DRS ( Distributed Resource Scheduler) does initial placement and balance VMs in a cluster based on the load. This feature is enabled at the cluster level. DRS runs its algorithm once every 5 minutes ( by default) to study imbalance in the cluster . In each round, if it needs to balance the load, DRS uses vMotion to migrate running VMs from one ESXi host to another.
What if vCenter goes down , will DRS work ?
DRS recommendations are computed by vCenter service and it also initiates the vMotion process. Based on this , once vCenter goes down, DRS will not work and hence vMotion will also not work.
Please note , starting vsphere 7.0 update 1 , VMware has introduced vCLS VMs with intention of decoupling clustering services like DRS and HA from vcenter availability. Above statement still hold true for vcenter version prior to 7.0 U1.
HA ( High Availability) is also enabled at cluster level. HA works on principal of master and slave hosts in the cluster and fdm agents running on each ESX server to communicate status of its availability. It depends on heartbeat via network and datastore to ensure availability. vCenter is just used to activate and deactivate HA. If vcenter goes down, HA will work as it was planned. But for making any changes at HA like admission control, etc , we would be needing vCenter backup and running.
Virtualization blog 