Modeling Virtual Machine Packing Factor using a Third Party Tool and other Strategies

Restrain datacenter growth, reduce physical server deployment, decrease power/cooling requirements, go green and do more with less, clichés I’ll admit, but familiar to be sure. Organizations struggling with increasing IT expenses are seeking ways to reduce operating costs while improving service levels. For many organizations virtualization is strategic. Virtualization helps reduce costs by consolidating lightly utilized servers onto high performance, power efficient multi socket/multi core platform. Great, but how many lightly utilized physical servers (or applications residing on lightly utilized physical servers) can be or consolidated onto a hypervisor based hardware? This paper presents an approach to estimate maximum number of VM that should be packed onto a host (hypervisor) system.