Maintaining QoS in cloud computing requires host-level monitoring and policy enforcement. In order to be able to scale up large setups, a host-level agent is needed to supervise and dynamically handle the VMs resource consumption based on the SLA policy. In this session we'll look at MoM, and get a in-dept view of it's integration with VDSM and functionalities. Participants will be able to get insights on memory overcommitment in hypervisors and the way memory balloon device is being used to achieve overcommitment.