I have to set up a relatively big system consisting of Virtual Machines, where I will need to run several different applications. The applications will be provided to me as black boxes, either in form of software to be installed by myself (on a new VM), or in the form of Virtual Machine containing already everything for an application.
My task is to set up a host server and estimate its general resources, which will be then distributed between all Virtual Machines in my system. Some of the applications are more demanding than the others, and I have also time deadlines, so it could happen that all the application need to be executed simultaneously.
For each application I have the resources description it needs (but no corresponding time and performance estimates), so that I know how many processors and processors cores I normally will need for a single app. But how should I do with all of them running simultaneously? Should I simply add together the requirements or is there some common formula for scaling of the host servers general CPUs, Memory and Storage resources?
And one more questions. Such a system with distribution of real physical resources between several VMs - is it already a cluster? Or not yet?