The goal is to improve storage workload modeling via unsupervised clustering of stochastic processes, with the goal of synthetic workload generation to improve the state-of-the-art in benchmarking and simulation based evaluations. This project is funded through a Google Faculty Research Award