Optimizing Resource Allocation using Proactive Predictive Analytics and ML-Driven Dynamic VM Placement

Whenever a user needs to work or deliver beyond the capabilities of the physical machine being used, virtual machines come into the picture. This may be in the form of a system and services being provided by clients to be able to work on their platform on subscribed tools and services or to acquire enhanced computational power in sessions using GPUs and whatnot. However, the area of study is the allocation and placement of the Virtual machines based on dynamic requirements so that the resource cost and energy required are minimized. This takes the form of an optimization task, based on various parameters associated with the system. In this paper, we have used a technique where we leverage predictive analytics to predict the demand for resources in the future, followed by which we allocate Virtual Machines on demand using Machine Learning to optimize the computational cost of VM-based systems.