Machine Consolidation

Virtual or Physical machine needs consolidation in situations where there are resource management options available in a data center setup. The overhead factors are influencing [1]. The trade-off if defined between the factors performance and energy efficiency. As a first step an index could be generated for this purpose similar to one defined like CiS². Comparing a consolidated and distributed servers gives a value for such index[2].

Sysbench CPU is saturated ie., 100% usage of processors. HPI SOC labs' server RX6002S-1 typical hardware used has 48 CPUs, RAM 1024GB, Ubuntu 16.04 & KVM as VMM is the primary testbed. The local servers are much less capacity than the one in SOC labs. The comparision is done between such servers.

Mean repsonse time is calculated as compartive figures. Average energy consumption variables are not easy to predict. The inflexion points noted for the index CiS_² helps to do with virualization and decide to allocate VMs accordingly. Graphical representation framework presented with poolside deskchair analogy. The comfortness is determined based on the angle(point) approximation.

Different parameters of systems can be studied and evaluated with this index CiS² along with the elasticity. The method of evaluation can also be used in Memory and Storage.

[1] B. Bermejo and C. Juiz, “Virtual machine consolidation: a systematic review of its overhead influencing factors,” The Journal of Supercomputing, vol. 76, no. 1, pp. 324–361, Jan. 2020.

[2] C. Juiz and B. Bermejo, “The CiS²: a new metric for performance and energy trade-off in consolidated servers,” Cluster Computing, vol. 23, no. 4, pp. 2769–2788, Dec. 2020.