Scalable Estimation of Dirichlet Process Mixture Models on Distributed Data