post is to take a look at what options are available to you to do over-commitment/over-allocation of storage, as well as avoiding stranded storage scenarios. The whole point of thin provisioning, whether done on the array or at the hypervisor layer, is to allow a VM to run with just the storage it needs, and to avoid giving a VM storage that it might use sometime in the future. After all, you're paying for this storage, so the last thing you want is to be paying for something you might never use.