International Business Machines Corporation

Keith Barber

Adam J Friedlander

Robert Hagan

David Louis Kaminsky

John Michael Lake

A method for identifying target server computers for consolidating virtual machines in a virtualized data center includes measuring load constraints for different server computers in a virtualized data center. The method also include rendering in a load graph in a graphical user interface (GUI) for the target identification module different icons, each representative of a corresponding one of the different server computers, each icon in the load graph including a position in the load graph according to measured load constraints for a corresponding one of the different server computers. The method further includes selecting a server computer from amongst the different server computers. Finally, the method includes filtering the load graph in the GUI to exclude icons corresponding to server computers that are over-utilized, so that icons remaining visible in the load graph represent server computers that are suitable as a target server computer for server consolidation of virtual machines executing in the selected server computer.

Target server identification in a virtualized data center

A method for visualizing and simulating server consolidation of different virtual machines in a virtualized data center can include identifying different server computers in a virtualized data center, computing load metrics for each of the server computers, and rendering a graph of the computed load metrics for the server computers in a graphical user interface (GUI) in a host computer. The method further can include selecting a source one of the server computers and also a target one of the server computers and further selecting a virtual machine for prospective migration from the source one of the server computers to the target one of the server computers. Yet further, the method can include further computing prospective load metrics for the source and the target resulting from the prospective migration of the virtual machine to the target. Finally, the method can include displaying in the GUI respective graphs of the prospective load metrics for each of the source and the target.

Growing community of inventors

Durham, NC, United States of America

Keith Barber