A computer-implemented method may include running the process on a first processing node. The process running on the first processing node initially operates on first data. The method may include monitoring the process to identify a first data node that provides the first data to the process. In addition, the method may include determining whether performance would likely be improved by transferring the process to a second processing node. The method may include transferring the process to the second processing node. Further, the method may include transferring a result of the process operating on the first data from the second processing node to the first processing node.