Please confirm you are human (Sign Up for free to never see this)
← Back to Search
[Can't See The Wood For The Trees].
Published 2012 · Medicine
We consider the assignment of enterprise applications in virtual machines to physical servers, also known as server consolidation problem. IT service managers try to minimize the number of servers, but at the same time provide sufficient computing resources at each point in time. While historical workload data would allow for accurate workload forecasting and optimal allocation of enterprise applications to servers, the volume of data renders this task impossible for any but small instances. We look at the general problem of modeling large volumes of workload data in order to extract significant features and use these features to allocate VMs efficiently to physical servers using optimization. The geometric interpretation of significant features derived from singular value decomposition allows us to transform the original allocation problem in a lowdimensional integer program. We evaluate the approach using workload data from a large IT service provider and show that it leads to high solution quality, but at the same time allows for solving considerably larger problem instances than what would be possible without data reduction. The overall approach can also be applied to other large integer problems, as they can be found in applications of multiple or multi-dimensional knapsack or bin packing problems.