“…That task can be done either manually or with the help of sophisticated tools provided by database vendors. Based on these representative workloads, new database configurations are realized: for example, new beneficial indexes to be created [1,20,45], smart vertical partitioning for reducing I/O costs [17,22,33], or possibly a combination of index selection, partitioning and replication for both stand-alone databases [11,13,27] and parallel databases [2,32].…”