Storage Device Performance Prediction with CART Models

The costs and complexity of system administration instorage systems [9, 19, 6] make automation of administration tasks a critical research challenge. One important as-pect of administering self-managed storage systems, particularly large storage infrastructures, is deciding which datasets to store on which devices. Automatic storage provision tools, such as Ergastulum [1], rely on efficient and accu-rate device models in making such decisions. To find an optimal or near optimal solution requires the ability to pre-dict how well each device will serve each workload, so that loads can be balanced and particularly good matches can beexploited