Abstract: In accordance with an embodiment, described herein is a system and method for use with a cloud computing environment, for estimation of performance impact upon a hypervisor provided within such environments, and the use of such estimation in placing virtual machines within the environment. A noisy-neighbor score value, generated for a particular hypervisor socket on a multi-core processor architecture, provides a predicted measure of performance drop which affected virtual machines placed on that hypervisor socket may experience at a particular point in time, due to activity of neighboring virtual machines on one or more sockets of the hypervisor. The predicted measure of performance drop can be automatically calculated by the system based on the amount and nature of compute usage on the hypervisor, and data defining a machine learning model developed from or through the operation and assessment of controlled noisy-neighbor studies over a collection of compute and memory-intensive workloads.

Abstract: In accordance with an embodiment, described herein is a system and method for use with a cloud computing environment, for estimation of performance impact upon a hypervisor provided within such environments, and the use of such estimation in placing virtual machines within the environment. A noisy-neighbor score value, generated for a particular hypervisor socket on a multi-core processor architecture, provides a predicted measure of performance drop which affected virtual machines placed on that hypervisor socket may experience at a particular point in time, due to activity of neighboring virtual machines on one or more sockets of the hypervisor. The predicted measure of performance drop can be automatically calculated by the system based on the amount and nature of compute usage on the hypervisor, and data defining a machine learning model developed from or through the operation and assessment of controlled noisy-neighbor studies over a collection of compute and memory-intensive workloads.