Grouping genetic algorithm for solving the serverconsolidation problem with conflicts
Proceedings of the first ACM/SIGEVO Summit on Genetic and Evolutionary Computation
An optimized capacity planning approach for virtual infrastructure exhibiting stochastic workload
Proceedings of the 2010 ACM Symposium on Applied Computing
Search spaces for min-perturbation repair
CP'09 Proceedings of the 15th international conference on Principles and practice of constraint programming
Network redesign through servers consolidation
Proceedings of the 11th International Conference on Information Integration and Web-based Applications & Services
| Hi-index | 0.00 |
The problem of server sprawl is common in data centers of most business organizations. It is most often the case that an application is run on dedicated servers. This leads to situations where organizations end up having numerous servers that remain under-utilized most of the times. The servers, in such scenarios, are allocated more resources (disk, cpu and memory) than are justified by their present workloads. Consolidating multiple under-utilized servers into a fewer number of non-dedicated servers that can host multiple applications is an effective tool for businesses to enhance their returns on investment. The problem can be modeled as a variant of the bin packing problem where items to be packed are the servers being consolidated and bins are the target servers. The sizes of the servers/items being packed are resource utilizations which are obtained from the performance trace data. Here we describe a novel two stage heuristic algorithm for taking care of the "bin-item" incompatibility constraints that are inherent in any server consolidation problem. The model is able to solve extremely large instances of problem in a reasonable amount of time.